Fertility and Birth Outcomes in Women With Epilepsy Seeking Pregnancy

Key Points

Question

Do women with epilepsy without a prior diagnosis of infertility or related disorders have the same likelihood of achieving pregnancy as their peers without epilepsy?

Findings

This multicenter cohort study of 89 women with epilepsy and 108 control women found no difference in the proportion of women who achieved pregnancy less than 12 months after enrollment in the study. Among women with epilepsy, 54 (60.7%) achieved pregnancy vs 65 (60.2%) among control women.

Meaning

Women with epilepsy without a history of infertility or related disorders seeking pregnancy had similar likelihood of achieving pregnancy compared with their peers without epilepsy.

This multicenter cohort study aims to determine if women with epilepsy without a prior diagnosis of infertility or related disorders are as likely to achieve pregnancy within 12 months as their peers without epilepsy.

Abstract

Importance

Prior studies report lower birth rates for women with epilepsy (WWE) but have been unable to differentiate between biological and social contributions. To our knowledge, we do not have data to inform WWE seeking pregnancy if their likelihood of achieving pregnancy is biologically reduced compared with their peers.

Objective

To determine if WWE without a prior diagnosis of infertility or related disorders are as likely to achieve pregnancy within 12 months as their peers without epilepsy.

Design, Setting, and Participants

The Women With Epilepsy: Pregnancy Outcomes and Deliveries study is an observational cohort study comparing fertility in WWE with fertility in control women (CW) without epilepsy. Participants were enrolled at 4 academic medical centers and observed up to 21 months from November 2010 to May 2015. Women seeking pregnancy aged 18 to 40 years were enrolled within 6 months of discontinuing contraception. Exclusion criteria included tobacco use and a prior diagnosis of infertility or disorders that lower fertility. Eighteen WWE and 47 CW declined the study, and 40 WWE and 170 CW did not meet study criteria. The Women With Epilepsy: Pregnancy Outcomes and Deliveries electronic diary app was used to capture data on medications, seizures, sexual activity, and menses. Data were analyzed from November 2015 to June 2017.

Main Outcomes and Measures

The primary outcome was proportion of women who achieved pregnancy within 12 months after enrollment. Secondary outcomes were time to pregnancy using a proportional hazard model, pregnancy outcomes, sexual activity, ovulatory rates, and analysis of epilepsy factors in WWE. All outcomes were planned prior to data collection except for time to pregnancy.

Results

Of the 197 women included in the study, 142 (72.1%) were white, and the mean (SD) age was 31.9 (3.5) years among the 89 WWE and 31.1 (4.2) among the 108 CW. Among 89 WWE, 54 (60.7%) achieved pregnancy vs 65 (60.2%) among 108 CW. Median time to pregnancy was no different between the groups after controlling for key covariates (WWE: median, 6.0 months; 95% CI, 3.8-10.1; CW: median, 9.0 months; 95% CI, 6.5-11.2; P = .30). Sexual activity and ovulatory rates were similar in WWE and CW. Forty-four of 54 pregnancies (81.5%) in WWE and 53 of 65 pregnancies (81.5%) in CW resulted in live births. No epilepsy factors were significant.

Conclusions and Relevance

Women with epilepsy seeking pregnancy without prior known infertility or related disorders have similar likelihood of achieving pregnancy, time to pregnancy, and live birth rates compared with their peers without epilepsy.

Introduction

Approximately 12.5 million women of childbearing age worldwide have epilepsy.¹ Most studies suggest birth rates in women with epilepsy (WWE) to be 37% to 88% of other groups.^2,3 Conversely, the Northern Finland Birth Cohort⁴ reported the number of children born to WWE did not differ from the reference group overall, although epilepsy not in remission was associated with fewer children. None of these studies collected information about the desire or attempts to achieve pregnancy. Birth rates could be lower in WWE because of social factors (eg, lower marriage rates or lower rates of seeking pregnancy) and/or biological factors (eg, decreased ovulatory rates). A UK survey⁵ reported 33% of WWE respondents were not considering having children because of their epilepsy. Neurologists do not have the information needed to counsel their female patients who desire pregnancy whether they are as likely to achieve pregnancy and to have a live birth as their peers.

