Professional Documents
Culture Documents
Oral Contraceptive and Intrauterine Device Use and The Risk of Cervical Intraepithelial Neoplasia Grade III or Worse A Population Based Study
Oral Contraceptive and Intrauterine Device Use and The Risk of Cervical Intraepithelial Neoplasia Grade III or Worse A Population Based Study
ScienceDirect
Original Research
a
Department of Obstetrics and Gynaecology, Radboud Institute for Molecular Life Sciences, Radboud university medical
center, PO Box 9101, 6500HB, Nijmegen, the Netherlands
b
Department of Biostatistics, Radboud Institute for Health Sciences, Radboud university medical center, PO Box 9101,
6585KM, Nijmegen, the Netherlands
c
Department of Medical Microbiology, Radboud university medical center, PO Box 9101, 6500HB, Nijmegen, the
Netherlands
d
Department of Obstetrics and Gynaecology, Radboud university medical center, PO Box 9101, 6500HB, Nijmegen, the
Netherlands
e
Department of Obstetrics and Gynaecology, Catharina Hospital, PO Box 1350, 5602ZA, Eindhoven, the Netherlands
f
Department of Pathology, Radboud university medical center, PO Box 9101, 6500HB, Nijmegen, the Netherlands
g
PALGA, Randhoeve 225a, 3995 GA, Houten, the Netherlands
KEYWORDS Abstract Objective: Hormonal contraceptive use has been associated with the development
Cervical intraepithelial of cervical cancer, although inconsistent results are reported on the association with intrauter-
neoplasia; ine device (IUD) use. The aim of this study was to evaluate the association between the type of
Contraceptives; contraceptive use and the development of cervical intraepithelial neoplasia grade III or worse
Abbreviations: ASC-H, atypical squamous cells e cannot rule out high-grade squamous intraepithelial lesion; ASC-US, atypical squamous cells of
undetermined significance; CIN, cervical intraepithelial neoplasia; hrHPV, high-risk human papillomavirus; HSIL, high-grade squamous intra-
epithelial lesion; IUD, intrauterine device; LSIL, low-grade squamous intraepithelial lesion; NILM, negative for intraepithelial lesion or malig-
nancy; NOS, not otherwise specified; OC, oral contraceptive; PALGA, the nationwide network and registry of histo- and cytopathology in the
Netherlands.
* Corresponding author: Department of Obstetrics and Gynaecology, Radboud Institute for Molecular Life Sciences, Radboud university medical
center, PO Box 9101, 6500HB, Nijmegen, the Netherlands. Fax: þ31 243668597.
E-mail addresses: diede.loopik@radboudumc.nl (D.L. Loopik), Joanna.intHout@radboudumc.nl (J. IntHout), willem.melchers@radboudumc.
nl (W.J.G. Melchers), leon.massuger@radboudumc.nl (L.F.A.G. Massuger), ruud.bekkers@radboudumc.nl (R.L.M. Bekkers), bert.siebers@
radboudumc.nl (A.G. Siebers).
1
Present address: GROW, School for Oncology & Developmental Biology, Maastricht University Medical Centre, PO Box 616, 6200MD,
Maastricht, the Netherlands.
https://doi.org/10.1016/j.ejca.2019.10.009
0959-8049/ª 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).
D.L. Loopik et al. / European Journal of Cancer 124 (2020) 102e109 103
(CIN3þ).
Follow-up studies;
Methods: A retrospective population-based cohort study including women aged 29e44 years
Intrauterine devices;
attending the cervical cancer screening program with normal cytology between 2005 and 2009
Risk assessment;
identified from the Dutch Pathology Registry. Subgroups with at least 5 years registered use of
Retrospective studies;
an oral contraceptive (OC) or IUD were compared with non-users. Risk ratios of CIN3þ were
Uterine cervical
estimated per contraceptive type.
neoplasms
Results: 702,037 women were included with a median follow-up of 9.7 years, of which 6705
(0.96%) and 559 (0.08%) women developed CIN3 and cervical cancer, respectively. IUD use
was associated with an increased risk of developing CIN3þ (risk ratio (RR) 1.51, 95% confi-
dence interval (CI) 1.32e1.74), and OC use was associated with an increased risk of devel-
oping CIN3þ (RR 2.77, 95%CI 2.65e3.00) and cervical cancer (RR 2.06, 95%CI 1.52
e2.79). The risk of developing CIN3þ and cervical cancer was higher for OC users compared
with IUD users (RR 1.83, 95%CI 1.60e2.09 and RR 1.70, 95%CI 1.00e2.90, respectively).
