Cervical surgery for cervical intraepithelial neoplasia and prolonged time to conception of a live birth: a case–control study

Authors


Correspondence: Dr AF Saftlas, Department of Epidemiology, University of Iowa College of Public Health, 105 River Street, S427 CPHB, Iowa City, IA 52242, USA. Email audrey-saftlas@uiowa.edu

Abstract

Objective

To determine whether women with a history of surgery for cervical intraepithelial neoplasia (CIN) are at an increased risk of subfertility, measured as a time to pregnancy of more than 12 months.

Design

Case–control study.

Setting

Iowa Health in Pregnancy Study (IHIPS), a population-based case–control study of preterm and small-for-gestational-age (SGA) live birth outcomes (from May 2002 through June 2005) in the USA.

Sample

Women with an intended pregnancy and a history of either one prior cervical surgery (n = 152), colposcopy only (n = 151), or no prior cervical surgery or colposcopy (n = 1021).

Methods

Cervical treatment history, pregnancy intention, time to pregnancy, and other variables were self-reported by computer-assisted telephone interviews. Odds ratios were calculated using logistic regression to estimate the risk of prolonged time to pregnancy among women with a history of cervical surgery or colposcopy alone, compared with untreated women (control group).

Main outcome measure

Prolonged time to pregnancy (i.e. >1 year).

Results

Prolonged time to pregnancy was most prevalent among treated women (16.4%), compared with untreated women (8.4%) and women with colposcopy only (8.6%) (= 0.039). After adjusting for covariates, women with prior cervical surgery had a more than two-fold higher risk of prolonged time to pregnancy compared with untreated women (aOR 2.09, 95% CI 1.26–3.46). In contrast, women with a history of colposcopy only had a risk equivalent to that found among untreated women (aOR 1.02, 95% CI 0.56–1.89).

Conclusions

Women with a history of cervical treatment for CIN are at increased risk of subfertility, measured as a time to pregnancy of more than 12 months.

Introduction

When premalignant cervical lesions are found in reproductive-aged women, several surgical treatment modalities are available to remove these lesions, including conisation, cryosurgery, laser vaporisation, and loop electrosurgical excision procedure (LEEP). The goal of these procedures is to remove dysplastic tissue; however, varying quantities of healthy endocervical tissue are also removed. The endocervix contains a large number of mucus-producing glands and crypts (i.e. secretory cells) that facilitate sperm transport from the vagina to the uterus. Loss of these structures may lead to a reduction in cervical mucus production or alterations in the composition of cervical mucus, which could inhibit sperm penetration and result in subfertility or infertility.[1] Removal of healthy tissue may also lead to changes in cervical architecture, such as cervical stenosis, which can also lead to infertility.[2, 3]

Although previous publications have reported little to no effect of cervical surgery on fertility outcomes,[2-16] several of these were case series studies,[4, 6, 8, 14, 15] and only two examined time to pregnancy.[5, 13] Of these two studies, only the study by Bigrigg et al. had a comparison group of untreated women.[5] This study reported on 72 women who had a LEEP and 62 control women who could recall their time to pregnancy. There was no difference between the two groups on time to conception, but the study only had enough power to detect a three-fold difference in infertility rates.[5] Very little information regarding the effect of other modalities of cervical treatment on time to pregnancy is available.

As treatments for cervical intraepithelial neoplasia (CIN) typically occur in young, reproductive-aged women,[17] it is important to fully understand the potential fertility impact of cervical surgeries. With such limited data currently available, healthcare providers are unable to fully counsel their patients on the long-term outcomes of their treatment options. Understanding the impact of cervical surgeries for cervical dysplasia on future fertility can guide women and their providers in selecting the most appropriate procedure.

The objective of this study is to determine whether women with a history of excisional or ablative cervical surgery are at increased risk of subfertility, assessed as time to the conception of a live birth.

