Data Sharing Statement.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Herzog AG, Mandle HB, MacEachern DB. Association of Unintended Pregnancy With Spontaneous Fetal Loss in Women With Epilepsy: Findings of the Epilepsy Birth Control Registry. JAMA Neurol. 2019;76(1):50–55. doi:10.1001/jamaneurol.2018.3089
Is planned pregnancy associated with spontaneous fetal loss in women with epilepsy?
In the Epilepsy Birth Control Registry survey study, results indicated that unplanned pregnancy may double the risk of spontaneous fetal loss in women with epilepsy. Interpregnacy interval less than 1 year and conception age younger than 18 years or older than 37 years were also factors.
The survey finding that unplanned pregnancy may double the risk for spontaneous fetal loss in women with epilepsy warrants prospective investigation with reliable outcome verification.
If unintended pregnancy is common among women with epilepsy and is associated with increased risk of spontaneous fetal loss (SFL), it is important to develop guidelines for safe and effective contraception for this community.
To assess whether planned pregnancy is a determinant of SFL in women with epilepsy.
Design, Setting, and Participants
The Epilepsy Birth Control Registry conducted this web-based, retrospective survey between 2010 and 2014. It gathered demographic, epilepsy, antiepileptic drug (AED), contraceptive, and reproductive data from 1144 women with epilepsy in the community between ages 18 and 47 years. Data were analyzed between March 2018 and May 2018.
Main Outcomes and Measures
The primary outcome was the risk ratio (RR) with 95% confidence intervals for SFL in unplanned vs planned pregnancies. The secondary outcome was the identification of some potentially modifiable variables (maternal age, pregnancy spacing, and AED category) of SFL vs live birth using binary logistic regression.
The participants were proportionally younger (mean [SD] age, 28.5 [6.8] years), and 39.8% had household incomes of $25 000 or less. Minority women represented only 8.7% of the participants. There were 530 of 794 unplanned pregnancies (66.8%) and 264 of 794 planned pregnancies (33.2%). The risk for SFL in 653 unaborted pregnancies in women with epilepsy was greater for unplanned (n = 137 of 391; 35.0%) than planned (n = 43 of 262; 16.4%) pregnancies (RR, 2.14; 95% CI, 1.59-2.90; P < .001). Regression analysis found that the risk for SFL was greater when planning was entered alone (odds ratio [OR], 2.75; 95% CI, 1.87-4.05; P < .001) and more so when adjusted for maternal age, interpregnancy interval, and AED category (OR, 3.57; 95% CI, 1.54-8.78; P = .003). Interpregnancy interval (OR, 2.878; 95% CI, 1.8094-4.5801; P = .008) and maternal age (OR, 0.957; 95% CI, 0.928-0.986 for each year from 18 to 47 years; P = .02), but not AED category, were also associated. The risk was greater when interpregnancy interval was less than 1 year (n = 56 of 122; 45.9%) vs greater than 1 year (n = 56 of 246; 22.8%) (RR, 2.02; 95% CI, 1.49-2.72; P < .001). Relative to the younger than 18 years cohort (n = 15 of 29; 51.7%), the risks were lower for the intermediate older cohort aged 18 to 27 years (n = 118 of 400; 29.5%; RR, 0.57; 95% CI, 0.39-0.84; P < .004) and the cohort aged 28 to 37 years (n = 44 of 212; 20.8%; RR, 0.40; 95% CI, 0.26-0.62; P < .001) but not significantly different for the small number of participants in the aged 38 to 47 years cohort (n = 3 of 12; 25.0%). No individual AED category’s SFL frequency differed significantly from the no AED category.
Conclusions and relevance
The Epilepsy Birth Control Registry retrospective survey finding that unplanned pregnancy in women with epilepsy may double the risk for SFL warrants prospective investigation with outcome verification.