Conclusions: Both OC use and IUD use were associated with an increased risk of developing
CIN3þ. However, for women with a contraceptive wish, an IUD seems safer than an OC as
the risk of developing CIN3þ and cervical cancer was higher for OC users.
ª 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC
BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Table 1
Baseline characteristics and highest cervical abnormality during follow-up per type of contraceptive.
Parameters No contraceptive usej Oral contraceptive usek Intrauterine device usel Total included womenm
for 5 years n Z 178,545 for 5 years n Z 85,823 for 5 years n Z 27,509 N Z 702,037 (100%)
(25.4%) (12.2%) (3.9%)
Age, years
Median (range) 39 (29e44) 35 (29e44) 35 (29e44) 35 (29e44)
Age group, years, No. (%)
29 9213 (5.2) 9222 (10.8) 1708 (6.2) 54,498 (7.8)
30e34 47,572 (26.6) 31,519 (36.7) 9362 (34.0) 230,447 (32.8)
35e39 52,064 (29.2) 24,265 (28.3) 8860 (32.2) 196,419 (28.0)
40e44 69,696 (39.0) 20,817 (24.3) 7579 (27.6) 220,673 (31.4)
Follow-up, years
Median (range) 9.6 (5.0e13.1) 9.7 (5.0e13.2) 9.4 (5.0e13.1) 9.7 (5.0e13.2)
Cytologya, No. (%)
Normal b 162,624 (91.1) 76,405 (89.0) 24,822 (90.2) 618,145 (88.1)
Low-gradec 8451 (4.7) 4884 (5.7) 1438 (5.2) 38,375 (5.5)
High-graded 196 (0.1) 91 (0.1) 29 (0.1) 728 (0.1)
Unknowne 184 (0.1) 90 (0.1) 15 (0.1) 531 (0.1)
Total 171,455 (96.0) 81,470 (94.9) 26,304 (95.6) 657,779 (93.7)
Histologyf, No. (%)
Normal 4042 (2.3) 1220 (1.4) 410 (1.5) 23,479 (3.3)
Low-gradeg 1210 (0.7) 858 (1.0) 292 (1.1) 7865 (1.1)
High-gradeh 1823 (1.0) 2270 (2.6) 502 (1.8) 12,811 (1.8)
Unknowni 15 (0.0) 5 (0.0) 1 (0.0) 103 (0.0)
Total 7090 (4.0) 4353 (5.1) 1205 (4.4) 44,258 (6.3)
ASC-H: atypical squamous cellsdcannot rule out high-grade squamous intraepithelial lesion; ASC-US: atypical squamous cells of undetermined
significance; CIN: cervical intraepithelial neoplasia; hrHPV: high-risk human papillomavirus; HSIL: high-grade squamous intraepithelial lesion;
LSIL: low-grade squamous intraepithelial lesion; NILM: negative for intraepithelial lesion or malignancy; NOS: not otherwise specified.
a
Highest cytologic result during follow-up in case of absent histologic results.
b
Normal: including NILM and negative hrHPV result.
c
Low-grade: including hrHPV positive result, ASC-US and LSIL.
d
High-grade: including ASC-H and HSIL.
e
Unknown: including inadequate cytology.
f
Highest histologic result during follow-up.
g
Low-grade: including CINeNOS and CIN1.
h
High-grade: including CIN2, CIN3 and cervical cancer.
i
Unknown: including missing data.
j
Women who did not use any type of contraceptive for at least five uninterrupted years during follow-up.
k
Women who used an oral contraceptive for at least five uninterrupted years during follow-up.
l
Women who had an intrauterine device for at least five uninterrupted years during follow-up.
m
All women included in the study, independent from contraceptive use.
cumulative incidence of CIN2 and CIN3. The cu- 1.00e2.90, respectively). Women who did not use any
mulative incidence of CIN2 was similar (0.94% and type of contraceptive had a significantly lower risk of
0.92%, respectively; p0.41), whereas the cumulative developing CIN3þ (RR 0.61, 95%CI 0.58e0.64) and
incidence of CIN3 was significantly higher in the OC cervical cancer (RR 0.63, 95%CI 0.53e0.75) compared
group (0.83% and 1.62%, respectively; p < 0.01). with the total group of included women. When strati-
The RR of CIN3þ stratified by age and contraceptive fying cervical cancer by histological type, 347 women
type is shown in Fig. 3. Older age in the IUD group was (62.1%) had a squamous cell carcinoma, 183 women
associated with a higher risk of developing CIN3þ, (32.7%) an adenocarcinoma and 29 women (5.2%) had
therefore the age-adjusted RR per contraceptive type another type of cervical cancer.
