ORIGINAL ARTICLE

Maternal Use of Acetaminophen, Ibuprofen, and

Acetylsalicylic Acid During Pregnancy and Risk of
Cryptorchidism
Morten Sndergaard Jensen,a,b Cristina Rebordosa,c Ane Marie Thulstrup,b Gunnar Toft,b
Henrik Toft Srensen,d,e Jens Peter Bonde,f Tine Brink Henriksen,a and Jrn Olseng,h
Background: Cyclooxygenase (COX) inhibitorsacetaminophen,
ibuprofen and acetylsalicylic acid have endocrine-disruptive properties in the rainbow trout. In humans, aspirin blocks the androgen
response to human chorionic gonadotropin (hCG), and, because
hCG-stimulated androgen production in utero is crucial for normal
testicular descent, exposure to COX inhibitors at vulnerable times
during gestation may impair testicular descent. We examined
whether prenatal exposure to acetaminophen, ibuprofen, and acetylsalicylic acid was associated with increased occurrence of cryptorchidism.
Methods: Our study used data on 47,400 live-born singleton sons of
mothers enrolled in the Danish National Birth Cohort during 1996
2002. Cryptorchidism was identified in 980 boys during childhood,
of whom 565 underwent orchiopexy. The use of acetaminophen,
ibuprofen, and acetylsalicylic acid during pregnancy was assessed in
3 computer-assisted telephone interviews and 1 self-administered
questionnaire. We estimated adjusted hazard ratios (HRs) of cryptorchidism by Cox regression analysis.
Results: Exposure to acetaminophen during both the first and
second trimesters was associated with increased occurrence of

Submitted 20 March 2010; accepted 7 June 2010; posted 30 August 2010.

From the aPerinatal Epidemiology Research Unit, Departments of Obstetrics and Pediatrics, Aarhus University Hospital, Skejby, Denmark;
b
Department of Occupational Medicine, Aarhus University Hospital,
Aarhus, Denmark; cArizona Respiratory Center, University of Arizona, Tucson, AZ; dDepartment of Clinical Epidemiology, Aarhus
University Hospital, Aarhus, Denmark; eDepartment of Epidemiology, Boston University, Boston, MA; fDepartment of Occupational
and Environmental Medicine, Bispebjerg Hospital, University of
Copenhagen, Copenhagen, Denmark; gThe Institute of Public Health,
University of Aarhus, Aarhus, Denmark; and hDepartment of Epidemiology, School of Public Health, UCLA, Los Angeles, CA.
The Danish National Research Foundation has established the Danish Epidemiology Science Centre that initiated and created the Danish National Birth
Cohort. The cohort is furthermore a result of a major grant from this
Foundation. Additional support for the Danish National Birth Cohort is
obtained from the Pharmacy Foundation, the Egmont Foundation, the March
of Dimes Birth Defects Foundation, the Augustinus Foundation, and the
Health Foundation. Aarhus University, Faculty of Health Sciences, supported Morten Sndergaard Jensen with a PhD scholarship.
Correspondence: Morten Sndergaard Jensen, Perinatal Epidemiology Research
Unit, Departments of Obstetrics and Pediatrics, Aarhus University Hospital,
Skejby, Brendstrupgaardsvej 100, DK-8200 Aarhus N, Denmark. E-mail:
morten@sondergaard-jensen.dk.
Copyright 2010 by Lippincott Williams & Wilkins
ISSN: 1044-3983/10/2106-0779
DOI: 10.1097/EDE.0b013e3181f20bed

Epidemiology Volume 21, Number 6, November 2010

cryptorchidism (HR 1.33 95% confidence interval 1.00

1.77). Exposure for more than 4 weeks within the postulated
time-window of programming testicular descent (gestational weeks
8 14) was associated with a HR of 1.38 (1.051.83) for cryptorchidism. Exposure to ibuprofen and acetylsalicylic acid was not associated with cryptorchidism.
Conclusion: Maternal intake of acetaminophen for more than 4
weeks during pregnancy, especially during the first and second
trimesters, may moderately increase the occurrence of
cryptorchidism.
(Epidemiology 2010;21: 779 785)

ryptorchidism (undescended testis) is one of the most

common abnormalities in newborn boys worldwide,
with etiology in nonsyndromic cases largely unknown.1 A
prevalence of 2%3% has been reported among newborn
boys in Denmark.2 Maternal lifestyle factors and environmental exposures during pregnancy are suspected to interfere
with normal testicular descent and to possibly increase the
risk of cryptorchidism.1,3
Rat experimental models indicate the existence of a
time window in which normal androgen action is crucial to
the programming of future testicular descent.4 For humans, it
is hypothesized that androgen activity around weeks 8 14 of
gestation programs later transinguinal descent during weeks
26 35, which would move the susceptible time window to a
much earlier gestational age than previously assumed.4 In
humans, late transinguinal testicular descent is androgendependent and is most likely the part of the descent process
that is most commonly affected in cryptorchid boys.5 Androgen production in the developing male gonad is driven by
human chorionic gonadotropin (hCG) and luteinizing hormone (LH).6 8 Prostaglandins from the cyclooxygenase
(COX) pathway have been associated with testosterone production in male rats,9 and in animal Leydig cells hCGinduced testosterone production is modulated by prostaglandins.10,11 The mechanisms in humans, however, are not yet
fully elucidated.12 The COX inhibitorsacetaminophen, ibuprofen, and acetylsalicylic acid have shown endocrine-disruptive properties in the rainbow trout, affecting steroid
www.epidem.com | 779

