Keywords
hormonal contraception - human papillomavirus - HPV - ethinylestradiol
Palavras-chave
contracepção hormonal - vírus do papiloma humano - HPV - etinilestradiol
Introduction
Human papillomavirus (HPV) infection is the most common sexually transmitted disease
(STD), affecting ∼ 50% of the world's population.[1] It is estimated that between 75 and 80% of sexually-active individuals will acquire
some subtype of HPV throughout life.[2] In Brazil, the prevalence rate of HPV varies from 13.7 to 54.3%, according to the
population and region studied.[3]
[4]
Most genital infections are asymptomatic, but clinical forms are usually associated
with low-risk oncogenic HPV and tend to be benign, whereas subclinical forms may include
benign and/or malignant lesions and are usually caused by high-risk oncogenic HPV.[5]
Among the factors associated with the increase of HPV infection are the number of
sexual partners, STD, multiparity, age of onset of sexual activity[6]
[7] and smoking.[2]
[8]
[9]
[10]
[11] There is no consensus in the literature on the association of hormonal contraceptives
with the prevalence and/or persistence of cervical lesions induced by HPV. Numerous
hypotheses attempt to justify the connection between the use of hormonal contraceptives
and these aspects, such as the possibility of exogenous steroids acting on the HPV
genome, causing mutations and the onset of cervical cancer, and the fact that progesterone
increases the transcription of certain types of HPV, including HPV-16, through mediation
by glucocorticoid-responsive elements that regulate virus transcription.[12] Furthermore, immune responses in the female genital tract are regulated by endogenous
and exogenous sex hormones, and antigen presentation, cytokine production, immunoglobulin
production and transport, and induction of tolerance have all been shown to be influenced
by variations in the levels of sex hormones.[13] Users of combined oral contraceptives (COCs) have a decrease in immunoglobulin A
(IgA) and immunoglobulin G (IgG) levels during the pause period in the cyclic schemes,
thus providing a favorable environment for the appearance of HPV lesions.[11]
[14]
[15]
[16]
[17]
Due to the many divergences found in the literature, it is extremely important to
try to clarify if there is a relationship between these factors, in order to enable
physicians to provide better guidance and information to the users of this class of
drugs so they may choose the best contraceptive option. Thus, this study aims to evaluate
the association between hormonal contraception and the appearance of HPV-induced lesions
in the uterine cervix of patients assisted at a school outpatient clinic at the Gynecology
and Obstetrics Service of the Medical Ambulatory of Specialties of Universidade do
Sul de Santa Catarina (AME,/USNISUL in the Portuguese acronym).
Methods
This study was based on the ethical principles of Resolution 466/12 of the Brazilian
National Health Council and the Code of Ethics of the Declaration of Helsinki, and
it was approved by the Research Ethics Committee of Universidade do Sul de Santa Catarina
(UNISUL, in the Portuguese acronym), under the CAAE no. 17596313.9.0000.5369.
A case-control study was performed with women in fertile age at the Gynecology and
Obstetrics Service of the Medical Ambulatory of Specialties (AME, in the Portuguese
acronym) of UNISUL, located in the municipality of Palhoça, state of Santa Catarina,
in the period between 2012 and 2015.
Patients were selected according to the International Statistical Classification of
Diseases and Related Health Problems, 10th Revision (ICD-10) on an electronic file.
For the case group, the ICD-10 was used for cervical lesions (N87) and, for the control
group, it was used for a gynecological revision (Z01.4). The choice of medical records
was made using a systematic technique that will be explained subsequently.
In the case group, patients aged 18 to 45 years, who were submitted to cervical biopsy
for alterations suggestive of HPV lesion, confirmed by anatomopathological examination,
were included in the study. These patients were classified, according to the Bethesda
classification, as having low-grade squamous intraepithelial lesion (LSIL), high-grade
squamous intraepithelial lesion (HSIL) and invasive cancer.[18] In the control group, patients within the same age range as the case group, who
underwent routine gynecological evaluation, and whose oncotic colpocytology was normal,
were included. In both groups, the medical records of immunocompromised patients,
smokers or patients whose diagnosis of injury occurred during pregnancy, and the medical
records that did not contain all the necessary information for the study, were excluded.
