A prospective study of treatments for cervical intraepithelial neoplasia and fecundability

Abstract

BACKGROUND:

Treatments for cervical intraepithelial neoplasia remove precancerous cells from the cervix by excising or ablating the transformation zone. Most studies show no association between cervical intraepithelial neoplasia treatments and fertility outcomes. However, only 2 studies have examined time to pregnancy, both using retrospective study designs, with 1 study showing no association and the other showing a 2-fold increased risk of infertility (time to pregnancy >12 months) following excisional or ablative treatment.

OBJECTIVE:

We examined the association between cervical intraepithelial neoplasia treatments and fecundability.

MATERIALS AND METHODS:

We analyzed data from Pregnancy Study Online (PRESTO), a prospective cohort study of North American pregnancy planners enrolled during 2013–2019. At baseline, women reported whether they ever had an abnormal Papanicolaou test result, the number of abnormal Papanicolaou test results, and their age at first abnormal Papanicolaou test result. They also reported whether they underwent diagnostic (colposcopy) or treatment (excisional or ablative) procedures, and their age at each procedure. We restricted analyses to 8017 women with 6 or fewer cycles of attempt time at enrollment who reported receiving a Papanicolaou test in the previous 3 years. We estimated fecundability ratios and 95% confidence intervals using proportional probabilities models adjusted for sociodemographics, healthcare use, smoking, number of sexual partners, history of sexually transmitted infections, and human papillomavirus vaccination.

RESULTS:

A history of abnormal Papanicolaou test results showed little association with fecundability (fecundability ratio, 1.00; 95% confidence interval, 0.95–1.06). Likewise, receipt of colposcopy or treatment procedures, and time since treatment were not materially associated with fecundability. Results were similar when stratified by age and smoking status.

CONCLUSION:

We observed no appreciable association of self-reported history of abnormal Papanicolaou test results, colposcopy, treatments for cervical intraepithelial neoplasia, or recency of treatment with fecundability. These results agree with the majority of previous studies in indicating little effect of cervical intraepithelial neoplasia treatments on future fertility.

About 10–15% of couples in the United States experience infertility, defined as the inability to conceive within 12 months of regular unprotected intercourse.^1,2 Known or suspected risk factors for female infertility include advanced age,^3,4 genetic defects,⁵ cigarette smoking,^6–8 obesity,⁹ gynecologic disease,¹⁰ and selected environmental agents.¹¹ Use of assisted reproductive technologies, such as in vitro fertilization (IVF), has increased markedly in recent years,¹² with US healthcare costs for infertility treatment exceeding $5.5 billion.¹³ Identifying modifiable determinants of infertility may curtail these costs. Although much research has been devoted to understanding female factors relating to the uterus, tubes, and ovaries, less is known about the contribution of cervical factors to infertility.¹⁴

The cervix protects the genital tract from infections and facilitates sperm transport into the fallopian tubes for fertilization.¹⁵ Cervical intraepithelial neoplasia (CIN) is defined as abnormal growth of cells at the transformation zone of the cervix, where the columnar endocervical epithelium meets the squamous ectocervical epithelium, that could potentially lead to cervical cancer.¹⁵ Suspected CIN is identified on a Papanicolaou (Pap) test. Almost all CIN is caused by infection with human papillomavirus (HPV).¹⁶ HPV infection has been associated with reduced fertility in some studies.^17,18 The HPV vaccine, which became available in 2006, protects against strains of HPV that may cause cervical cancer (eg, types 16 and 18 in the quadrivalent vaccine, and 5 additional oncogenic types in the 9-valent vaccine).^19–21 In the sole study of HPV vaccination and fecundability, a history of HPV vaccination had little effect on fecundability overall, but vaccination was associated with higher fecundability among women with a history of sexually transmitted infections (STIs).²²

Treatment for suspected CIN involves several steps. First, a colposcopy is performed to visually examine the cervix, vagina, and vulva.¹⁷ Biopsy samples are taken of all abnormal-appearing areas, and histopathologic diagnosis of biopsy specimens is then performed to identify CIN.¹⁶ CIN is classified by morphologic and histochemical characteristics into 2 main types: low-grade (CIN1) and high-grade (CIN2 and CIN3). CIN1 is considered a marker of an active HPV infection and usually regresses without treatment, although patients usually undergo a repeat Pap test, with or without an HPV test, 12 months after diagnosis to monitor the abnormal cervical cells.¹⁸ Although CIN2 and CIN3 (also called histologic high-grade squamous intraepithelial lesion [HSIL]) may also regress, they are considered to be markers of precancerous cells. To treat CIN2 and CIN3, local ablative or excisional methods are used to destroy or remove the affected cervical cells, depending on the extent of disease.¹⁸ Ablative procedures destroy precancerous cells, most commonly with freezing (cryotherapy). No pathology sample is obtained with ablative procedures, and they are most often used with lower-grade or smaller lesions. Excisional procedures, including loop electrosurgical excision procedure (LEEP) or cold knife cone, remove the affected tissue, which is then sent for pathologic examination, and thus provide additional diagnostic precision and treatment. Conization can be performed with laser, although laser is more commonly used for ablation.²³ Of these procedures, LEEP has been used for 25 years in the United States and is the most common treatment for CIN 2/3.²³