Epidemiology Volume 21, Number 6, November 2010

Jensen et al

hormone synthesis.13 Studies in men have shown that, aspirin

(acetylsalicylic acid) inhibits the androgen response to
hCG.14 Since hCG-stimulated androgen production in utero is
crucial to normal testicular descent,7 exposure to any COX
inhibitor at vulnerable times during gestation may interfere
with normal masculinization of the fetus. Acetaminophen,
ibuprofen, and acetylsalicylic are widely used during pregnancy, and are known to cross the placental barrier.15,16
NSAIDs inhibit prostaglandin synthesis through inhibition
of COX enzymes, with variable potency for COX-1 and
COX-2.17 Acetaminophen exerts its COX-2 inhibitory effect primarily through a mechanism different from that of
NSAIDs.18 20
To our knowledge, there are no published studies on the
risk of cryptorchidism in humans following prenatal exposure
to COX inhibitors. One recent analysis of acetaminophen and
congenital malformations was conducted from the same
cohort as used in the present study.21 That study was not
designed to examine cryptorchidism specifically, but rather
major congenital abnormalities, according to the European
Surveillance of Congenital Abnormalities Classification
(EUROCAT).22 Cryptorchidism is considered a minor malformation, and was registered only if it co-occurred with a
major malformation. This excluded most nonsyndromic cryptorchidism cases. In the previous study, the hazard ratio (HR)
of cryptorchidism in the presence of first-trimester acetaminophen exposure was 1.24 (0.79 1.94). We examined whether
prenatal exposure to acetaminophen, ibuprofen, and acetylsalicylic acidincluding exposures during the suggested
male programming window (gestational weeks 8 14)is
associated with an increased occurrence of cryptorchidism.

METHODS
Study Population
The study is based on data from the Danish National
Birth Cohort, a population-based cohort of children born to
women who were pregnant during 1996 2002 and who
intended to carry their pregnancy to term.23 At their first
antenatal visit, pregnant women were invited by their general
practitioner to join the cohort. Approximately 50% of all
general practitioners in Denmark participated and nearly 60%
of invited women consented to a series of telephone interviews on exposures during pregnancy and their childs health
status. The interviews occurred around gestational weeks 17
and 32, and at 6 and 18 months after birth. Participants also
completed a self-administered questionnaire at enrollment
(between recognition of pregnancy and first telephone interview). Information provided on this questionnaire was
checked and elaborated during the first telephone interview.
Participants were asked to have prescriptions or medication
packets available at the time of each interview. All Regional Science Ethics Committees in Denmark have approved the Danish national Birth Cohort; before we initi780 | www.epidem.com

ated this study, we obtained approval from the Danish Data

Protection Agency.

Exposure Assessment
We defined exposure to acetaminophen, ibuprofen, or
acetylsalicylic acid as self-reported use of drugs containing
any of these active substances at least once during pregnancy.
Data on exposure were collected both prospectively (before
cryptorchidism could be recognized), by using the enrollment
questionnaire and the 2 telephone interviews during pregnancy, and retrospectively (after cryptorchidism could be
recognized) during the telephone interview that took place 6
months postpartum. The retrospective data concerned only
the period between the second telephone interview (around
gestational week 32) and delivery. Study questionnaires elicited the name of the drug and the timing of use on a weekly
basis, and, taken together, covered from 4 weeks before
pregnancy until delivery. Women were asked, Have you
taken any kind of pain killers? If they answered yes, they
were asked to specify which drug they used from a list of 44
specific painkillers, including acetaminophen, ibuprofen, and
acetylsalicylic acid alone or in combination (both over-thecounter and prescribed drugs). Respondents were also given
the option to provide the names of painkillers or other drugs
not included in the list. For each drug reported, they were
asked to specify in which gestational weeks they were used.
Complementary questions were Have you taken any medicine during pregnancy that we have not yet talked about? A
positive response then elicited the name and gestational
weeks of exposure. Additional questions addressed antirheumatic drugs used for muscle or joint diseases and drugs used
to treat fever, inflammation, or infections. The enrollment
questionnaire asked women about any drug use from 4 weeks
before the last menstrual period until 14 weeks afterward,
including drug name and timing of use. Drugs containing
acetaminophen, ibuprofen, or acetylsalicylic acid were classified according to the gestational week of reported use.
Exposure status in a particular gestational week was rarely
reported in more than 1 interview; when this occurred, exposure was considered present if reported in either of the 2
interviews. Discrepancies of exposure status between interviews for a given gestational week did not affect the trimester-specific definition of exposure. Gestational age was calculated from the first day of the last menstrual period and
from routine obstetric ultrasonography examinations. The
gestational age determined by ultrasonography was used in
case of discrepancy between the 2 measures. We focused on
ibuprofen and acetylsalicylic acid among the NSAIDs because they are the most commonly used in this population and
because they are also available over-the-counter in Denmark.