has been estimated (Fig. 4). The use of an IUD was
associated with an increased risk of developing CIN3þ
(RR 1.51, 95%CI 1.32e1.74) but did not significantly 4. Discussion
influence the risk of cervical cancer (RR 1.21, 95%CI
0.71e2.07). The use of an OC was associated with an This study, including 702,037 screened women with a
increased risk of developing CIN3þ (RR 2.77, 95%CI lowerisk profile for the development of cervical ab-
2.65e3.00) and cervical cancer (RR 2.06, 95%CI normalities, confirms that OC use is associated with an
1.52e2.79). The risk of developing CIN3þ and cervical increased risk of developing CIN3þ, but in contrast to
cancer was higher for OC users compared with IUD previous studies, we also found this increased risk in
users (RR 1.83, 95%CI 1.60e2.09 and RR 1.70, 95%CI IUD users compared with non-users. The risk of
106 D.L. Loopik et al. / European Journal of Cancer 124 (2020) 102e109
Fig. 3. Risk ratio and 95%CI of CIN3þ stratified by age. CI: confidence interval; CIN: cervical intraepithelial neoplasia; RR: risk ratio.
D.L. Loopik et al. / European Journal of Cancer 124 (2020) 102e109 107
Fig. 4. Risk ratio and 95%CI of CIN3þ and cervical cancer adjusted for age. CI: confidence interval; CIN: cervical intraepithelial
neoplasia; RR: risk ratio.
A possible mechanism that could explain the poten- not to use a type of contraceptive could be a reflection of
tial protective effect of an IUD is through a device- adopted sexual behaviour. Unfortunately, we could not
related inflammatory response in the endocervical canal adjust for sexual behaviour, hrHPV status or other po-
[8]. Copper-IUDs release ions which increase prosta- tential confounders, such as smoking status, as this was
glandin levels causing chronic inflammation, while a retrospective quantitative study. Therefore, contra-
levonorgestrel-releasing IUDs decrease prostaglandin ceptive use could be a surrogate marker for a certain
levels suppressing the local immunity [15]. In contrast to sexual behaviour pattern.
combined hormonal contraceptive use, progesterone In our study, the risk of developing CIN3þ was
only may have a protective effect by increasing the higher for OC users compared with IUD users. A similar
number of Langerhans cells, which are important for sexual behaviour pattern has been found in women with
immunosurveillance [16]. However, progesterone has OC use and the use of another type of contraceptive [14].
also been shown to increase E6/E7 oncogene transcrip- The main difference between the IUD and OC group
tion in cell lines with integrated HPV-16 [17]. was found between the cumulative incidence of CIN2
A meta-analysis of 28 studies found a two-fold risk of and CIN3. The cumulative incidence of CIN2 was
developing cervical cancer in women with long duration similar, whereas the cumulative incidence of CIN3 was
of hormonal contraceptives, also after adjustment for significantly higher in the OC group (Fig. 2). These data
HPV status, smoking status and sexual behaviour [6]. may show that IUD use is more associated with pro-
Another study with 17,032 women found a persistent ductive and transient hrHPV infections and OC use with
increased risk for cervical cancer for many years after transforming and integrated hrHPV infections. Com-
cessation of use [18]. However, a study about potential bined hormonal contraceptives may stimulate the inte-
confounding effects concluded that sexual behaviour is gration of hrHPV-DNA into the host genome, which
different among OC users and non-users, and they could enhances the expression of the E6 and E7 HPV onco-
not confirm that the use of OCs was an independent risk proteins. These oncoproteins stimulate the degradation
factor for developing CIN [14]. The decision to use or of P53 tumour suppressor genes and enhance the ability
108 D.L. Loopik et al. / European Journal of Cancer 124 (2020) 102e109
of the viral DNA to transform cells and induce carci- out. This manuscript will not be published elsewhere in
nogenesis [16,19]. This could explain the similar preva- the same form.
lence of hrHPV between the different subgroups, as a
DNA hrHPV test does not differ between integrated and Role of the funding source
not integrated hrHPV.