The sample size was calculated in the OpenEpi 2.3.1 software (Open Source Epidemiologic
Statistics for Public Health, Atlanta, Georgia, EUA) using the formula for case-control
studies with the following parameters: 95% confidence interval (95%CI), 80% test power,
60% exposed proportion between cases, 1:1 ratio of controls for cases and odds ratio
(OR) of 2.2 for a unicaudal test. The last two parameters were used according to the
article by Ajah et al, who investigated a similar outcome in a multicenter study.[19] The procedure resulted in a final sample of 104 patients per group.
The extracted data was recorded in a data collection instrument specially developed
for the present study. Afterwards, it was inserted in an electronic database of the
Microsoft Excel (Microsoft, Redmond, WA, US) software, and exported to the Statistical
Package for the Social Sciences (SPSS, IBM Corp. Armonk, NY, US) software, version
24.0, in which it was analyzed.
Qualitative variables were described by absolute and relative frequencies, while quantitative
variables were described as medians, means and standard deviations for a subsequent
bivariate analysis. To test the homogeneity of the proportions, the chi-square (χ2)
test was used for the qualitative variables, and the Student t-test, for the quantitative
variables.
Results
A total of 202 medical records were included and divided into 101 cases and 101 controls.
The average age of the patients with HPV cervical lesion was of 29.7 ± 8.8 years,
while, among those without lesions, it was of 32.7 ± 10.5 years. This result (p = 0.687), which is similar to what was observed for the other sociodemographic characteristics
evaluated in the study (ethnicity, marital status and level of schooling), did not
differ statistically between the two groups.
Regarding the prevalence of hormonal contraceptive use in the study population, 101
patients (50%) did not use it, while 101 patients (50%) did. The proportion of use
between groups was of 55.4% and 44.6% respectively, in the case and control groups
(p = 0.157). The most commonly used route of administration for hormonal contraceptives
in the studied population was the oral route (90.1%), while the route of administration
of 9.9% of the women was intramuscular (p = 0.283). There was no report of contraceptive use by the transdermal, vaginal, subcutaneous
or intrauterine routes.
The average time of use of hormonal contraceptives was 5 years. There was no association
between this time of use and the occurrence of cervical lesion by HPV, even when other
time stratifications were evaluated. When analyzing the characteristics of time of
use, route of administration ([Table 1]) and formulation of hormonal contraceptives ([Table 2]), compared to non-users, there was no difference between the two groups studied.
Table 1
Association between contraceptive route of administration and HPV lesions compared
to non-users in the study patients
|
Route of Administration
|
Case
n (%)
|
Control
n (%)
|
Total
n (%)
|
p-value
|
OR (95%CI)
|
|
Non-user
|
45 (44.5)
|
56 (55.4)
|
101 (50.0)
|
–
|
1
|
|
User
|
|
|
|
|
|
|
Oral
|
51 (50.5)
|
40 (86.9)
|
91 (45.0)
|
0.283
|
0.6443 (0.3648–1.1378)
|
|
Intramuscular
|
5 (5.0)
|
5 (5.0)
|
10 (5.0)
|
0.715
|
0.8214 (0.2240–3.0126)
|
Abbreviations: 95%CI, 95% confidence interval; OR, odds ratio.