Covariates
We decided a priori to adjust for the following potential
confounders: maternal age in years at time of childbirth
2010 Lippincott Williams & Wilkins

Epidemiology Volume 21, Number 6, November 2010

(25, 2529, 30 34, and 35 years)24; household occupational status (higher- or middle-grade professionals, skilled
workers and students, and unskilled workers and unemployed); parity before birth of the index boy (0 and 1)25,26;
time to pregnancy (0 5, 6 12, 13 months); treatment of
infertility (yes vs. no)24,27; and maternal smoking during
pregnancy (none, 110 and 11 cigarettes per day).27 Table
1 includes the approximate percentages of missing responses
for each covariate.

Outcome Measures
Information on cryptorchidism, other congenital malformations, and surgical procedures undergone by boys in the
cohort was obtained from the Danish National Patient Registry. This registry contains information on all inpatient and
outpatient clinic diagnoses and surgeries performed during
the follow-up period.28 The registry covered 100% of Denmarks hospitals during the study period. The study used 2
endpoints with increasing diagnostic specificity: first, boys
with a diagnosis of cryptorchidism coded in accordance with
the International Classification of Diseases 10 (ICD10 codes:
Q53, Q531, Q531A, Q532, Q532A, and Q539); second, boys
with a diagnosis of cryptorchidism who also underwent
orchiopexy (codes KKFH00, KKFH01, and KKFH10 in the
Nordic Classification of Surgical Procedures). Orchiopexy
indicates that the cryptorchidism is persistent, while the

Prenatal Exposure to Weak Analgesics and Cryptorchidism

diagnosis of cryptorchidism also covers transient cases of

undescended testes that descend spontaneously.

Statistical Methods
Although cryptorchidism is considered a congenital
malformation, not all cases are identified at birth. Some,
including recurrences, are diagnosed and treated throughout
childhood.29 The boys in the Danish National Birth Cohort
had not reached an age where all cryptorchidism cases had
been identified,30 nor had they all attained the same age by
the end of follow-up: the youngest boy was 6 years old and
the oldest 13 years old. To account for variation in follow-up
periods, we estimated crude and adjusted HRs by means of
Cox regression models, using boys age as the time variable.
The boys entered the risk set at birth and were followed until
their age at first diagnosis, death, emigration from Denmark,
or end of follow-up (21 October 2009), whichever came first.
The proportional hazards assumption was checked by visual
inspection of log-minus-log plots.
We analyzed separate and combined exposures to acetaminophen, ibuprofen, and acetylsalicylic acid. For acetaminophen, we were able to study exposure by trimester. For
ibuprofen and acetylsalicylic acid, the number of exposed
boys was small, allowing only dichotomous analysis of exposure. For acetaminophen exposure during the entire pregnancy, we compared no exposure (no acetaminophen expo-

TABLE 1. Characteristics of 47,400 Mothers and Boys in the Danish National Birth Cohort, by Their Use of Acetaminophen,
Ibuprofen, and Acetylsalicylic Acid During Pregnancy, 1996 2002

Characteristic
Maternal age (years)a
25
2529
3034
35
Middle- or higher-grade professionalb
Primiparousc
Time to pregnancy 12 monthsb
Treatment of infertilityb
Smoking during pregnancyc
Diabetes mellitusb
Fever during pregnancyc
Muscle or joint disease during pregnancyc
Inflammation or infection during pregnancyb
Gestational age 37 weeksd
Birth weight 2500 gd
Singleton brother in cohorta
Other congenital malformationsa
Cryptorchidisma,e
Orchiopexya,e

Unexposed
(n 21,504)
%

Acetaminophen
(n 22,449)
%

Ibuprofen
(n 2333)
%

Acetylsalicylic Acid
(n 3135)
%

9.8
39.2
36.6
14.5
64.5
48.2
12.0
5.3
23.0
1.3
21.2
13.4
11.0
5.2
3.1
3.6
7.0
2.0
1.1

9.7
38.1
37.4
14.8
62.5
42.1
11.9
5.4
29.2
1.5
33.2
20.2
16.6
4.3
2.6
3.9
6.5
2.1
1.2

9.5
38.3
36.9
15.3
64.7
52.6
11.9
5.2
32.2
1.7
30.6
24.9
16.5
5.2
3.3
2.2
7.5
1.8
1.0

7.7
35.6
38.4
18.3
66.8
44.5
11.7
4.9
28.8
1.3
30.0
17.3
15.4
4.2
2.9
2.4
6.4
2.4
1.4

Percent of missing responses, coded using a missing category, for each covariate: a0%; b5%7%; c1%5%; d1%.
e
Cumulative incidence proportion at end of follow-up.