Methods

We examined the association of cervical surgery with subsequent time to pregnancy within the Iowa Health in Pregnancy Study (IHIPS), a population-based case–control study of preterm and small-for-gestational-age (SGA) birth outcomes among residents of four Iowa counties who delivered a live birth over the period, from May 2002 through June 2005.[18] All potential case and control subjects were identified and selected from the Iowa electronic birth certificate file, and were then sent an introductory letter inviting them to participate in the study (n = 7202). The letter was followed up with a phone call inviting them to be screened for eligibility. Women under the age of 18 years at delivery, women with twins and higher-order births, non-English speakers, and women with a diagnosis of pre-gestational type-I or type-II diabetes mellitus, systemic lupus, or chronic renal disease were excluded from the study. A total of 4250 (59%) potential subjects selected from live-birth records were reached by telephone; of these, 836 (19.7%) refused to participate and 548 (12.9%) were ineligible. All protocols and informed consent procedures were approved by the University of Iowa Institutional Review Board.

Eligible women who provided verbal consent responded to a 45-minute computer-assisted telephone interview (CATI; = 2709). Over 90% of participating women provided signed consent, allowing the research team to review their prenatal and hospital delivery records and to validate the preterm delivery and SGA outcomes (n = 2308). As excisional cervical surgery was an emerging risk factor for preterm delivery,[18] all women were asked a series of nine questions regarding their history of positive pap smears, colposcopy with punch biopsy, and excisional surgery for the treatment of CIN (Table S1). Because very few studies had evaluated the effect of cervical surgery on subsequent fertility, we also asked women who had a planned pregnancy to report on how many months they took to conceive upon the cessation of contraceptive use: ‘How many months of having sex without using birth control did it take for you to get pregnant?’

To determine the influence of a history of cervical surgery on time to pregnancy, we restricted the analysis to the subgroup of women who had completed the interview, had an intended pregnancy (and therefore could provide data on time to pregnancy), and for whom medical chart data had been abstracted (n = 1632). Excluded from the final analysis were women who: had received fertility treatment to conceive the index pregnancy (n = 216); had a history of two or more cervical surgeries (n = 24); could not recall the type of cervical treatment (= 11); or had missing data on time to pregnancy (n = 57). This left a total of 1324 women for analysis. Of these, 1021 (77.1%) women had never received any form of cervical treatment, 151 (11.4%) women had a history of colposcopy or punch biopsy only, and 152 (11.5%) had cervical surgery for CIN (46 cone biopsies, 45 LEEPs, 44 cryosurgeries, and 17 laser vaporisations). As a result of the small number of women receiving the individual treatment modalities, we analysed cone biopsy, LEEP, cryosurgery, and laser vaporization as a single treatment category.

Twelve months of intercourse with no contraception is defined as the lower reference limit for infertility by the World Health Organization, and by many countries.[19] For this reason, we defined subfertility as being present in women who required at least 12 months of sexual intercourse with no contraception in order to conceive.

Covariates

Information was also collected on a wide variety of demographic, reproductive, lifestyle, and other characteristics, including maternal age, maternal race (non-Hispanic white, non-Hispanic black, Hispanic, Asian/Pacific Islander and other), education (less than or equal to high school, some college, college graduate – bachelor degree, or college graduate – masters or professional degree), household income, prepregnancy body mass index (BMI < 24.9, 25–29.9, 30+), gravidity, and parity. Pregnancy intention was measured using a question from the Pregnancy Risk Assessment Monitoring System 2000–2003 Core Questionnaire: ‘Thinking back to just before you became pregnant, which of the following best describes how you felt about becoming pregnant?’.[20]

Analysis

To identify study factors associated with increased time to pregnancy, we compared the characteristics of participants stratified by time to pregnancy (≤12 months versus >12 months), and tested for statistical significance using chi-square tests for categorical variables and the Student's t-test for continuous variables.

