Jensen et al
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
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
(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
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.
www.epidem.com | 781
Jensen et al
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
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
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
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
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
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
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
www.epidem.com | 785