2010 Lippincott Williams & Wilkins

www.epidem.com | 781

Jensen et al

sure at any time during pregnancy) with exposure only during

1 trimester (each separately), exposure only during both the
first and second trimesters, exposure only during both the
second and third trimesters, and exposure during all 3 trimesters. By applying this grouping, each woman was counted
only once in this model. Exposure during weeks 8 14 (the
suggested male-programming window) was analyzed in a
separate model, including observations already counted in the
analysis of acetaminophen exposure during entire pregnancy.
In a dose-response analysis, we compared mothers
with increasing number of exposed weeks during gestation
(1, 2 4, 5 8, 9 12, and 12 weeks) to unexposed mothers.
Although the comparison is crudefor example, 8 tablets a
day for 5 weeks was grouped together with 2 tablets a day for
5 weeksit gives some information on accumulated weeks
of exposure. A dose-response analysis was also done for
weeks 8 14 with a cut point of 4 weeks chosen prior to
analysis. In both analyses on weeks of exposure, we excluded
7294 women who (due to changes in the initial enrollment
questionnaire) lacked information on a week-by-week basis
during the first trimester. Combined exposure to more than 1
drug at any time during pregnancy was also analyzed by
comparing those unexposed to any drug to those exposed to
1 drug or to 2 or more drugs. All analyses were adjusted for
potential confounders as described above (in the section on
covariates). We also assumed robust standard errors in all
adjusted analyses to account for clustering in observations
(boys) with the same mothers (n 1715). All analyses were
done separately for the 2 endpoints cryptorchidism diagnosis and orchiopexy.
We performed secondary analyses to determine the
consistency of our results when changes were made in the
analytical approach. Birth weight, gestational age, and other
congenital malformations are associated with cryptorchidism,25,26,31 and we assumed that these are intermediary variables rather than confounders. However, some may argue that
these factors reflect the presence of other unmeasured confounders at baseline that led to the outcomes of interest. We
therefore evaluated whether inclusion of these variables in the
models changed the estimates. We repeated our analyses in a
population of boys without other congenital malformations
(n 44,174) to evaluate potential effects on nonsyndromic
cryptorchidism and to exclude observations already reported
in a previous study.21 To control for confounding by indication, we fitted a model that included 3 variables describing
diseases or conditions that may trigger use of NSAIDs or
acetaminophen: disease in muscles and joints during pregnancy (yes/no), fever during pregnancy (yes/no), and inflammation or infection during pregnancy (yes/no). In addition to
assuming robust standard errors, we performed analyses restricted to the first-born boy in the cohort (n 45,685).
Incomplete information on covariates was handled by including a missing category for each variable in the models to
782 | www.epidem.com

Epidemiology Volume 21, Number 6, November 2010

retain observations in the analysis and thus to avoid bias from

exclusion due to incomplete information. We also performed
analyses restricted to those with complete information on all
covariates (n 42,891) to see if this changed our estimates.
Finally, we excluded a small group of mothers (n 663) with
self-reported diabetes mellitus, because this condition may
act as a confounder.32 We used restriction instead of adjustment because strata with diabetic mothers were small. Statistical analysis was done using Stata 11.0 software (Statacorp, College Station, TX).

RESULTS
Of all pregnancies registered in the Danish National
Birth Cohort (n 101,091), we excluded those that did not
end with at least 1 live birth (n 6380), those with female
offspring (n 46,207), those with twin or higher multiple
births (n 997), and pregnancies where boys or mothers
were not uniquely identifiable (n 107). The remaining
47,400 mother-boy pairs were eligible for analysis. Of these,
46,500 (98%) were followed until study closure on 21 October 2009, 645 were censored due to emigration and 255 were
censored because they died during the study period. Altogether, 980 boys (2.1%) were recorded in the Danish National
Patient Registry with a cryptorchidism diagnosis. Of these,
565 underwent an orchiopexy. Of the mothers, 21,504 (45%)
were not exposed to acetaminophen, ibuprofen, or acetylsalicylic acid during pregnancy; 22,449 (47%) were exposed
some time during pregnancy to acetaminophen, 2333 (5%) to
ibuprofen and 3135 (7%) to acetylsalicylic acid. Among the
mothers using ibuprofen and acetylsalicylic acid during pregnancy, 57% and 50%, respectively, also used acetaminophen
at some time during pregnancy. Characteristics of the study
population are presented in Table 1. Maternal age, presence
of a brother in the cohort, a report of fever, muscle or joint
disease, inflammation or infection, and maternal smoking
were all associated with use of the drugs.
Table 2 presents crude and adjusted HRs for cryptorchidism and orchiopexy according to exposure to acetaminophen, ibuprofen, and acetylsalicylic acid. Exposure to
acetaminophen in both the first and second trimesters was
associated with increased occurrence of a cryptorchidism
diagnosis (HR 1.33 95% confidence interval {CI}
1.00 1.77) and orchiopexy (1.26 0.86 1.84). Exposure in
all 3 trimesters also was associated with an increased (but
weaker) risk of both a cryptorchidism diagnosis (1.17 0.94
1.46) and orchiopexy (1.06 0.78 1.43). Exposure within
any single trimester alone was only weakly associated with
cryptorchidism, if at all. Exposure within gestational weeks
8 14 (the suggested male programming window) was weakly
associated with a cryptorchidism diagnosis (1.14 0.971.34)
and orchiopexy (1.10 0.89 1.36). Exposure to ibuprofen or
acetylsalicylic acid was not consistently associated with cryptorchidism.
2010 Lippincott Williams & Wilkins