Table 2
Association between contraceptive composition and HPV lesions compared to non-users
in the study patients
|
Composition
|
Case
n (%)
|
Control
n (%)
|
Total
n (%)
|
p-value
|
OR (95%CI)
|
|
Non-user
|
45 (44.5)
|
56 (55.4)
|
101 (50.0)
|
–
|
1
|
|
User
|
|
|
|
|
|
|
EE + associations
|
48 (47.6)
|
39 (39.6)
|
86 (43.1)
|
0.147
|
1.5316 (0.8604–2.7263)
|
|
E2 + associations
|
6 (5.9)
|
4 (4)
|
10 (4.9)
|
0.355
|
1.8667 (0.4964–7.0201)
|
|
DMPA
|
1 (1.0)
|
–
|
1 (0.5)
|
0.615
|
0.2743 (0.0109–6.8944)
|
|
DSG
|
1 (1.0)
|
2 (2.0)
|
3 (1.5)
|
0.452
|
2.4643 (0.2479–24.4951)
|
Abbreviations: 95%CI, 95% confidence interval; DSG, desogestrel; DMPA, depot medroxyprogesterone
acetate; EE, ethynilestradiol; E2, estradiol; OR, odds ratio.
When considering the types of progesterone used and the degree of HPV lesions, 70
(69.3%) women in the case group and 45 (44.6%) patients in the control group used
some type of progesterone, but there was no difference between users and non-users,
regardless of the type of progesterone ([Table 3]).
Table 3
Association between the type of progesterone present in hormonal contraceptives and
the degree of HPV lesions compared to non-users in the study patients
|
Progesterone
|
Case
n (%)
|
Control
n (%)
|
Total
n (%)
|
p-value
|
OR (95%CI)
|
|
LSIL (n = 70/98)
|
|
Non-user
|
31 (31.6)
|
56 (57.1)
|
87 (51.8)
|
–
|
1
|
|
User
|
|
|
|
|
|
|
CPA
|
6 (8.6)
|
7 (7.1)
|
15 (8.9)
|
0.466
|
1.5484 (0.4780–5.0160)
|
|
DSG
|
22 (31.4)
|
19 (19.4)
|
41 (24.4)
|
0.055
|
2.0917 (0.9836–4.4482)
|
|
DHPA
|
2 (2.9)
|
1 (1.0)
|
3 (1.8)
|
0.302
|
3.6129 (0.3148–41.4622)
|
|
DMPA
|
1 (1.4)
|
–
|
1 (0.6)
|
0.307
|
5.3810 (0.2128–136.0563)
|
|
DRSP
|
3 (4.3)
|
1 (1.0)
|
4 (2.4)
|
0.150
|
5.4194 (0.5404–54.3438)
|
|
GST
|
3 (4.3)
|
10 (10.2)
|
13 (7.7)
|
0.378
|
0.5419 (0.1387–2.1174)
|
|
LNG
|
1 (1.4)
|
4 (4.1)
|
5 (3.0)
|
0.485
|
0.4516 (0.0483–4.2203)
|
|
NET-EN
|
1 (1.4)
|
–
|
1 (0.6)
|
0.307
|
5.3810 (0.2128–136.0563)
|
|
HSIL (n = 28/98)
|
|
Non-user
|
12 (42.8)
|
56 (57.1)
|
68 (54.0)
|
–
|
1
|
|
User
|
|
|
|
|
|
|
CPA
|
6 (21.4)
|
7 (7.1)
|
15 (11.9)
|
0.070
|
3.3333 (0.9029–12.3056)
|
|
DSG
|
4 (14.3)
|
19 (19.4)
|
23 (18.2)
|
0.977
|
0.9825 (0.2827–3.4138)
|
|