We designed this study to determine whether WWE without a preexisting diagnosis of infertility or associated disorders who are following their clinically determined treatment regimen differ from women without epilepsy when attempting to achieve pregnancy. The primary aim was the proportion of women achieving pregnancy within 1 year. Secondary aims were comparisons between WWE and women without epilepsy for (1) live birth rates, (2) time to achieve pregnancy, (3) sexual activity rates, and (4) ovulatory rates. An additional secondary aim was to explore seizure and medication factors within WWE that could affect fertility.

To our knowledge, no prior studies prospectively compared WWE with women without epilepsy attempting to conceive. Our study approach excluded women with known infertility diagnosis or diagnoses associated with reproductive disorders (eg, polycystic ovary syndrome [PCOS]). This study specifically addressed the woman who arrives in clinic with no reason to suspect infertility other than an epilepsy diagnosis.

Methods

We enrolled WWE and control women (CW) without epilepsy seeking pregnancy at 4 academic medical centers in the United States: Brigham and Women’s Hospital, Boston, Massachusetts, New York University Medical Center, New York, Northwell Health (formerly North Shore–Long Island Jewish Health System), Great Neck, New York, and the University of Pittsburgh, Pittsburgh, Pennsylvania. We recruited WWE and CW from flyers and brochures placed in neurology and gynecology clinics at the clinical sites and in local restaurants, churches, childcare centers, hospital research bulletins, email listserves, and online postings. Institutional review board approvals were obtained from Partners HealthCare System for Brigham and Women’s Hospital, New York University, Northwell Health Institutional Review Board (formerly North-Shore Long Island Jewish Hospital Institutional Review Board), and University of Pittsburgh Medical Center. All participants provided written informed consent. The study was registered on clinicaltrials.gov (NCT01259310).

Sample Size and Power Analysis

The rate of pregnancy in WWE was hypothesized to be two-thirds the rate for CW. Pregnancy rate by 12 months after enrollment was chosen as the primary end point given its clinical relevance and consistency with the World Health Organization criteria for a diagnosis of infertility.⁶ We used published 3-month pregnancy rates to conservatively calculate sample size for pregnancy rates by 12 months after enrollment. The population-based estimate of pregnancy rate over 3 menstrual cycles is approximately 55%⁷; thus, we estimated that the rate among WWE was expected to be approximately 36%. Based on a 2-tailed comparison of rates, we estimated that a sample size of 100 women in each group would be needed to provide 79% power to detect this difference at a significance level of .05. To account for dropout, we planned to recruit a total of 110 women in each group.

Women with epilepsy and CW were included if they were aged 18 to 40 years, expressed a desire to conceive, had a steady male partner, and had discontinued any method of contraception within the previous 6 months or planned to do so within 6 months of providing informed consent. The study protocol commenced after contraception was discontinued in the latter case. For intramuscular medroxyprogesterone acetate, a woman could not be enrolled until 6 months after last injection and menses were regular. Exclusion criteria were a current attempt to achieve pregnancy for 6 or more months; demonstrated infertility with the same partner anytime in the past⁶; breastfeeding; diagnoses of infertility, PCOS, severe endometriosis, menopause, untreated thyroid disease, nipple discharge, hyperprolactinemia, or other pituitary disease; smoking more than 10 cigarettes per day; or planning to use medication to conceive or a sperm donor. Women were also excluded if their male partner had a history of abnormal semen analysis results, vasectomy reversal, or male factor infertility.

At the enrollment visit, demographic information and medical history were obtained, including results of ancillary studies, seizure types, and epilepsy treatments; a fertility questionnaire was administered; physical and neurological examinations were performed; and seizures, etiology, and epilepsy syndromes were classified using the National Institute of Neurological Disorders and Stroke Epilepsy Common Data Elements.⁸ Irody Inc designed an electronic daily diary app and the cloud data system (the Women With Epilepsy: Pregnancy Outcomes and Deliveries [WEPOD] study diary) for participants (Figure 1). We installed the app on a provided iPod Touch (Apple) or personal smartphone and instructed all participants on how to track menses and sexual intercourse as well as medication use and seizures for the epilepsy group. Participants could also enter data via a web-based interface or choose to use a paper diary.

Figure 1. Electronic Diary App for the Women With Epilepsy: Pregnancy Outcomes and Deliveries Study.

Participants had monthly visits in person or via telephone for review of their diary. Participants were instructed to conduct a home pregnancy test if menses did not begin by day 35 of their cycle. If positive, they had an initial pregnancy visit and a 1-month postpartum visit. Pregnancy losses (ie, ectopic pregnancy, miscarriage, fetal death, and abortions) were included in the positive pregnancy group. Participants were asked to sign a release to obtain medical records from their health care team and the delivery hospital to verify pregnancy outcomes. Each participant contributed only 1 pregnancy to these analyses.