Although the goal of all surgical procedures for CIN is to remove or destroy abnormal cervical cells, healthy endocervical tissue is also removed in order to obtain negative margins and to avoid disease recurrence. The volume of tissue removed is generally greatest for conization (a procedure usually reserved for severe precancerous lesions),²³ followed by LEEP or laser cone, followed by ablative procedures (laser ablation or cryotherapy).²² Some women with persistent or recurrent CIN2/3 may require repeat procedures, resulting in the loss of additional cervical length. Although regeneration of the cervical tissue is possible, the more tissue that is removed, the lower the likelihood of regeneration.²⁴ Excision and ablation procedures can lead to reduced cervical mucus production, decreased sperm motility, reduced protection against infection,²⁴ and, in extreme cases, cervical stenosis,²⁵ all of which could deleteriously affect fecundability. On the other hand, effective removal of HPV-affected tissue could improve fecundity, and thus the direction of the anticipated association is unclear.

A 2014 meta-analysis of 15 studies concluded that there was no evidence that treatments for CIN adversely affect fertility or early pregnancy outcomes.²⁵ Seven of the 15 studies investigated fertility outcomes, including total pregnancy rates (n = 4), pregnancy rates in women wishing to conceive within an unspecified period (n = 1), and time to pregnancy (TTP) (n = 2).^26,27 However, heterogeneity between studies was high.²⁵ In one of the two TTP studies, a retrospective cohort study,²⁶ there was little difference in TTP between the 72 women with a history of LEEP and 62 unexposed women. In contrast, in a 2013 retrospective case-control study,²⁷ women who underwent a CIN procedure (ie, LEEP, cryosurgery, cone biopsy, or laser vaporization) had 2 times the odds of TTP >12 months (odds ratio [OR], = 2.09; 95% confidence interval [CI], 1.26, 3.46) compared with untreated women. Positive associations were observed for LEEP (OR, 2.47; 95% CI, 1.10, 5.55), cryosurgery (OR, 1.72; 95% CI, 0.69, 4.25), cone biopsy (OR, 1.64; 95% CI, 0.70, 3.86), and laser vaporization (OR, 3.81; 95% CI, 1.17, 12.37), but only 151 women reported a cervical surgery.²⁷ Little effect was observed for colposcopy with biopsy alone (OR, 0.91; 95% CI, 0.49, 1.70).²⁷ Previous studies generally had limited data on potential confounders, such as cigarette smoking and history of STIs.

The present analysis uses self-reported data from a prospective preconception cohort study of North American pregnancy planners to examine the association between history of abnormal Pap test results and fecundability, defined as the average per-cycle probability of conception among noncontracepting couples. The study also evaluates the extent to which frequency and recency of abnormal Pap test results, colposcopy, and type of procedure to treat CIN influence fecundability.

Materials and Methods

Study population

Pregnancy Study Online (PRESTO) is an ongoing Web-based prospective cohort study of women who are planning their pregnancy in the United States and Canada.²⁸ The study began enrollment in June 2013, and participants are recruited mainly through advertising on social media and pregnancy-related websites. Women were eligible if they are between the ages of 21 and 45 years, not using contraception or fertility treatments (eg, hormones to trigger ovulation or assisted reproductive technologies), in a stable relationship with a male partner, and not currently pregnant. Participants completed a baseline questionnaire and bimonthly follow-up questionnaires online for up to 12 months. More than 80% of participants completed at least 1 follow-up questionnaire.²⁸ The study protocol was approved by the Boston University Medical Campus Institutional Review Board, and all participants provided informed consent.

Assessment of abnormal Pap test results, cervical procedures, and covariates

At baseline, participants reported data on sociodemographics, anthropometric variables, behavioral factors, medication use, and reproductive and medical history, including whether they had had an abnormal Pap test result, how many, and at what age(s). They also provided information on whether their abnormal Pap test result was investigated with colposcopy, whether they received treatment with either an ablation or excisional procedure, and at what age the procedure(s) occurred. Participants were asked to provide age at their first procedure. The exact phrasing of questions is shown in Figure 1.