DHPA
|
–
|
1 (1.0)
|
1 (0.8)
|
0.805
|
1.5067 (0.0579–39.1998)
|
|
DRSP
|
1 (3.6)
|
1 (1.0)
|
2 (1.6)
|
0.287
|
4.6667 (0.2723–79.9627)
|
|
GST
|
1 (3.6)
|
10 (10.2)
|
11 (8.7)
|
0.486
|
0.4667 (0.0545–3.9988)
|
|
LNG
|
1 (3.6)
|
4 (4.1)
|
5 (4.0)
|
0.894
|
1.1667 (0.1195–11.3870)
|
|
NET-EN
|
3 (10.7)
|
–
|
3 (2.4)
|
0.776
|
0.6457 (0.0313–13.3099)
|
|
Invasive cancer (n = 3/98)
|
|
Non-user
|
2 (66.7)
|
56 (57.1)
|
58 (57.4)
|
–
|
1
|
|
User
|
|
|
|
|
|
|
CPA
|
–
|
7 (7.1)
|
7 (6.9)
|
0.797
|
1.5067 (0.0658–34.4808)
|
|
DSG
|
1 (33.3)
|
19 (19.4)
|
20 (19.8)
|
0.757
|
1.4737 (0.1264–17.1847)
|
|
DHPA
|
–
|
1 (1.0)
|
1 (1.0)
|
0.250
|
7.5333 (0.2410–235.4646)
|
|
DRSP
|
–
|
1 (1.0)
|
1 (1.0)
|
0.250
|
7.5333 (0.2410–235.4646)
|
|
GST
|
–
|
10 (10.2)
|
10 (9.9)
|
0.963
|
1.0762 (0.0481–24.0580)
|
|
LNG
|
–
|
4 (4.1)
|
4 (4.0)
|
0.570
|
2.5111 (0.1039–60.6610)
|
Abbreviations: 95%CI, 95% confidence interval; CPA, cyproterone acetate; DHPA, dihydroxyprogesterone
acetate; DMPA, depot medroxyprogesterone acetate; DRSP, drospirenone; DSG, desogestrel;
GST, gestodene; HSIL, high-grade squamous intraepithelial lesion; LNG, levonorgestrel;
LSIL, low-grade squamous intraepithelial lesion; NET-EN, norethisterone enantate;
OR, odds ratio.
When evaluating the dose of ethinylestradiol (EE) present in the COCs, users of 0.03 mg
EE had a 1.9-fold increased risk of developing cervical lesions induced by HPV when
compared with non-users of contraceptives (p = 0.039) ([Table 4]); in these cases, the risk of developing LSIL was 2.1 times higher, but with no
impact on HSIL and on the occurrence of invasive cancer ([Table 5]).
Table 4
Association between doses of ethinylestradiol present in oral combined hormonal contraceptives
and HPV lesions compared to non-users in the study patients
|
EE Dose (mg)
|
Case n = 93
|
Control n = 95
|
Total n = 188
|
p-value
|
OR (95%CI)
|
|
Non-user
|
45 (48.0%)
|
56 (58.9%)
|
101 (53.7%)
|
|
1
|
|
User
|
|
|
|
|
|
|
≥ 0.03 mg
|
38 (40.8%)
|
24 (25.3%)
|
62 (33.0%)
|
0.039
|
1.9704 (1.0345–3.7529)
|
|
≤ 0.02 mg
|
10 (10.7%)
|
15 (15.8%)
|
25 (13.3%)
|
0.681
|
0.8296 (0.4829–2.0265)
|
Abbreviation: 95%CI, 95% confidence interval; EE, ethinylestradiol; OR, odds ratio.