For each participant, time to pregnancy was calculated as the interval between the end contraception date and 14 days after the first day of the last menstrual period. Women who did not achieve pregnancy within 12 months of beginning the protocol or who dropped out without becoming pregnant were considered censored in the analysis.

Sexual Activity

For each participant, months of daily electronic diary data with more than 80% of days with tracked sexual activity were analyzed. Sexual activity rate was calculated as the number of days with at least 1 episode of sexual intercourse divided by the number of days in eligible months prior to pregnancy diagnosis or study end.

Ovulation Status

We obtained venous blood for serum samples within 1 day of menstrual day 21 for up to 2 cycles. Samples collected within 14 days of the next menstrual period were considered in the luteal phase and included in this analysis. Progesterone level was measured in the Brigham Research Assay Core in Boston, Massachusetts, via access chemiluminescent immunoassay in batched samples. Intra-assay variation was reported as 6.11% to 11.19% and interassay variation as 6.59% to 9.57%. The menstrual cycle was determined as ovulatory if the progesterone level was 3.0 ng/mL or greater.

Statistical Analyses

We summarized all categorical variables by proportions and continuous variables by means, medians, standard deviations, and interquartile ranges. The primary analysis used a χ² test to compare the proportion of women achieving pregnancy between groups. In secondary analysis, we used Kaplan-Meier curves, the log-rank test, and Cox proportional hazards models to assess the association between time to pregnancy and epilepsy status while controlling for baseline characteristics. All factors significant at the .20 level in univariate analyses were considered in multivariable models, as were covariates that are consistently shown to affect fertility (ie, age and body mass index). The importance of each variable was assessed through likelihood ratio tests at the .10 significance level. The process of removing variables, refitting, and retesting was continued until all variables in the model had a P value of .05 or less.

Pregnancy outcomes were analyzed as proportions of miscarriages to live births across groups using a χ² test. Rates of sexual activity were estimated and compared between the WWE and CW using Poisson regression. Ovulatory status was analyzed using logistic regression and Generalized Estimating Equations to account for the intraparticipant correlation.

Separate analyses assessed the additional contribution of epilepsy factors in the epilepsy cohort only, including seizures in the 9 months prior to enrollment (yes/no) and enzyme-inducing antiepileptic drug (EIAED) use. Carbamazepine, phenytoin, phenobarbital, primidone, and oxcarbazepine were categorized as EIAEDs. Statistical significance was set at P < .05, and all P values were 2-tailed. All analyses used SAS statistical software version 9.2 (SAS Institute).

Sensitivity Analyses

Because WWE and CW were recruited from different sources out of necessity, a higher percentage of WWE signed informed consent prior to stopping contraception. To rule out any obscuration of a difference between the groups in time to pregnancy, a sensitivity analysis was performed by removing the participants enrolled prior to stopping contraception.

Race and ethnicity were self-reported. National survey data of infertility in the United States examined racial and ethnic differences. The only category that differed was non-Hispanic black women, who were more likely to be infertile than non-Hispanic white women.⁹ As a sensitivity analysis, we repeated the analysis after removing non-Hispanic black women.

Prior studies demonstrate that the probability of conception is highest in the first cycle without contraception and drops off with each subsequent cycle.^7,10,11 Because participants could enroll until 6 months after contraception discontinuation, we compared the proportion of WWE and CW enrolling less than 3 months and 3 to 6 months after discontinuing contraception.

Results

We enrolled 89 WWE and 108 CW (Table 1) from November 2010 through February 2014. Primary reasons for not enrolling in the WEPOD study after initial inquiry are listed in Table 2. Recruitment was stopped below the planned 100 WWE because of slow enrollment but was partially compensated by enrollment of additional CW. There were 18 early discontinuations from the epilepsy group and 15 from the control group, who were censored at the time of dropout.

Table 1. Demographic and Clinical Characteristics^a.