FIGURE 1.

Survey questions about abnormal Papanicolaou (Pap) test results and cervical diagnosis and treatment procedures in Pregnancy Study Online (PRESTO)

Assessment of time to pregnancy

At baseline, participants reported their date of last menstrual period (LMP), typical menstrual cycle length (if regular), and the number of menstrual cycles during which they had been trying to conceive. For women with irregular menses, we used follow-up data to estimate their typical menstrual cycle length. On every follow-up questionnaire, female participants reported their pregnancy status and LMP date. Additional pregnancy data were obtained via linkage with birth registries in selected states (California, Florida, Massachusetts, Michigan, Ohio, Pennsylvania, and Texas) and via online searches (eg, baby registries). We calculated TTP in discrete menstrual cycles as follows: cycles trying to conceive at study entry + ((LMP date from most recent follow up questionnaire − date of baseline questionnaire)/cycle length) + 1. Participants contributed cycles from study entry until conception, fertility treatment, no longer trying to conceive, loss to follow-up, or 12 cycles, whichever occurred first.

Validation study

We carried out a validation study of self-reported data on abnormal Pap test results and cervical treatments. Institutional review board approval was obtained for this study as part of the main informed consent form. We mailed letters and medical record release forms to 518 women enrolled during January 1, 2018, to April 25, 2019, who reported an abnormal Pap test result within the previous 10 years (to ensure that records were still available). In the letter, we requested permission to obtain medical records from their provider to verify their information on abnormal Pap test results, CIN diagnoses, and treatment procedures. Twelve women reported that their records were no longer available or that they had switched medical providers. To date, 47 medical record release forms have been received from participants, and 38 medical records have been received from physician offices. All (100%) of the self-reported abnormal Pap test results were confirmed (ie, abnormal cytology). For the procedures and treatments, we confirmed by medical record 97% (30/31) of positive reports of colposcopy, 86% (6/7) of positive LEEP reports, 100% (4/4) positive reports of cryotherapy, and 0% (0/11) positive reports of conization (“cone biopsy”). There were no reports of laser ablation. Given the poor reporting accuracy of cone biopsy (most likely because participants misinterpreted the question to be asking about “colposcopic biopsy”), and the rarity of conization as a procedure in North America,²³ we did not consider this treatment further in our analysis. Nevertheless, of the women who reported a cone biopsy, 100% also reported a colposcopy, 29% reported LEEP, 4% reported laser treatment, and 16% reported cryotherapy; thus these women were analyzed in other exposure groups.

Exclusions

As of September 2019, a total of 11,120 eligible women had enrolled and completed a baseline questionnaire. We excluded 169 women because they had a baseline LMP >6 months prior to baseline, and 30 women with no prospective LMP dates or pregnancy over follow-up. To reduce the potential for selection bias, we excluded 2181 women because they had pregnancy attempt times of more than 6 cycles at study entry. Finally, to reduce the potential for detection bias for this analysis, we excluded 684 women who did not report having had a Pap test in the 3 years before enrollment. A total of 8017 women comprised the analytic sample.

Data analysis

We used life-table methods to calculate the percentage of women who conceived during 12 months of follow-up.²⁹ We evaluated the association of abnormal Pap test result (ever vs never), age at first abnormal Pap test result (years), and frequency of abnormal Pap test result (0, 1, ≥2) with fecundability. We then evaluated the associations between each of the cervical diagnostic and treatment procedures (Figure 1) and fecundability. Finally, we examined time since treatment to the date at which these women first started trying to conceive. We reasoned that if the treatment had a deleterious effect on fecundability, the effect might wane over time if cervical tissue healed or regenerated over time. The reference category for all analyses were women who did not report any abnormal Pap test results. We used proportional probabilities regression models to estimate fecundability ratios (FR) and 95% confidence intervals (CI). The FR represents the average per-cycle probability of conception comparing exposed with unexposed women. We fit restricted cubic splines to allow for nonlinear relations between time since treatment and fecundability.^30,31 In secondary analyses, we used a log-binomial regression model to estimate the risk ratio for the association between the exposure variables and the clinical definition of infertility (TTP ≥12 vs 0 cycles).