Epidemiology Volume 21, Number 6, November 2010

Prenatal Exposure to Weak Analgesics and Cryptorchidism

TABLE 2. Hazard Ratios and 95% Confidence Intervals for Cryptorchidism and Orchiopexy According to Mothers Use of
Acetaminophen, Ibuprofen, and Acetylsalicylic Acid During Pregnancy, Danish National Birth Cohort, 1996 2002
Cryptorchidism Diagnosis
Exposure and Timing
Acetaminophen during pregnancyc
Unexposed throughout pregnancyd
1st trimester only
2nd trimester only
3rd trimester only
Both 1st and 2nd trimester
Both 2nd and 3rd trimester
All 3 trimesters
Acetaminophen during weeks 814c,e
Unexposedd
Exposed
Ibuprofenc
Unexposedd
Exposed
Acetylsalicylic acidc
Unexposedd
Exposed

Orchiopexy
a

No. Cases
(Person-years)

Crude HR

Adjusted HR
(95% CI)b

No. Cases
(Person-years)

Crude HR

Adjusted HRa
(95% CI)b

480 (212,717)
92 (43,683)
58 (21,758)
101 (41,408)
53 (17,664)
28 (13,737)
95 (35,779)

1.00
0.94
1.16
1.07
1.33
0.89
1.16

1.00
0.94 (0.751.17)
1.17 (0.891.54)
1.08 (0.871.33)
1.33 (1.001.77)
0.90 (0.611.32)
1.17 (0.941.46)

274 (214,078)
60 (43,897)
31 (21,911)
60 (41,692)
29 (17,833)
16 (13,806)
49 (36,047)

1.00
1.07
1.10
1.12
1.27
0.90
1.06

1.00
1.06 (0.801.41)
1.11 (0.771.61)
1.12 (0.841.48)
1.26 (0.861.84)
0.90 (0.541.49)
1.06 (0.781.43)

766 (338,222)
189 (71,688)

1.00
1.14

1.00
1.14 (0.971.34)

443 (340,340)
105 (72,201)

1.00
1.10

1.00
1.10 (0.891.36)

937 (400,179)
43 (20,573)

1.00
0.89

1.00
0.88 (0.641.19)

541 (402,725)
24 (20,712)

1.00
0.86

1.00
0.84 (0.561.27)

904 (392,185)
76 (28,567)

1.00
1.18

1.00
1.18 (0.931.49)

522 (394,626)
43 (28,810)

1.00
1.14

1.00
1.15 (0.841.56)

Adjusted for maternal age at childbirth, household occupational status, parity before the index boy, time to pregnancy, treatment of infertility, and smoking during pregnancy.
95% confidence interval calculated from robust standard errors that account for dependency among boys born to the same mother (clustering).
Number with exposure information: acetaminophen (46,413), ibuprofen, and acetylsalicylic acid (47,400).
d
Reference category.
e
Estimated in a separate model, ie including observations already counted in the analysis of acetaminophen exposure during entire pregnancy.
b
c

Analyses of cumulative acetaminophen exposure and

combined exposure to the 3 drugs are presented in Table 3.
No clear dose-dependent pattern was observed, although
cumulative acetaminophen exposure during pregnancy of
more than 4 weeks was associated with HR point estimates
above 1.0. For example, 5 8 weeks of exposure was associated with increased occurrence of a cryptorchidism diagnosis
(HR 1.32 CI 0.971.78) and orchiopexy (1.63 1.13
2.34). Exposure during more than 4 weeks within the suggested male programming window was also associated with
increased occurrence of a cryptorchidism diagnosis (1.38
1.051.83) and orchiopexy (1.44 1.00 2.06). Combined
exposure to more than 1 drug did not add to the risk of
cryptorchidism (Table 3).
We performed secondary analyses to test the validity of
our analytical approach, including (1) adjustment for birth
weight, gestational age, and other congenital malformations,
(2) adjustment for diseases in muscles or joints, fever, or
infections during pregnancy, (3) restriction to the first of
brothers born in the cohort, (4) restriction to mother-boy pairs
with complete information on covariates, (5) restriction to
boys without other congenital malformations, and (6) exclusion of mothers with self-reported diabetes mellitus. The
results of these secondary analyses were all very similar to
those presented and did not alter our conclusions.
2010 Lippincott Williams & Wilkins