HrHPV positive women with a contraceptive wish This research did not receive any specific grant from
should be counseled about the possible positive and funding agencies in the public, commercial or not-for-
negative side-effects of the different contraception op- profit sectors.
tions and may be counseled against long-term use of an
OC and may benefit from an IUD. On the other hand,
Ethics approval
OCs may have beneficial effects on other type of can-
cers, such as ovarian, endometrial and colorectal cancer
All the authors report adherence to ethical standards in
[20,21]. These known health benefits should be weighed
the conception of the work, data collection and writing
against the increased risk of developing cervical cancer.
of the manuscript. The study was approved by the sci-
Nonetheless, all women should be encouraged to
entific committee of the Dutch Pathology Registry
participate in the cervical screening program and further
(PALGA). The study was exempt from institutional
research about the benefits of intensified screening in
review board approval because data were gathered
women with long-term use of an OC is warranted.
retrospectively and analysed anonymously.
Beside the lack of information about other behav-
ioural factors, the selection of subgroups with at least 5
years of reported use of one type of contraception is not Conflict of interest statement
a hundred percent certain as most women participate in
the screening program every 5 years, and women may None declared.
have changed their use of contraceptive type tempo-
rarily. However, the differences between the subgroups
were statistically significant, regardless of potential Acknowledgements
misreporting.
The strengths of our study includes its population- None.
based design with a large sample size of women with a
lowerisk profile for the development of cervical ab- References
normalities with a median follow-up of 9.7 years, the
differentiation between the high-grade outcomes: CIN2, [1] Bosch FX, Lorincz A, Munoz N, Meijer CJ, Shah KV. The causal
CIN3 and cervical cancer and the comparison of women relation between human papillomavirus and cervical cancer. J
with OC use and IUD use to non-users. Clin Pathol 2002;55(4):244e65.
[2] Einstein MH, Schiller JT, Viscidi RP, Strickler HD, Coursaget P,
Studies focusing on behavioural patterns (including Tan T, et al. Clinician’s guide to human papillomavirus immu-
contraceptive use), the incidence and persistence of hrHPV nology: knowns and unknowns. Lancet Infect Dis 2009;9(6):
infection and progression of CIN are needed. Also studies 347e56. https://doi.org/10.1016/s1473-3099(09)70108-2.
exploring the biological role of OCs and IUDs on acqui- [3] Pett M, Coleman N. Integration of high-risk human papilloma-
sition, persistence and integration of hrHPV infections are virus: a key event in cervical carcinogenesis? J Pathol 2007;212(4):
356e67. https://doi.org/10.1002/path.2192.
important to fully understand the association. [4] Castellsague X, Bosch FX, Munoz N. Environmental co-factors
in HPV carcinogenesis. Virus Res 2002;89(2):191e9.
[5] Munoz N, Castellsague X, de Gonzalez AB, Gissmann L. Chapter
5. Conclusions 1: HPV in the etiology of human cancer. Vaccine 2006;24(Suppl
3). https://doi.org/10.1016/j.vaccine.2006.05.115. S3/1-10.
Both OC use and IUD use were associated with an [6] Smith JS, Green J, Berrington de Gonzalez A, Appleby P, Peto J,
increased risk of developing CIN3þ compared with Plummer M, et al. Cervical cancer and use of hormonal contra-
non-users. However, for women with a contraceptive ceptives: a systematic review. Lancet (London, Engl) 2003;
361(9364):1159e67.
wish, an IUD seems safer than an OC as the risk of [7] Averbach SH, Ma Y, Smith-McCune K, Shiboski S,
developing CIN3þ and cervical cancer was higher for Moscicki AB. The effect of intrauterine devices on acquisition and
OC users. clearance of human papillomavirus. Am J Obstet Gynecol 2017;
216(4):386. https://doi.org/10.1016/j.ajog.2016.11.1053. e1 - e5.
[8] Castellsague X, Diaz M, Vaccarella S, de Sanjose S, Munoz N,
Submission declaration and verification Herrero R, et al. Intrauterine device use, cervical infection with
human papillomavirus, and risk of cervical cancer: a pooled
This manuscript has not been published previously and analysis of 26 epidemiological studies. Lancet Oncol 2011;12(11):
1023e31. https://doi.org/10.1016/s1470-2045(11)70223-6.
is not under consideration for publication elsewhere, [9] Roura E, Travier N, Waterboer T, de Sanjose S, Bosch FX,
and this publication is approved by all authors and the Pawlita M, et al. The influence of hormonal factors on the risk of
responsible authorities where the work has been carried developing cervical cancer and pre-cancer: results from the EPIC
D.L. Loopik et al. / European Journal of Cancer 124 (2020) 102e109 109