Selection of potential confounders was guided by the literature, the drawing of causal diagrams, and associations observed in the data (Table 1). We controlled for age (<25, 25–29, 30–34, 35–39, ≥40 years), non-Hispanic white (yes vs no), married to partner (yes vs no), household income (<$50,000, $50,000–$99,000, $100,000–$149,000, ≥$150,000 USD), education (<12, 12, 13–15, 16, ≥17 years), region of residence (US Midwest, Northeast, South, West; Canada), body mass index (BMI) (<18.5, 18.5–24.9, 25–29.9, 30–34.9, ≥35 kg/m²), smoking (current, past, never), hormonal last method of contraception (yes vs no), frequency of physician visits in the previous year (<2, 2–3, ≥4 times), health insurance (private via employment, private via direct purchase, government, free clinic, pay completely out of pocket), lifetime number of sexual partners, history of STIs or pelvic inflammatory disease (PID) (yes vs no), and HPV vaccination before the age of 18 years (yes vs no). We performed separate analyses with further adjustment for parity (parous vs nulliparous), to minimize concerns about partially controlling for the outcome.³² Further control for male circumcision made little difference in the effect estimates, so we did not adjust for it in final models. We further stratified the data by age and smoking, as these variables could plausibly modify the associations between CIN-related variables and health outcomes, including fecundability. We also performed secondary analyses in which we restricted the outcome to “viable” pregnancies, censoring women with spontaneous abortion at the time of their conception.

TABLE 1.

Baseline characteristics of women by abnormal Papanicolaou (Pap) test result

Characteristica	Abnormal Pap test result
	Ever
	Untreated		Treatedb	Total (treated + untreated)	Never
Participants, n (%)	1,812		1,009	2,821 (35.2)	5,196 (64.8)
Age (y) mean)	30.5		31.3	30.8	29.5
Partner’s age (y) (mean)	32.2		32.5	32.3	31.8
Married to partner (%)	88.8		88.3	88.7	91.3
Non-Hispanic white (%)	92.5		93.0	92.8	93.5
Household income <$50,000 USD (%)	17.7		19.7	18.1	19.8
Less than a college degree (%)	27.2		32.5	28.5	26.0
Geographic region (%)
Midwest	23.1		22.7	22.9	20.5
Northeast	25.0		21.2	23.7	22.2
South	25.4		24.6	25.3	23.7
West	14.3		14.2	14.2	17.1
Canada	12.2		17.3	14.0	16.5
Physical activity (MET-hours/wk, mean)	34.3		35.4	34.5	33.9
Body mass index (kg/m2) (%)
<25	46.0		48.3	46.8	44.3
25–29	25.6		25.1	25.6	24.0
≥30	28.4		26.6	27.6	31.7
Daily use of multivitamins or folate (%)	81.9		78.9	80.8	81.6
Insurance type (%)
Private insurance purchased individually	15.0		15.8	15.3	14.8
Private insurance through employer	78.3		70.0	75.5	73.8
Government program	28.6		38.0	31.7	34.5
Free clinic or health department	8.3		9.6	8.7	9.1
Completely out-of-pocket	0.2		0.1	0.5	0.7
Primary care provider frequency per year (%)
<2	41.2		35.8	39.4	41.4
2–3	42.1		46.9	43.6	41.5
≥4	16.7		17.3	17.0	17.1
Current regular or occasional smoker (%)	11.6		15.7	12.8	8.8
Past smoker (%)	16.7		20.9	18.1	12.9
Alcohol intake (drinks/wk) (mean)	3.7		3.7	3.7	2.9
Perceived stress scale score (mean)	16.3		16.3	16.3	16.1
Gravid (%)	52.0		57.3	53.6	51.1
Parous (%)	34.2		36.5	34.9	34.1
Irregular menses (%)	16.6		18.5	17.2	16.9
Doing something to improve conception chances (%)	76.7		77.9	77.1	77.3
Intercourse frequency <1 time/wk (%)	19.9		19.8	20.1	21.3
Intercourse frequency ≥4 times/wk (%)	15.7		16.7	16.0	16.1
Hormonal last method of contraception (%)	41.9		41.7	42.0	37.4
Maternal history of miscarriage (%)	40.2		42.0	40.8	42.2
History of infertility (%)	11.0		13.6	11.9	13.5
Male partner circumcised (%)	78.5		76.6	77.9	75.2
Number of sexual partners (mean)	9.9		10.4	10.1	6.9
History of sexually transmitted infection (%)
Chlamydia	12.4		13.0	12.6	5.9
Genital warts	4.9		7.9	6.0	1.6
Herpes	6.1		8.0	6.8	3.4
Pelvic inflammatory disease (%)	1.6		2.0	1.7	0.9
HPV vaccination age <18 y	14.2		9.5	12.9	14.8

HPV, human papillomavirus; MET, metabolic equivalent of task.