DISCUSSION
Maternal use of ibuprofen and acetylsalicylic acid was
not associated with excess occurrence of cryptorchidism, but
exposure to these drugs are not common among pregnant
Danish women, and only crude exposure groups were available for analysis. In contrast, approximately 47% of the
cohort reported use of acetaminophen, allowing analyses of
timing of exposure and cumulative exposure. Overall, we
observed no markedly increased occurrence of cryptorchidism with fetal acetaminophen exposure, but cumulative exposure of more than 4 weeks and exposure during the first and
second trimesters was associated with increased HRs for
cryptorchidism. We found no dose-response associations,
which may be explained by crude dose measurements. The
finding of an excess occurrence among those exposed during
the suggested male programming window (gestational weeks
8 14), is consistent with studies indicating that diethylstilbestrol exposure must start before week 11 to increase the
risk of cryptorchidism.33 However, too few women were
exposed only during weeks 8 14 to allow for estimation of
effects specifically for this time window. The associations for
gestational weeks 8 14 may consequently be confounded by
acetaminophen exposure during other parts of pregnancy.
It is unlikely that selection into this cohort affected the
associations under study, and losses to follow-up were minimal.
www.epidem.com | 783

Epidemiology Volume 21, Number 6, November 2010

Jensen et al

TABLE 3. Hazard Ratios and 95% Confidence Intervals for Cryptorchidism and Orchiopexy According to Weeks of Use of
Acetaminophen and Combined Exposure to Acetaminophen, Ibuprofen, and Acetylsalicylic Acid During Pregnancy, Danish
National Birth Cohort, 1996 2002
Cryptorchidism Diagnosis

Weeks of acetaminophen exposure

Throughout pregnancyc
0 weeksd
1 week
24 weeks
58 weeks
912 weeks
12 weeks
During gestational weeks 814c
0 weeksd
14 weeks
4 weeks
Combined exposure during pregnancy
Unexposedd
1 drug
2 drugs

Orchiopexy

Crude HR

Adjusted HR
(95% CIb)

No. Cases
(Person-years)

Crude HR

Adjusted HRa
(95% CIb)

480 (212,717)
119 (47,139)
86 (36,057)
47 (15,322)
23 (7284)
46 (16,123)

1.00
1.09
1.02
1.32
1.34
1.23

1.00
1.09 (0.891.34)
1.01 (0.811.28)
1.32 (0.971.78)
1.35 (0.882.05)
1.23 (0.901.66)

274 (214,078)
65 (47,476)
44 (36,293)
33 (15,420)
14 (7349)
24 (16,250)

1.00
1.06
0.93
1.65
1.46
1.14

1.00
1.05 (0.801.38)
0.92 (0.671.26)
1.63 (1.132.34)
1.46 (0.852.50)
1.13 (0.741.71)

665 (286,735)
99 (38,226)
54 (16,363)

1.00
1.08
1.40

1.00
1.07 (0.871.33)
1.38 (1.051.83)

380 (288,590)
53 (38,486)
32 (16,506)

1.00
1.02
1.46

1.00
1.02 (0.761.36)
1.44 (1.002.06)

428 (190,405)
467 (195,326)
61 (25,319)

1.00
1.06
1.08

1.00
1.06 (0.931.21)
1.07 (0.821.40)

242 (191,606)
276 (196,556)
31 (25,522)

1.00
1.11
0.96

1.00
1.11 (0.931.32)
0.95 (0.651.39)

No. Cases
(Person-years)

Adjusted for maternal age at childbirth, household occupational status, parity before the index boy, time to pregnancy, treatment of infertility, and smoking during pregnancy.
95% confidence interval calculated from robust standard errors that account for dependency among boys born to the same mother (clustering).
Models excluding 7294 women who lacked first trimester exposure information on a week-by-week basis; women included during entire pregnancy (n 38,365); women
included during gestational weeks 8 14 (n 39,123).
d
Reference category.
b
c

Only 900 (2%) boys were lost to follow-up because of death or

emigration. Furthermore, previous studies have demonstrated
that nonparticipation in cohort studies probably has only small
effects on the internal validity.34 Exposure was assessed prospectively at each interview during pregnancy based on detailed
information concerning periods of use, and included prescribed
and over-the-counter use of acetaminophen, ibuprofen, and acetylsalicylic acid. There may be differential recall bias of exposures reported (retrospectively) in the interview 6 months postpartum, but this would affect only exposures late in the third
trimester. The studys conclusions are unlikely to be affected by
recall bias, because data for most of pregnancy (from 4 weeks
prior to the last menstrual period to around gestational week 32)
were collected before cryptorchidism could be recognized. Even
in the data collected after recognition was possible, differential
reporting of exposures by boys cryptorchidism status is unlikely. It has been shown that the rigorous assessment of drug
usage in our study cohort yields higher rates of reported use than
data from prescription databases. This is not unexpected because
most of these analgesics are sold over-the-counter.35 The inherent imprecision of gestational age assessment would tend to
attenuate associations due to misclassification of the timing of
exposure, and may lead to underestimation of effects. Any
exposure misclassification by mothers regarding usage, names
of drugs, and timing of exposure would also most likely lead to
underestimation of effects.
784 | www.epidem.com