To estimate the association of treatment for cervical surgery (colposcopy/punch biopsy only, cervical surgery, or untreated control group) with risk of prolonged time to pregnancy, we calculated crude and adjusted odds ratios using logistic regression analysis. Covariates that changed the odds ratios by at least 10% were retained in the final regression model. All data analysis was conducted using sas® 9.3 (SAS Institute Inc., Cary, NC, USA).

Results

A total of 9.4% (n = 124) of women with an intended pregnancy required more than 12 months to conceive. The comparable proportion among women treated with excisional or ablative surgery was 16%, which is nearly two times higher than the rate observed in the untreated (8.4%) and the colposcopy-only (8.6%) groups. Characteristics of the study population stratified by time to conception of >1 year (yes/no) are shown in Table 1. Compared with women who had a longer time to pregnancy, those who required <12 months to conceive were significantly more likely to have a college degree, a normal pre-pregnancy BMI, to have been a non-smoker prior to pregnancy, and were less likely to have had cervical surgery for the treatment of CIN. There were no significant differences between the two groups according to maternal age, household income, race, parity, and case–control status (i.e. preterm, SGA, or control).

Table 1. Characteristics of the 1324 study participants by time to pregnancy in the Iowa Health in Pregnancy Study, Iowa City, IA, USA, 2002–2005
 ≤12 months to pregnancy (n = 1200)>12 months to pregnancy (n = 124)Pa
  1. a

    Chi-square P value for the group difference ≤12 months to conception versus >12 months to conception.

Age
18–19 years63 (5.3)8 (6.5)0.36
20–24 years134 (11.2)9 (7.3)
25–29 years450 (37.5)41 (33.1)
30–34 years389 (32.4)43 (34.7)
35–45 years164 (13.7)23 (18.6)
Education
High school graduate or less136 (11.3)30 (24.2)0.0002
Some college, vocational education, or Associate's degree343 (28.6)30 (24.2)
College graduate (Bachelor's degree)494 (41.2)37 (29.8)
One or more years graduate/professional school227 (18.9)27 (21.8)
Household income
<$30 000144 (12.5)21 (17.7)0.06
$30 000–50 000190 (16.5)29 (24.4)
$50 000–75 000350 (30.4)32 (26.9)
$75 000–100 000255 (22.1)21 (17.7)
>$100 000214 (18.6)16 (13.5)
Race
White1071 (89.3)107 (86.3)0.32
Non-white129 (10.8)17 (13.7)
Parity
0542 (45.2)61 (49.2)0.83
1440 (36.7)42 (33.9)
2157 (13.1)16 (12.9)
3–661 (5.1)5 (4.0)
Prepregnancy BMI
Underweight (<18.5)76 (6.3)10 (8.1)0.03
Normal (18.5–25)682 (56.8)55 (44.4)
Overweight (25–30)251 (20.9)28 (22.6)
Obese (>30)191 (15.9)31 (25.0)
Prepregnancy smoking
Yes216 (18.0)34 (27.4)0.01
No984 (82.0)90 (72.6)
Cervical surgery
No cervical surgery935 (77.9)86 (69.4)0.04
Colposcopy or punch biopsy138 (11.5)13 (10.5)
Cone biopsy39 (3.3)7 (5.7)
LEEP/LLETZ37 (3.1)8 (6.5)
Cryosurgery38 (3.2)6 (4.8)
Laser vaporization13 (1.1)4 (3.2)
Case status
Preterm delivery316 (26.3)37 (29.8)0.33
Small for gestational age192 (16.0)24 (19.4)
Normal delivery (control group)692 (57.7)63 (50.8)

As shown in Table 2, even after adjustment for confounding variables, the risk for subfertility was similar between women who had a colposcopy or punch biopsy and untreated women (aOR 0.91, 95% CI 0.49–1.70). In contrast, women treated with excisional or ablative therapy had a significantly increased risk of longer time to pregnancy compared with women in the control group (aOR 2.09, 95% CI 1.26–3.46). Specifically, LEEP and laser vaporisation were associated with a significantly increased time to pregnancy (aOR 2.47, 95% CI 1.10–5.55, and aOR 3.81, 95% CI 1.17–12.37, respectively). Women treated with cone biopsy or cryosurgery were also at an increased risk for longer time to pregnancy, although this association was not statistically significant (aOR 1.64, 95% CI 0.70–3.86, and aOR 1.72, 95% CI 0.69–4.25, respectively). Parity was also examined as a possible confounding factor, but it did not alter the risk estimates.