Characteristic	No. (%)
	Women With Epilepsy (n = 89)	Control Women (n = 108)
Age, y
Mean (SD)	31.9 (3.5)	31.1 (4.2)
Median (range)	32 (24-40)	31 (23-40)
Raceb
Asian	5 (5.6)	17 (15.7)
Black	1 (1.1)	17 (15.7)
White	78 (87.6)	64 (59.3)
Other	5 (5.6)	10 (9.3)
Ethnicity
Hispanic or Latino	11 (12.4)	14 (13.0)
Highest education level
<High school	1 (1.1)	0
High school	8 (9.0)	6 (5.6)
Some college	8 (9.0)	10 (9.3)
Associate’s degree	2 (2.2)	9 (8.3)
Bachelor’s degree	30 (33.7)	37 (34.3)
Advanced degree	39 (43.8)	46 (42.6)
Missing	1 (1.1)	0
Employment status
Student (not currently employed)	2 (2.2)	9 (8.3)
Unemployed	19 (21.3)	11 (10.2)
Employed part-time	9 (10.1)	14 (13.0)
Employed full-time	57 (64.0)	74 (68.5)
Missing	2 (2.2)	0
Marriedb	79 (88.8)	81 (75.0)
Body mass index, mean (SD)c	25.3 (5.3)	25.2 (4.7)
Prior pregnancy	46 (51.7)	63 (57.8)
Epilepsy and treatment characteristicsd
Seizure in the 9 mo prior to enrollment
Yes	39 (44.8)	NA
No	48 (55.1)	NA
Seizure type
Generalized only	27 (31.0)	NA
Focal only	57 (65.5)	NA
Unspecified	4 (4.6)	NA
AED monotherapy
Lamotrigine	39 (44.8)	NA
Levetiracetam	25 (28.7)	NA
Strong enzyme-inducing AED	11 (12.6)	NA
Other	3 (3.4)	NA
AED polytherapy
Strong enzyme-inducing AED	5 (5.7)	NA
Other	4 (5.0)	NA

Abbreviations: AED, antiepileptic drug; NA, not applicable.

^aComparisons were made between groups using χ² for race, ethnicity, educational level (less than bachelor’s degree vs bachelor’s degree and above), employment status (less than full-time vs full-time), marital status, and prior pregnancy and using t tests for age and body mass index.

^bRace and marital status significantly differed between the 2 groups.

^cBody mass index calculated as weight in kilograms divided by height in meters squared.

^dTwo participants were excluded from epilepsy factor analysis because they were not taking AEDs.

Table 2. Primary Reason for Women Not Participating in the Women With Epilepsy: Pregnancy Outcomes and Deliveries Study by Study Group^a.

Reason	No.
	Women With Epilepsy	Control Women	Total
Became pregnant before visit 1	3 (1)	3 (1)	6 (2)
Breastfeeding	0	3 (1)	3 (1)
Canceled enrollment visit	0	3	3
Current partner diagnosed with infertility	0	1	1
Demographic restrictions	0	14	14
Diagnosis of infertility	3	20	23
Diagnosis of polycystic ovary syndrome	0	1 (1)	1 (1)
Endometriosis	2	0	2
Moved out of state	3	0	3
No menstrual periods within the last year	0	17	17
Not planning pregnancy	2	6	8
Discontinued contraception >6 mo prior to enrollment	11	62	73
Other	1	1	2
Outside of age range	1	6	7
Separated from partner	1 (1)	0	1 (1)
Smoke >10 cigarettes/d	0	1	1
Still taking contraception without clear timeline to stop	7	24 (1)	31 (1)
Untreated nipple discharge	0	2	2
Untreated high prolactin level	0	1	1
Untreated thyroid problem	0	0	0
Use of sperm donor and/or fertility medications	6	5	11
Declined study, did not show, or did not respond to further contact requests	18	47	65
Total	58 (2)	217 (4)	275 (6)

^aNumber of women in each study group who did not meet study criteria by specific primary reason. Most women were excluded during the prescreening phase. The number in parentheses is the subtotal of women who signed informed consent but did not meet study criteria during the enrollment visit or shortly thereafter.

For the primary aim, we found no differences between the epilepsy and control groups in the proportion who achieved pregnancy (WWE, 60.7% [54 of 89]; CW, 60.2% [65 of 108]; χ² = 0.28; P = .60). Median time to pregnancy was 6.0 months (95% CI, 3.8-10.1) in WWE and 9.0 months (95% CI, 6.5-11.2) in CW (P = .30) (Figure 2). Women who had features of higher parity, were white, and were married were more likely to achieve pregnancy, and women with a higher age at enrollment were less likely to achieve pregnancy (eTable in the Supplement). Time to pregnancy was no different after controlling for parity, race/ethnicity, marital status, age, and body mass index.