^aAll characteristics are standardized to age distribution of cohort at baseline;

^bLoop electrosurgical excision procedure (LEEP), laser treatment, or cryotherapy.

We performed a probabilistic quantitative bias analysis³³ in which we corrected the FR for exposure misclassification of abnormal Pap test results using validation data from other study populations.^34–38 As data on history of abnormal Pap test results and cervical treatments were ascertained prior to conception, we assumed that exposure misclassification was nondifferential. Using prior validation data on women of similar age for parameters,^35,37 we defined trapezoidal probability distributions for both sensitivity and specificity. Parameters for sensitivity ranged from 0.60 to 0.80 with lower and upper modes of 0.65 and 0.75. Parameters for specificity ranged from 0.88 to 1.00 with lower and upper modes of 0.91 and 0.93. We sampled from this distribution 500 times and calculated negative and positive predictive values. We used these values to calculate a corrected data set and to estimate adjusted FRs and simulation intervals.

We used multiple imputation to impute missing exposure, covariate, and outcome data. For women who did not complete any follow-up questionnaires, we assigned them 1 cycle of follow-up and imputed their pregnancy status (yes vs no). We generated 5 imputation data sets and combined point estimates and standard errors from each data set. Missing data for cervical procedures ranged from 6% (laser treatment, cryosurgery) to 9% (loop excision, abnormal Pap test results). Missing data for age at each cervical procedure ranged from 1% (loop excision, abnormal Pap test results) to 11% (colposcopy). Missingness for covariates ranged from <0.1% (BMI and current smoking) to 4% (household income). There were no missing values for age. All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).³⁹

Results

In the analytic cohort, 8017 women contributed 4794 pregnancies and 30,553 menstrual cycles. Using life-table methods, 73% of women conceived during the 12-month follow-up period. Regarding the participants who did not conceive, 164 (2%) were still contributing follow-up, 215 (3%) stopped trying to conceive, 555 (7%) initiated fertility treatment, 1161 (14%) were lost to follow-up, and 1128 (14%) completed 12 cycles of attempt time. Women with and without a history of abnormal Pap test results were similar in age (30.8 vs 29.5 years) (Table 1). A history of abnormal Pap test results was positively associated with current or past smoking, number of sexual partners, irregular menses, and a history of STI or PID. A history of abnormal Pap test results was inversely associated with being married, income, BMI, education, doing something to improve one’s chances of conception (eg, monitoring cervical mucus, using ovulation predictor kits), and a history of miscarriage or infertility.

Among the 2821 women (35% of cohort) who reported a history of abnormal Pap test results, almost all women reported a colposcopy procedure (99.7%). Among women reporting a colposcopy procedure, 1049 (37%) of women reported any treatment procedure. Among women reporting treatment(s), LEEP (26.6%) and cryosurgery (13.9%) were the 2 most common treatments, followed by laser treatment (3.8%). The percentages of women reporting 1, 2, or ≥3 abnormal Pap test results were 18.4%, 8.3%, and 8.4%, respectively. The median age at first abnormal Pap test result was 23 years (interquartile range [IQR], 20–26 years).

The FR for the association between a history of abnormal Pap test and fecundability was 1.00 (95% CI, 0.95, 1.06) (Table 2). The FR did not differ materially when we examined number of abnormal Pap tests (1 vs none: FR, 0.99; 95% CI, 0.92, 1.06; ≥2 vs none: FR, 1.02; 95% CI, 0.94, 1.09). As nearly all women with a history of abnormal Pap test reported undergoing colposcopy (99%), the FR for any colposcopy (vs no abnormal Pap test result) was identical to the FR for abnormal Pap test (1.00; 95% CI, 0.95, 1.06), and the FR for women reporting colposcopy only (ie, with no further treatment) was 1.00 (95% CI, 0.93, 1.07). Fecundability was also similar among women reporting any form of cervical treatment (LEEP, laser, or cryosurgery) (FR, 1.00; 95% CI, 0.91, 1.10), as well as excisional (LEEP) treatment (FR, 1.02; 95% CI, 0.91, 1.13) and ablative treatments (laser treatment: FR, 1.13; 95% CI, 0.88, 1.46; cryotherapy: FR, 0.94; 95% CI, 0.82, 1.08). In a model that mutually adjusted for each treatment, the results were similar but less precise. Further adjustment for parity made little difference in the effect estimates.

TABLE 2.