Table 5
Association between doses of ethinylestradiol present in oral combined hormonal contraceptives
and the degree of HPV lesions compared to non-users in the study patients
|
EE Dose (mg)
|
Case n (%)
|
Control n (%)
|
Total n (%)
|
p-value
|
OR (95%CI)
|
|
LSIL (n = 70/95)
|
|
Non-user
|
31 (44.3)
|
56 (58.9)
|
87 (52.7)
|
–
|
1
|
|
User
|
|
|
|
|
|
|
≥ 0.03
|
28 (40.0)
|
24 (25.3)
|
52 (31.5)
|
0.036
|
2.1075 (1.0467–4.2434)
|
|
≤ 0.02
|
11(15.7)
|
15 (15.8)
|
26 (15.3)
|
0.537
|
1.3247 (0.5423–3.2363)
|
|
HSIL (n = 28/95)
|
|
Non-user
|
12 (42.9)
|
56 (58.9)
|
68 (55.3)
|
–
|
1
|
|
User
|
|
|
|
|
|
|
≥ 0.03
|
10 (35.7)
|
24 (25.3)
|
34 (27.6)
|
0.177
|
1.9444 (0.7401–5.1083)
|
|
≤ 0.02
|
6 (21.4)
|
15 (15.8)
|
21 (17.1)
|
0.280
|
1.8667 (0.6008–5.7995)
|
|
Invasive cancer (n = 3/95)
|
|
Non-user
|
2 (66.7)
|
56 (58.9)
|
58 (59.2)
|
–
|
1
|
|
User
|
–
|
–
|
–
|
–
|
–
|
|
≥ 0.03
|
–
|
24 (25.3)
|
24 (24.5)
|
0.621
|
0.46120 (0.0213–9.9676)
|
|
≤ 0.02
|
1 (33.4)
|
15 (15.8)
|
16 (13.3)
|
0.620
|
1.8667 (0.1583–22.0070)
|
Abbreviations: 95%CI, 95% confidence interval; EE, ethinylestradiol; HSIL, high-grade
squamous intraepithelial lesion; LSIL, low-grade squamous intraepithelial lesion;
OR, odds ratio.
Discussion
The present study demonstrated, for the first time in the Brazilian population, the
association between the use of oral hormonal contraceptives based on EE at the dose
of 0.03 mg and the appearance of HPV-induced LSIL in the uterine cervix.
When analyzing the sociodemographic variables of the patients in the present study,
no characteristic was considered significant either for the development or non-development
of the lesions. These findings were similar to those described in 2016 by Abouzeid
and El-Agwany,[10] who recorded, in a case-control study with 250 women, that there was no significant
statistical difference in age, marital status or parity.
Other studies evaluated the association of cervical lesions with sociodemographic
variables not investigated in the present study. Among the women younger than 35 years
of age, women who lived in rural areas, and women with no fixed partners, women with
a lower educational level and multiparous women, a greater probability of developing
cervical neoplasm was observed.[20] Another study aimed at determining the demographic and behavioral factors associated
with HPV positivity, a prevalence of 25.3% of HPV lesions was identified in patients
aged between 31 and 35 years.[21] In 2010, when investigating the short-term persistence of HPV infection among 2,408
women with low-grade cervical lesions or other cytological abnormalities, Maucort
et al found a greater probability of HPV lesion persistence in white women aged between
20 and 29-years.[5]
In addition, there was no difference between the groups when considering the characteristics
of the patients observed in the present study, regarding the prevalence of use, route
of administration and time of use of hormonal contraceptives. Finally, no association
was found between these characteristics of the use of hormonal contraception and the
presence of cervical lesions, as well as no significant difference between the case
and control groups. These findings are similar to those of Westreich et al[21] in 2014, who analyzed the impact of the use of depot medroxyprogesterone (DMPA),
norethisterone enanthate (NET-EN) and COCs separately, on the incidence and progression
of cervical lesions, without finding significant differences. Similarly, Binesh et
al,[22] in 2013, in a cross-sectional study, found no association between COC consumption
and changes in cervical cytology, agreeing with the findings of Sammarco et al,[8] in 2013, when studying the persistence and clearance of HPV in users of COCs, after
evaluating and controlling possible confounding factors.
In contrast, in 2011, Marks et al[23] identified an association between the appearance of new HPV lesions in women using
COCs, while Mitchell et al,[20] in 2014, and Jensen et al,[24] in 2013, demonstrated that the use of any hormonal contraceptive increases the probability
of the persistence of HPV carcinogenic viruses when compared with non-users, and Maucourt-Bouch
et al, in 2010, observed a slight increase in the risk of persistent lesions in injectable
contraceptive users.[5]
Regarding the time of use of contraceptives, there was no statistical significance
in the present study, a finding similar to those of the studies by Westreich et al,[22] in 2014, and Green et al,[25] in 2003. Watson-Jones et al,[6] in 2013, suggested that the use of hormonal contraceptives, both oral and intramuscular,
for less than four years, would serve as a protective factor against HPV lesions when
compared to condom use.