Ascertainment of cryptorchidism was based on routine

data from Danish nationwide registries independent of maternal
exposure. To our knowledge, the validity of the cryptorchidism
diagnosis in Danish registries has never been directly assessed,
but orchiopexy cases encompassing both diagnosis and corrective surgery are considered highly specific. Our cases of cryptorchidism were mostly persistent, since transient cryptorchidism with spontaneous descent within 3 6 months usually is not
reported to the registries. Some transient cases that reoccur as
ascensus testis (recurrent cryptorchidism) were probably included given the follow-up period of our study population.29
Overall, our findings were similar for the cryptorchidism and the
orchiopexy endpoints.
Women who use analgesics for an extended period of
time during pregnancy may be predominantly those with a
chronic disorder. Confounding by indication may be more
prominent in the regular users than among infrequent
users, who consume analgesics sporadically for conditions
such as headache. We aimed to adjust for confounding by
indication in the secondary analyses and results were
essentially unchanged. However, some unadjusted or residual confounding may still explain our findings.
To our knowledge, this is the first study that aimed
specifically to study associations between weak analgesics and
cryptorchidism. Previous studies have mostly addressed congenital anomalies overall, using, for example, the EUROCAT clas 2010 Lippincott Williams & Wilkins

Epidemiology Volume 21, Number 6, November 2010

sification that excludes minor malformations such as cryptorchidism unless they co-occur with major malformations.21,22
Animal and in vitro studies lend some support to the hypothesis
under study. The COX inhibitors, acetaminophen, ibuprofen,
and acetylsalicylic acid have shown endocrine-disruptive properties in the rainbow trout, affecting steroid hormone synthesis.13 Prostaglandins from the COX pathway seem to modulate
testosterone production in animal Leydig cells.9 11,36,37 However, some studies support the hypothesis that COX inhibition
leads to reduced testosterone production36 while others do
not,9 11,37 and the net effect of COX inhibition on testicular
steroid genesis is not clear. One human study in men showed
that COX inhibition by acetylsalicylic acid reduced production
of testosterone, 17OHprogesterone, androstenedione, and dehydroepiandrosterone in response to hCG.14 Whether this applies to other COX inhibitors and to the crucial hCG-induced
testosterone production in utero remains unknown.
In conclusion, our study indicates that cumulative acetaminophen exposure of more than 4 weeks duration, especially
during the first and second trimesters, may moderately increase
the occurrence of cryptorchidism. Exposure to ibuprofen and
acetylsalicylic acid was not associated with cryptorchidism.
REFERENCES
1. Chacko JK, Barthold JS. Genetic and environmental contributors to
cryptorchidism. Pediatr Endocrinol Rev. 2009;6:476 480.
2. Cortes D, Kjellberg EM, Breddam M, Thorup J. The true incidence of
cryptorchidism in Denmark. J Urol. 2008;179:314 318.
3. Jensen MS, Toft G, Thulstrup AM, et al. Cryptorchidism concordance in
mono- and dizygotic twin brothers, full brothers and half brothers. Fertil
Steril. 2010;93:124 129.
4. Welsh M, Saunders PT, Fisken M, et al. Identification in rats of a
programming window for reproductive tract masculinization, disruption
of which leads to hypospadias and cryptorchidism. J Clin Invest. 2008;
118:1479 1490.
5. Amann RP, Veeramachaneni DN. Cryptorchidism in common eutherian
mammals. Reproduction. 2007;133:541561.
6. Clements JA, Reyes FI, Winter JS, Faiman C. Studies on human sexual
development. III. Fetal pituitary and serum, and amniotic fluid concentrations of LH, CG, and FSH. J Clin Endocrinol Metab. 1976;42:9 19.
7. Toppari J, Kaleva M, Virtanen HE, Main KM, Skakkebaek NE.
Luteinizing hormone in testicular descent. Mol Cell Endocrinol. 2007;
269:34 37.
8. Huhtaniemi IT, Korenbrot CC, Jaffe RB. HCG binding and stimulation
of testosterone biosynthesis in the human fetal testis. J Clin Endocrinol
Metab. 1977;44:963967.
9. Didolkar AK, Gurjar A, Joshi UM, Sheth AR, Roychowdhury D. Effect
of prostaglandins A-1, E-2 and F-2 alpha on blood plasma levels of
testosterone, LH and FSH in male rats. Andrologia. 1981;13:50 55.
10. Frungieri MB, Gonzalez-Calvar SI, Parborell F, Albrecht M,
Mayerhofer A, Calandra RS. Cyclooxygenase-2 and prostaglandin F2
alpha in Syrian hamster Leydig cells: inhibitory role on luteinizing
hormone/human chorionic gonadotropin-stimulated testosterone production. Endocrinology. 2006;147:4476 4485.
11. Romanelli F, Valenca M, Conte D, Isidori A, Negro-Vilar A.
Arachidonic acid and its metabolites effects on testosterone production
by rat Leydig cells. J Endocrinol Invest. 1995;18:186 193.
12. Stocco DM, Wang XJ, Jo Y, Manna PR. Multiple signaling pathways
regulating steroidogenesis and steroidogenic acute regulatory protein
expression: more complicated than we thought. Mol Endocrinol. 2005;
19:26472659.
13. Gravel A, Vijayan MM. Salicylate disrupts interrenal steroidogenesis

2010 Lippincott Williams & Wilkins

Prenatal Exposure to Weak Analgesics and Cryptorchidism

14.