Association of abnormal Papanicolaou (Pap) test results and CIN procedures with fecundability

Characteristic	No. of pregnancies	No. of cycles	Unadjusted FR (95% CI)	Adjusted FR (95% CI)a	Adjusted FR (95% CI)b	Adjusted FR (95% CI)c
Abnormal Pap test result
None	3133	19,601	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)
Any	1661	10,952	0.97 (0.92, 1.02)	1.00 (0.95, 1.06)	1.00 (0.94, 1.06)	—
1	882	5820	0.97 (0.90, 1.04)	0.99 (0.92, 1.06)	0.99 (0.92, 1.06)	—
≥2	779	5132	0.97 (0.90, 1.04)	1.02 (0.94, 1.09)	1.01 (0.94, 1.09)	—
Procedures
Colposcopy (any)	1657	10,932	0.97 (0.92, 1.02)	1.00 (0.95, 1.06)	1.00 (0.94, 1.06)	1.00 (0.94, 1.07)
Colposcopy (only)	1054	6900	0.98 (0.91, 1.04)	1.00 (0.93, 1.07)	1.00 (0.93, 1.07)	—
LEEP	429	2734	0.98 (0.89, 1.09)	1.02 (0.91, 1.13)	1.00 (0.91, 1.11)	1.02 (0.91, 1.14)
Laser treatment	69	381	1.10 (0.84, 1.44)	1.13 (0.88, 1.46)	1.15 (0.90, 1.48)	1.16 (0.90, 1.50)
Cryotherapy	221	1634	0.89 (0.78, 1.03)	0.94 (0.82, 1.08)	0.95 (0.83, 1.09)	0.92 (0.80, 1.06)
LEEP, laser treatment, or cryotherapy	603	4032	0.96 (0.88, 1.05)	1.00 (0.91, 1.10)	1.00 (0.91, 1.09)	—
Time since first procedure (y)
Abnormal Pap test result (any)
<2	306	2116	0.92 (0.82, 1.04)	0.96 (0.85, 1.08)	0.96 (0.86, 1.09)
2–4	414	2586	1.05 (0.94, 1.17)	1.06 (0.95, 1.18)	1.06 (0.95, 1.19)
5–9	608	3783	1.00 (0.92, 1.09)	1.00 (0.92, 1.10)	0.99 (0.91, 1.09)
≥10	333	2467	0.88 (0.79, 0.98)	0.97 (0.86, 1.09)	0.96 (0.86, 1.08)
Colposcopy only (no treatment)
<2	163	1162	0.90 (0.77, 1.06)	0.94 (0.80, 1.11)	0.95 (0.81, 1.11)
2–4	276	1631	1.04 (0.92, 1.18)	1.06 (0.94, 1.20)	1.07 (0.95, 1.21)
5–9	398	2559	0.99 (0.90, 1.09)	0.98 (0.89, 1.09)	0.98 (0.88, 1.08)
≥10	217	1548	0.94 (0.82, 1.07)	1.01 (0.88, 1.16)	1.00 (0.87, 1.15)
LEEP
<2	92	631	0.90 (0.69, 1.18)	0.99 (0.81, 1.20)	1.00 (0.82, 1.22)
2–4	109	685	1.01 (0.84, 1.23)	1.05 (0.85, 1.31)	1.05 (0.84, 1.31)
5–9	147	826	1.03 (0.88, 1.20)	1.07 (0.88, 1.29)	1.04 (0.87, 1.26)
≥10	81	592	0.82 (0.65, 1.02)	0.90 (0.68, 1.18)	0.88 (0.67, 1.16)
Laser treatment
<2	21	81	1.27 (0.77, 2.10)	1.32 (0.83, 2.09)	1.37 (0.85, 2.22)
2–4	11	86	1.06 (0.48, 2.32)	1.07 (0.51, 2.25)	1.10 (0.53, 2.30)
5–9	23	132	0.96 (0.60, 1.53)	0.96 (0.61, 1.51)	0.95 (0.61, 1.48)
≥10	14	82	1.14 (0.58, 2.24)	1.27 (0.66, 2.43)	1.31 (0.68, 2.52)
Cryosurgery
<2	26	210	0.88 (0.55, 1.42)	0.90 (0.58, 1.42)	0.92 (0.59, 1.44)
2–4	29	281	0.93 (0.58, 1.48)	0.95 (0.60, 1.50)	0.98 (0.60, 1.59)
5–9	98	551	1.08 (0.90, 1.31)	1.09 (0.90, 1.32)	1.07 (0.88, 1.31)
≥10	68	592	0.69 (0.52, 0.94)	0.78 (0.58, 1.06)	0.79 (0.58, 1.07)
LEEP, laser treatment, or cryotherapy
<2	124	806	0.97 (0.81, 1.16)	1.01 (0.85, 1.21)	1.02 (0.86, 1.22)
2–4	127	899	1.06 (0.82, 1.37)	1.05 (0.82, 1.34)	1.05 (0.81, 1.35)
5–9	222	1279	1.03 (0.88, 1.21)	1.05 (0.89, 1.23)	1.03 (0.87, 1.21)
≥10	130	1048	0.77 (0.62, 0.95)	0.86 (0.69, 1.08)	0.86 (0.69, 1.07)

Women may contribute to ≥1 procedure category: 52 women reported LEEP and laser treatment, 121 women reported LEEP and cryotherapy, 37 women reported cryotherapy and laser trreatment, and 25 women reported LEEP, laser treatment, and cryotherapy.