Figure 2. Time to Pregnancy.

Cumulative proportional hazard survival for women with epilepsy (WWE) and control women (CW) of pregnancy after discontinuation of contraception. Participants who dropped out without becoming pregnant were considered censored in the analysis. The median time to pregnancy was 6.03 months (95% CI, 3.8-10.1) for WWE and 9.03 months (95% CI, 6.5-11.2) for CW (P = .30).

Pregnancy Outcomes

Pregnancy outcomes were similar, with 81.5% of pregnancies producing live births for each group (44 of 54 pregnancies in WWE and 53 of 65 pregnancies in CW). Miscarriage rates were similar, occurring in 8 pregnancies (14.8%) in WWE and in 12 pregnancies (18.5%) in CW. Additionally, of the WWE pregnancies, 1 ectopic pregnancy (1.9%) and 1 therapeutic abortion caused by genetic anomalies (1.9%) occurred.

Sexual Activity

A total of 145 participants had months with more than 80% daily tracking of sexual activity and were included in the analysis, including 57 of 89 WWE (64.0%) and 88 of 108 CW (81.5%). We found no difference in sexual activity rates between WWE and CW (Figure 3).

Figure 3. Distribution of Mean Sexual Activity Rates.

Distribution of mean sexual activity rates for women with epilepsy and control women. Sexual activity rate was calculated as the number of days with at least 1 episode of sexual intercourse divided by the number of days in eligible months prior to pregnancy diagnosis or study end.

Ovulation Status

A total of 243 cycles were included in the analysis. Six cycles in WWE and 17 cycles in CW were excluded because of the progesterone sample being drawn more than 14 days prior to the next menstrual period. In WWE, 92 of 104 cycles (88.5%) were ovulatory. In CW, 113 of 139 cycles (81.3%) were ovulatory. The relation between ovulation and study group was not significant.

Epilepsy Factors

In the analysis of the epilepsy cohort, 2 participants were removed because they were not taking antiepileptic drugs (AEDs) (Table 1). All other WWE continued to use their prescribed AEDs. Presence of active seizure in the prior 9 months did not influence the likelihood of pregnancy. Women receiving an EIAED were less likely to achieve pregnancy compared with women taking other AEDs (hazard ratio, 0.456; 95% CI, 0.193-1.081).

Sensitivity Analyses

Results of sensitivity analyses were similar to initial analyses. After removing women who stopped birth control after enrollment (14 WWE and 2 CW), adjusted times to pregnancy were a median of 7.3 months (95% CI, 3.8-12.2) in WWE and 9.0 months (95% CI, 6.5-11.2) in CW (P = .75). One non-Hispanic black WWE and 16 non-Hispanic black CW were removed from the time to pregnancy analysis; findings did not change. There was no difference between groups for duration not taking birth control prior to enrollment, with 74 of 89 WWE (83.1%) and 87 of 108 CW (80.6%) enrolled within 3 months of stopping birth control and the remainder enrolled 3 to 6 months after stopping birth control.

Discussion

In this prospective study of WWE seeking pregnancy, the first such study, to our knowledge, that included controls, we provide substantial data to support that WWE have similar pregnancy rates and time to achieve pregnancy as women without epilepsy of similar age. These findings are specific to women who do not have established diagnoses of infertility or associated disorders. Prior research has suggested WWE are at higher risk of reproductive endocrine dysfunction,¹² and PCOS occurs more frequently in WWE, especially with valproate use in idiopathic generalized epilepsy syndromes.^13,14,15,16 Therefore, we excluded women with PCOS to be able to determine if women without known PCOS have impaired fertility. Our findings are directly applicable to the more common clinical scenario of WWE without a preexisting diagnosis of infertility or associated disorders seeking information from their physicians when they decide they want to become pregnant.

The US Centers for Disease Control and Prevention report approximately 6% of married women in the United States are infertile,¹⁷ although other estimates are higher.¹⁸ The lower 12-month pregnancy rates in these groups of WWE and CW can be attributed to the fact that we enrolled women up to 6 months after birth control cessation. The probability of conception is highest in the first cycle without contraception and drops with each subsequent cycle.^7,10,11,19 The proportion of WWE and CW enrolling less than 3 months and 3 to 6 months after discontinuing contraception did not differ, and thus, this was not likely to bias our data. Additionally, the ages of participants were higher in this study than other studies, which is expected to lower pregnancy rates.^10,19 Furthermore, our reported pregnancy rates do not adjust for the 33 early discontinuations.