15.
16.

17.

18.
19.
20.

21.

22.
23.

24.

25.

26.

27.

28.

29.
30.
31.
32.

33.

34.
35.

36.

37.

and brain glucocorticoid receptor expression in rainbow trout. Toxicol

Sci. 2006;93:41 49.
Conte D, Romanelli F, Fillo S, et al. Aspirin inhibits androgen response
to chorionic gonadotropin in humans. Am J Physiol. 1999;277:E1032
E1037.
Schoenfeld A, Bar Y, Merlob P, Ovadia Y. NSAIDs: maternal and fetal
considerations. Am J Reprod Immunol. 1992;28:141147.
Rayburn W, Shukla U, Stetson P, Piehl E. Acetaminophen pharmacokinetics: comparison between pregnant and nonpregnant women. Am J
Obstet Gynecol. 1986;155:13531356.
Rao P, Knaus EE. Evolution of nonsteroidal anti-inflammatory drugs
(NSAIDs): cyclooxygenase (COX) inhibition and beyond. J Pharm
Pharm Sci. 2008;11:81s110s.
Graham GG, Scott KF. Mechanism of action of paracetamol. Am J Ther.
2005;12:46 55.
Anderson BJ. Paracetamol (Acetaminophen): mechanisms of action.
Paediatr Anaesth. 2008;18:915921.
Aronoff DM, Oates JA, Boutaud O. New insights into the mechanism of
action of acetaminophen: its clinical pharmacologic characteristics reflect its inhibition of the two prostaglandin H2 synthases. Clin Pharmacol Ther. 2006;79:9 19.
Rebordosa C, Kogevinas M, Horvath-Puho E, et al. Acetaminophen use
during pregnancy: effects on risk for congenital abnormalities. Am J
Obstet Gynecol. 2008;198:178.e1 e7.
Dolk H. EUROCAT: 25 years of European surveillance of congenital
anomalies. Arch Dis Child Fetal Neonatal Ed. 2005;90:F355F358.
Olsen J, Melbye M, Olsen SF, et al. The Danish National Birth
Cohortits background, structure and aim. Scand J Public Health.
2001;29:300 307.
McGlynn KA, Graubard BI, Klebanoff MA, Longnecker MP. Risk
factors for cryptorchism among populations at differing risks of testicular cancer. Int J Epidemiol. 2006;35:787795.
Jones ME, Swerdlow AJ, Griffith M, Goldacre MJ. Prenatal risk factors
for cryptorchidism: a record linkage study. Paediatr Perinat Epidemiol.
1998;12:383396.
Akre O, Lipworth L, Cnattingius S, Sparen P, Ekbom A. Risk factor
patterns for cryptorchidism and hypospadias. Epidemiology. 1999;10:
364 369.
Jensen MS, Toft G, Thulstrup AM, Bonde JP, Olsen J. Cryptorchidism
according to maternal gestational smoking. Epidemiology. 2007;18:
220 225.
Andersen TF, Madsen M, Jorgensen J, Mellemkjoer L, Olsen JH. The
Danish National Hospital Register. A valuable source of data for modern
health sciences. Dan Med Bull. 1999;46:263268.
Wohlfahrt-Veje C, Boisen KA, Boas M, et al. Acquired cryptorchidism
is frequent in infancy and childhood. Int J Androl. 2009;32:423 428.
Jensen MS, Bonde JP, Olsen J. Prenatal alcohol exposure and cryptorchidism. Acta Paediatr. 2007;96:16811685.
Weidner IS, Moller H, Jensen TK, Skakkebaek NE. Risk factors for
cryptorchidism and hypospadias. J Urol. 1999;161:1606 1609.
Virtanen HE, Tapanainen AE, Kaleva MM, et al. Mild gestational
diabetes as a risk factor for congenital cryptorchidism. J Clin Endocrinol
Metab. 2006;91:4862 4865.
Palmer JR, Herbst AL, Noller KL, et al. Urogenital abnormalities in men
exposed to diethylstilbestrol in utero: a cohort study. Environ Health.
2009;8:37.
Nohr EA, Frydenberg M, Henriksen TB, Olsen J. Does low participation
in cohort studies induce bias? Epidemiology. 2006;17:413 418.
Olesen C, Sondergaard C, Thrane N, Nielsen GL, de Jong-van den Berg
L, Olsen J. Do pregnant women report use of dispensed medications?
Epidemiology. 2001;12:497501.
Reddy GP, Prasad M, Sailesh S, Kumar YV, Reddanna P. Arachidonic
acid metabolites as intratesticular factors controlling androgen production. Int J Androl. 1993;16:227233.
Wang X, Dyson MT, Jo Y, Stocco DM. Inhibition of cyclooxygenase-2
activity enhances steroidogenesis and steroidogenic acute regulatory
gene expression in MA-10 mouse Leydig cells. Endocrinology. 2003;
144:3368 3375.

www.epidem.com | 785