CI, confidence interval; CIN, cervical intraepithelial neoplasia; FR, fecundability ratio; LEEP, loop electrosurgical excision procedure.

^aAdjusted for age, marital status, race/ethnicity, income, education, geographic region, body mass index, smoking status, history of sexually transmitted infections or pelvic inflammatory disease, hormonal last method of contraception, number of sexual partners, frequency of physician visits in previous year, type of health insurance, and history of human papillomavirus vaccination age before 18 years;

^bAdjusted for all variables in footnote “a” plus parity at enrollment;

^cMutually adjusted for other types of treatment (LEEP, laser, and cryotherapy) when applicable.

Median time since diagnosis or cervical procedure was 5.0 years (IQR, 2–9 years). Overall, there was little association between time since diagnosis/treatment and fecundability (Table 2, Figure 2). Considering all treatments combined, the FR for time since treatment of <2, 2–4, 5–9, and ≥10 years relative to no abnormal Pap test were 1.01 (95% CI, 0.85, 1.21), 1.05 (95% CI, 0.82, 1.34), 1.05 (95% CI, 0.89, 1.23), and 0.86 (95% CI, 0.69, 1.08), respectively. We observed some evidence of lower fecundability with increasing time since treatment among older women, but the numbers of exposed women were small (Supplemental Table 1). Results were relatively uniform when we stratified the data by cigarette smoking status (Supplemental Table 2). Nearly identical results were observed when we restricted the outcome to viable pregnancies (Supplemental Table 3).

FIGURE 2.

Restricted cubic spline of time since treatment and fecundability, with 4 knots at the 25^th, 50^th, 75^th, and 90^th percentiles and truncated at the 95^th percentile. Adjusted for adjusted for age, marital status, race/ethnicity, income, education, geographic region, body mass index, smoking status, history of sexually transmitted infections or pelvic inflammatory disease, hormonal last method of contraception, number of sexual partners, frequency of primary care physician visits in previous year, type of health insurance, and history of human papillomavirus vaccination age before 18 years

When we corrected our overall results for exposure misclassification, using sensitivity and specificity estimates from a previous validation study³⁵ and controlling for the same set of covariates, we reached a similar conclusion regarding the association between abnormal Pap test results and fecundability (FR, 0.96; simulation interval, 0.92, 1.00). Finally, results based on the clinical definition of infertility (TTP >12 cycles) were similar to those based on the continuous TTP variable (data not shown).

Comment

Principal findings

In this North American prospective preconception cohort study, we found little association between a history of abnormal Pap test results, colposcopy with biopsy, or treatments for CIN (whether excisional or ablative in nature) and fecundability. Recency of treatment did not appreciably affect these findings. Findings were generally similar across strata of age and smoking status, factors that were hypothesized to modify the association between CIN treatment and fecundability.

Results

A previous case-control study that analyzed CIN treatment and fecundability, which also relied on self-reported exposure and TTP, showed a 2-fold increase risk of infertility (TTP >12 months) for women with CIN treatments compared with untreated women.²⁷ The study indicated an association between excisional treatments and infertility (ORs ranged from 1.64 for cone biopsy to 3.81 for laser ablation), with no association for colposcopy only (OR, 0.91). The study was retrospective in design, had small numbers of exposed women (n = 17 for laser; n = 46 for cone biopsy), and was limited to women who were already pregnant, thereby excluding women who did not conceive.²⁷ In another retrospective cohort study of TTP,²⁶ the authors found no appreciable difference in fecundability when comparing women undergoing LEEP and untreated women. Results from Bigrigg et al²⁶ and our study are consistent with the majority of other studies demonstrating no clear adverse effects of CIN treatments on fertility.²⁵