A prospective cohort study in India²⁰ reported 3 or more AEDs and use of phenobarbital were associated with higher risk of infertility. Other studies reported that EIAEDs contribute to higher rates of menstrual disorders.^21,22 In the WEPOD study, only 9 WWE (10.1%) were taking AED polytherapy and none were taking phenobarbital, which may explain why the WEPOD study did not similarly show diminished fertility. The WEPOD study only enrolled 16 WWE (18.0%) taking an EIAED, reflective of the practices at the study sites. The hazard ratio for women taking EIAEDs to achieve pregnancy was 0.456 (95% CI, 0.193-1.081) but did not reach statistical significance. Our subgroup analyses in the epilepsy group were secondary and may not have shown differences because of low power. Therefore, the potential influence of EIAEDs on fertility is worth studying in a larger group of women.

Other studies suggest differences in fertility between active vs controlled epilepsy.^4,23 In this cohort, 39 of 87 WWE (44.8%) had seizures within the prior 9 months; we did not find a difference in pregnancy outcomes between these women and seizure-free women.

Because studies reported higher rates of anovulatory cycles and sexual dysfunction in WWE,^24,25,26 we explored these 2 important factors for pregnancy. We did not find differences in ovulation or sexual activity rates between the 2 groups. However, when the goal is to conceive, libido plays a less important role in rates of sexual activity, so our data cannot provide direct inferences regarding libido in WWE.

The number of live births and miscarriages did not differ between the 2 groups in the WEPOD study, and rates were within expectations for women who track prospectively prior to conception.^27,28 Although some studies suggest WWE are at higher risk for miscarriage,²⁹ to our knowledge, no prior studies exclusively enrolled women prior to pregnancy and likely missed capture of early miscarriages. The International Registry of Anti-Epileptic Drugs and Pregnancy study³⁰ reported increased miscarriage risk with antiepileptic polytherapy, and it is possible that our study had too few women taking polytherapy to detect this difference. However, all miscarriages in this study occurred in women on AED monotherapy. Our findings are similar to another study³¹ that found no difference in miscarriage rates in WWE (18%) compared with wives of men with epilepsy.

A unique characteristic of the WEPOD cohort is that all women planned their pregnancy in advance, differing from pregnancy registry or population studies. A strength of our study was use of a digital diary to capture medication use, seizures, sexual activity, and menstrual bleeding. This, coupled with progesterone level measurements to confirm ovulation, allowed us to address potential fertility factors that could not be studied otherwise. Women were very compliant with the diary, as has been published elsewhere.³²

Limitations

There were limitations in our study. This study does not answer if there are differences in rates of infertility or PCOS, as women with these diagnoses were excluded from both groups. We were not able to prospectively track data from the date of stopping contraception in most participants. Some women may have achieved pregnancy before they could be recruited, making pregnancy rates lower than expected. We accounted for all anticipated confounders associated with fertility. Our analysis showed expected associations (ie, age and gravidity) and some unanticipated associations (ie, race and marital status). Although inclusion criteria required a stable male partner, it is possible that being married was associated with shorter time to pregnancy because of greater availability for intercourse and/or stronger desire to achieve pregnancy expeditiously. In the enrolled women, higher rates of marriage were present in white participants (90.1% [128 of 142]) vs other races (58.2% [32 of 55]) and could have contributed to the unexpected finding of an association of white race with shorter time to pregnancy. The epilepsy group had greater proportions of white participants and married participants than the control group. However, because these covariates were incorporated into the model, it is unlikely any of these differences influenced the primary findings of no differences in pregnancy rates or time to pregnancy. The possibility remains that additional unmeasured differences between WWE and CW were not accounted for in the model and could have influenced the results, especially because different sources of recruitment were necessary.

Conclusions

In this prospective cohort study of women without an a priori diagnosis of infertility or related disorder, WWE seeking pregnancy experienced similar pregnancy rates compared with their peers without epilepsy. Secondary analyses also showed similar times to achieve pregnancy and pregnancy outcomes. Although this study design cannot exclude an increased risk for impaired fertility across all WWE, our results should reassure WWE without a prior diagnosis of infertility or related disorders and the clinicians who care for them when planning pregnancy.

Supplement.

eTable. Analysis of maximum likelihood estimates of time to pregnancy.