Table 2. Prior cervical excisional treatment associated with crude and adjusted odds of >12 months to conception, Iowa Health in Pregnancy Study, Iowa, IA, USA, 2002–2005
 UnadjustedFull modelb
OR95% CIaOR95% CIa
  1. a

    Wald statistic.

  2. b

    Full model includes household income, education (categorical), age, pre-pregnancy BMI, smoking (yes/no), and case status (PTD, SGA, or control).

  3. c

    Cervical surgeries includes cone biopsy, loop electrosurgical procedure (LEEP), cryosurgery, and laser vaporisation.

Cervical treatment (n  = 1324)
No cervical treatment1.00(Ref.)1.00(Ref.)
Colposcopy or punch biopsy1.02(0.56–1.89)0.91(0.49–1.70)
Any cervical surgeryc2.14(1.32–3.47)2.09(1.26–3.46)

As LEEP and other excisional or ablative cervical surgery is more frequent among women who deliver preterm,[18] and because IHIPS was a case–control study of preterm delivery (PTD) and SGA births, we tested for an interaction between cervical treatment and case status (PTD, SGA, and normal grown) on the risk of a longer time to pregnancy, and found none (= 0.86). As a further check of our findings, we restricted the analysis of cervical treatment to the subgroup of control women who delivered normal (non-SGA), term infants (delivered at ≥37 completed weeks gestation; n = 790) to determine whether the increased risk of subfertility associated with cervical treatment was primarily the result of a high proportion of women being at high obstetric risk in the study population. Findings from this analysis were comparable with those observed in the full analysis sample: treated women had a significantly greater risk (aOR 3.33, 95% CI 1.66–6.67) of longer time to pregnancy than untreated women, whereas the colposcopy-only group had a risk similar to that of untreated women (aOR 1.12, 95% CI 0.47–2.62).

Discussion

Our findings suggest that cervical treatment for CIN is associated with a longer time to pregnancy, even after adjusting for important confounding factors. Women who had a history of any type of excisional or ablative cervical treatment had a two-fold higher risk of a longer time to pregnancy than untreated women. Women who had colposcopy or punch biopsy were not at increased risk of a time to pregnancy of more than 1 year.

Several potential mechanisms may account for a longer time to pregnancy following cervical treatment. The endocervix contains a large number of mucus-producing glands and crypts, which promote fertility by filtering, capturing, and storing sperm, and by facilitating sperm transport from the vagina to the uterus.[21] The mucus-producing cells are most highly concentrated in the lower portion of the endocervix, next to the external os.[22] Thus, even the shallowest excisional or ablative procedures are likely to remove some quantities of mucus-producing tissue. Loss of these structures may alter the quality (e.g. pH, viscosity, and salinity), quantity, levels of cervical mucus types (i.e. L-, G-, P-, and S-mucus), and its immunologic composition. These changes may impair effective sperm–mucus interactions and resistance to upper and lower genital tract infections, which could, in turn, lead to an increased risk of subfertility or infertility.[1] The impact of mucus-gland destruction on the subsequent quality, immunologic composition, and quantity of cervical mucus, and on the subsequent risk of infertility, is not well established.

Cervical stenosis, or the narrowing of the endocervical canal, is a fairly infrequent sequelae of cervical surgery and another possible mechanism.[23, 24] This loss of cervical architecture could impair the sperm–mucus interaction and prevent sperm from entering the uterine cavity, resulting in subfertility. Finally, the possibility that the mechanism of subfertility is related to the presence of higher grades of CIN in treated women cannot be ruled out.