Strengths and limitations

An important study limitation is our reliance on self-reported data to ascertain a history of abnormal Pap test results, types of procedures and treatments, and ages at each treatment. We were unable to examine differences between low-grade and high-grade CIN, although most women who received treatment likely had CIN 2/3. In addition, we did not have information on the type of CIN lesion, HPV genotype, clonal/transforming infection, viral load in vaginal secretions, or partner’s HPV status, which could have independent effects on fecundability. Although we were able to verify self-report with medical record review for a portion of our sample, the majority of data are self-reported. Self-reported data on abnormal Pap test results has been shown to have mixed accuracy,^34–38 depending on the study population, ages of participants, and length of recall period; data on the accuracy of self-reported CIN treatments are limited. In these studies, the specificity and sensitivity of self-reported abnormal Pap test result history relative to medical record data in reproductive-aged women has ranged from 0.47³⁶ to 0.79³⁷ for sensitivity, and from 0.63³⁸ to 0.99³⁶ for specificity. In the present study, the accuracy of self-reported cone biopsy was particularly poor; the accuracy of all other tests and procedures assessed (eg, abnormal Pap test results, colposcopy, LEEP, and cryotherapy) ranged from 86% to 100%. Exposure misclassification would tend to be nondifferential because women were unaware of their future pregnancy chances at enrollment. Such misclassification could explain the largely null associations observed. Age is related to likelihood of abnormal Pap test results, CIN procedures, and fecundability, and the accuracy of recall tends to decrease with increasing time since treatment. Some women had as many as 20 years between their procedure and their current pregnancy attempt. When we stratified the data by age at enrollment, we observed little difference in our effect estimates. Nevertheless, unmeasured or residual confounding (eg, by completeness of HPV vaccination) and overcontrol for confounding (eg, control for factors that are down-stream effects of exposure) are possible, given that we ascertained most covariates at the time of the baseline questionnaire. Finally, there were no appreciable differences in loss to follow-up by history of abnormal Pap test results (ever, 4.2%; never, 4.5%) or cervical procedures (colposcopy, 4.2%; LEEP, 4.3%; laser, 2.8%; cryotherapy, 6.3%).

Confounding by healthcare use was an anticipated source of bias. If follow-up of abnormal Pap test results with treatment is related to underlying fecundability, and if such confounding is not adequately controlled, upward bias could explain our null results. In an effort to minimize biases due to differential healthcare use or treatment adherence, we restricted our analytic cohort to women with a Pap test within 3 years before enrollment, and we controlled for indicators of healthcare use (eg, frequency of primary care visits and type of health insurance). Finally, HPV infection in men may have deleterious effects on sperm^40–42 and potentially on male fecundity, and partners may share the same HPV infections; however, we were unable to account for the effects of male HPV infection on couple fecundability.

Our study was limited to pregnancy planners. If unplanned pregnancies are more common among more fertile women or women who never had an abnormal Pap test result, exclusion of unplanned pregnancies could have introduced selection bias. Nevertheless, we can think of no plausible reason why the relation between CIN treatments and fecundability would differ among pregnancy planners and nonplanners. Furthermore, as our study⁴³ and others^44,45 have shown, even when participation in a cohort study is associated with characteristics such as age, parity, and smoking, bias due to self-selection has little effect on the measures of association.

To our knowledge, this is the largest study on surgical treatments for CIN and fecundability, as well as the only study to analyze the effect of time since diagnosis or procedure. Other study strengths include its prospective design and geographically dispersed population, including participants residing in all US states and Canadian provinces, countries with similar guidelines on screening and treatment.²³ Data were collected on a wide range of potential confounders, including BMI, smoking, physical activity, socioeconomic position (eg, education, income), healthcare use, and male partner variables. Finally, medical record validation in a subset of participants showed high accuracy for most treatment variables.

Conclusion

In conclusion, we found no appreciable association between a history of abnormal Pap test results or treatments for CIN and fecundability. Time since the diagnosis or procedure was also unrelated to fecundability. Although the study relied primarily on self-reported data, which could explain the lack of association observed, the results are consistent with the majority of previous studies on this topic.²⁵

AJOG at a Glance.

Why was this study conducted?

Most studies show no association between treatments for cervical intraepithelial neoplasia and fertility outcomes. However, only 2 studies have examined time to pregnancy, both using retrospective study designs, with 1 study showing no association and the other showing a 2-fold increased risk of infertility (time to pregnancy >12 months) following excisional or ablative treatment. This is first prospective time to pregnancy study of the association between treatments for cervical intraepithelial neoplasia and fecundability.

Key findings

History of abnormal Papanicolaou test result, receipt of colposcopy or treatments for cervical intraepithelial neoplasia, and time since treatment were not materially associated with fecundability.

What does this add to what is known?

The results from this prospective time to pregnancy study agree with the majority of previous studies in indicating little effect of treatments for cervical intraepithelial neoplasia on fecundability.