On the other hand, Roura et al, in 2016, identified a strong association between the
time of COC use and the risk of developing HSIL and invasive cancer.[9] Marks et al,[2] in 2011, identified that current COC use for more than 6 years is associated with
an increased risk of developing persistent HPV infections; and Brinton,[26] in 1991, suggested that 10 years of COC use could present an increased risk of developing
cervical cancer, which are findings similar to those of other studies with variable
research designs that evaluated the long-term use of contraceptives.[19]
[27]
In the present study, no association was found between HPV lesions and the progesterone
types present in the contraceptives, either alone or associated with EE. Abouzeid
and El-Agwany[10] reached similar findings in 2016 in a case-control study including 200 users of
contraceptives containing progesterone alone and 50 non-hormonal contraceptive users,
data corroborated by Darwish et al[28] in 2004. Contradicting these findings, in 1990, Herrero et al[29] demonstrated that women receiving DMPA had a high risk of developing cervical cancer,
but this result was significant only in those patients who had used it for more than
five years, which was the median time of hormone use in the present study.
When comparing hormonal contraceptive types and associating them with the HPV lesions,
a statistical significance was not evidenced, but when comparing only the doses of
EE, separated in doses of 0.03 mg and 0.02 mg, there was a significant association.
Patients taking higher doses of estrogen are more likely to develop HPV-induced lesions,
especially LSIL. This fact could be justified by the stimulation of a cervical ectopy
secondary to higher concentrations of estrogen,[30] besides altering the immune system and inducing an increased concentration and activity
of pro-inflammatory cytokines,[31] facilitating the development of these lesions.
No studies have been found to systematically assess and compare the dose of EE present
in COCs and its association with HPV-induced lesions. In 2012, Aksoy et al,[32] when evaluating the effect of EE 0.30 mg + drospirenone 3 mg on the cervical mucus,
observed a statistically significant increase in mucoprotein 2 (MUC2) levels, suggesting
that this is related to the efficacy of COCs, and also speculating that the MUC2 increase
induced by the hormonal contraceptive may be the mechanism responsible for the cervical
carcinogenesis induced by this method, although they consider that large-scale longitudinal
studies are necessary to confirm these findings.[32] Mitrani-Rosenbaum et al[33] (1989) and Gadducci et al[34] (2011) have demonstrated that both estrogen and progesterone can affect cervical
cells by HPV mRNA transcription and by integrating it into the host DNA. In addition,
sex steroids could increase the expression of HPV E6 and E7 genes, leading to apoptosis
failure and promoting carcinogenesis. However, Webster et al[35] (2001) failed to demonstrate that estrogen or progesterone could interfere with
HPV cellular apoptosis, and Harris et al[36] (2009) found that recent use of concentrations of EE > 0.03 mg for more than 2 years
is not associated with high-grade cervical lesions.
The present study had some limitations, which were related but not restricted to the
convenience sampling, the number of cases raised, and the design adopted. The evaluation
of confounding factors was not possible because the study is retrospective and based
on data analysis of medical records. It is possible that the inconsistencies reported
in the association between hormonal contraception and cervical lesions caused by HPV
may be related in part to these confounding risk factors, including lifestyle, sexual
behavior and HPV oncogenic type, which are very difficult to control.
Conclusion
Although the results found in the present study suggest that the users of COCs with
concentrations of EE > 0.03 mg could develop LSIL, a cause-effect relationship could
not be determined due to the design of the study, and more studies are needed to conclude
causality. However, this finding seems even more relevant if we consider the median
time of 5 years of use of COCs with concentrations of EE > 0.03 mg, as well as the
average age of the users (27.8 ± 6.4 years), deserving professional attention regarding
the orientation and follow-up of these women. The mechanisms involved in the persistence
and incidence of HPV lesions are far from being clarified, and new studies are needed
to elucidate better approaches regarding the type of contraception, route of administration
and hormonal doses that are not associated with HPV-induced lesions.