Although previous studies reported no association between cervical surgery and time to pregnancy,[2-16] most were case series reports based on small numbers of treated women,[4, 6, 8, 14, 15] with incomplete follow-up,[4-6, 13, 14] and no comparison group of untreated women.[6, 8, 9, 15] In a prospective design, Ferenczy et al. monitored the time to conception from the date of a large loop excision of the transformation zone (LLETZ) procedure. They did not include a comparison group of untreated women, however, and followed the women for just 3 years, during which most women did not attempt to conceive.[13] Bigrigg et al. conducted an age-matched retrospective cohort to evaluate the association of LLETZ and an increased time to pregnancy, but the study was greatly underpowered and had incomplete follow-up.[5] Our study reports on the subfertility experience of a much larger group of women than previous studies, none of whom received fertility treatments to conceive the index pregnancy. We also conducted a more rigorous analysis by adjusting for confounding factors.

As our study may be the first to report that women with a history of cervical surgery have a significantly longer time to pregnancy, further research is needed to confirm and define the risks. A key limitation of this study is the self-reported nature of the exposure and outcome data, and the small numbers of women in each of the specific treatment modality groups. We did, however, perform analyses according to specific treatment and found an increased risk associated with each of the treatment types when compared with untreated women. Because we did not have access to study the treatment and pathology records of the women, we were unable to verify the type of cervical treatment or the cervical pathology findings; however, all but two treated women knew the type of procedure that they had received. We also minimised bias by excluding women who received fertility treatment. Of note, none of the women in our sample reported having a diagnosis of pelvic inflammatory disease within the 6 months prior to conception.

Recall bias is another potential threat to the validity of this study. Women were also asked to recall their exposure to cervical treatment and time to pregnancy an average of 9.6 months following delivery. Recall bias would be most likely to occur among women who had SGA and PTD births than among women with normal, term infants. If present, recall bias would lead to an overestimation of the risk associated with excisional or ablative treatments. To test for this bias, we looked for an association of cervical treatment with subfertility within the control group of women who delivered term and adequate-for-gestational age infants. These results are similar to those within the full study sample, indicating that recall bias is unlikely to be differential because the association of treatment for cervical dysplasia with time to pregnancy is as strong in control subjects as it is in case subjects.

The fact that women treated for cervical dysplasia are advised to abstain from sexual activity for 2–4 weeks is another possible factor that could bias results towards a longer time to pregnancy in these women. To take this delay into account, we re-ran the analyses and provided treated women with a ‘handicap’ of an additional 4 weeks to meet the definition of prolonged time to pregnancy; however, the odds ratios remained statistically significant.

Conclusion

Women with a history of cervical treatment for CIN are at increased risk of subfertility, measured as a time to pregnancy of more than 12 months. Understanding the impact of cervical surgeries on future fertility could be an important factor when deciding upon cervical treatment protocols, especially in young women. For women under evaluation for infertility, a history of destructive cervical therapies may indicate a need for an earlier use of fertility treatments such as intrauterine inseminations that bypass the cervix altogether. Therefore, a better understanding of the implications of cervical damage on future fertility is warranted.

Disclosure of interests

BJS is employed by Merck & Co., Rahway, NJ, USA. The work was completed prior to her employment and the opinions expressed do not necessarily reflect the opinions of Merck & Co.

Contribution to authorship

AFS was responsible for the concept and design of the study. AFS and KKH oversaw the data acquisition. CNS, AFS, and KKH were responsible for the overall analysis. All of the authors contributed to the interpretation of the data, and the drafting and approval of the article.

Details of ethics approval

All protocols and informed consent procedures were approved by the University of Iowa Institutional Review Board (IRB ID: 200003008; July 2012).

Funding

National Institute of Child Health and Human Development (NICHD), National Institutes of Health NICHD R01 HD39753. National Institute of Allergy and Infectious Diseases (NIAID), National Institutes of Health R21 AI1068111.

Acknowledgements

None.

Ancillary