The World Health Organization recognizes family planning as an important issue because it “allows individuals and couples to anticipate and attain their desired number of children and the spacing and timing of their births.”1Quiz Ref ID The Epilepsy Birth Control Registry (EBCR) retrospective survey between 2010 and 2014 found that 65% of pregnancies in women with epilepsy were unplanned.2 This very high rate may exceed the high 45% to 51% range reported for the general population in the United States between 2001 and 2011.2 As in the general population, the EBCR found that younger age and minority race/ethnicity women with epilepsy were at increased risk.2 It also found that 34.7% of the unintended pregnancies occurred in women with epilepsy who were taking no contraception and that seizure type and especially interactions between enzyme-inducing antiepileptic drugs (AEDs) and hormonal contraception constituted significant risks.2 The importance of unplanned pregnancy in the general population is that a high proportion end in abortion and some, but not all, studies in the general population find that unplanned pregnancies are associated with increased rates of spontaneous fetal loss (SFL), premature delivery, low birth weight, congenital anomalies, and infant mortality.3,4 There is evidence that a history of SFL in women with a family history of epilepsy may increase the cumulative incidence of having subsequent offspring with epilepsy, from 4.7% to 21.9% in comparison with women with neither risk factor, thereby constituting a remarkable rate ratio of 4.6.5 Retrospective studies have reported either no significant increase6 or significant increase7,8 in the risk for SFL in women with epilepsy as compared with women in the general population. Although regional variations in the rates of SFL among women with epilepsy may complicate the issue,9 a prospective US study of women seeking pregnancy found no significant difference in the rates of achieving pregnancy within 12 months or in SFL between women with epilepsy and women without epilepsy.10Quiz Ref ID Maternal exposure to AED polytherapy, the presence of a major congenital malformation in one of the parents, previous SFL, and maternal age have been identified as possible predictors of SFL by European Registry of Antiepileptic Drugs and Pregnancy.11 There has also been a finding that AED use, but not epilepsy, is a risk.12 Valproate in particular has been implicated in adverse pregnancy outcomes in general but not proven to be a significant risk factor specifically for SFL.13
The EBCR is a web-based survey and educational site that gathers demographic, epilepsy, AED, contraceptive, folic acid use, and reproductive data from women with epilepsy in the community (as opposed to a clinic population).14 In this analysis, we examined the effects of pregnancy planning, ie, whether the pregnancy was planned or unplanned, on pregnancy outcomes in terms of live birth, induced abortion, and SFL in this large-scale, community-based study of women with epilepsy. We also analyzed pregnancy outcomes in association with other possible modifiable variables such as the maternal age, pregnancy spacing, and use of AEDs in general, as well as in particular AED categories.
These retrospective data come from the 1144 women with epilepsy, aged 18 to 47 years, who completed the online survey between 2010 and 2014. Women younger than 18 years were excluded from the survey owing to the difficulty in verifying online consent of minors and their guardians. In this web-based survey, all women who answered “yes” to having epilepsy and then answered “yes” to being between ages 18 and 47 years were included in the study. Those who did not meet these criteria were redirected to the option of viewing our educational material. Referral sources included Facebook (n = 581 [50.8%]), Epilepsy Foundation/epilepsy.com (n = 304 [26.6%]), internet research (n = 54 [4.7%]), neurologists (n = 52 [4.5%]) and other (n = 118 [10.3%], with each comprising <2%). Three percent provided no response. The study was approved by the Western institutional review board (Puyallup, Washington).
Reproductive data consisted of the number of pregnancies, whether the pregnancies were planned or unplanned, the types of AEDs used during pregnancies, and the outcomes of the pregnancies (live birth, induced abortion, SFL, or decline to answer). We included any loss of pregnancy before delivery, except for induced abortion, in the SFL category. We used the term unaborted to refer to pregnancies that were not medically terminated.
We categorized AED treatment as none, monotherapy, or polytherapy. Antiepileptic drugs were further grouped into 6 categories based on their effects on enzymatic metabolism: (1) no AED; (2) enzyme-inducing AEDs, which included phenobarbital, phenytoin, carbamazepine, oxcarbazepine, and topiramate in dosages greater than 200 mg daily; (3) glucuronidated AEDs, which included only lamotrigine; (4) non–enzyme-inducing AEDs, which included levetiracetam, zonisamide, gabapentin, topiramate in dosages up to 200 mg daily, lacosamide, clobazam, pregabalin, and tiagabine; (5) enzyme-inhibiting AEDs, which included only valproate; and (6) mixed categories. Valproate was listed in the enzyme-inhibiting AED category and oxcarbarbazepine in the enzyme-inducing AED category, although valproate and the 10-monohydroxy metabolite of oxcarbazepine are also glucuronidated. When there was a combination of a category that affected enzymes and a non–enzyme-inducing AED, the combination was listed by the AED category that affected enzymes. If the combination was composed of 2 or more categories that affected enzymes, they were listed as mixed.
We determined frequencies, proportions, and risk ratios (RRs) with 95% confidence intervals for live births, induced abortions, and SFL for planned and unplanned pregnancies. Spontaneous fetal loss is reported as a proportion of unaborted as well as the total number of pregnancies. We compared proportions among categorical variables using χ2 analysis. We carried out binary logistic regression analysis to determine the significance of some modifiable factors associated with SFL vs live birth in unaborted pregnancies using pregnancy planning, maternal age, interpregnancy intervals, AED use, and AED categories as predictor variables. We considered 2-tailed P values of .05 or less to be significant.
We have reported the demographic characteristics of the EBCR population previously and noted that the participants in the web-based survey were disproportionately younger and better educated but with lower household income and minority representation than women in the general population.14 The 1144 women with epilepsy reported 951 past pregnancies. Eight hundred four occurred after the age at their first seizure and were the focus for further analysis. Distribution by pregnancy age cohort was as follows: 47 (5.8%) occurred at younger than 18 years, 492 (61.2%) in the cohort aged 18 to 24 years, 253 (31.5%) in the cohort aged 28 to 37 years, and 12 (1.5%) in the cohort aged 38 to 47 years. Among 794 reported outcomes, there were 473 live births (59.6%), 141 induced abortions (17.8%), and 180 SFL (22.7%). Ten women with epilepsy declined to report their pregnancy outcomes. All 10 reported unintended pregnancies.
The outcomes of the 794 total pregnancies for which we had outcome and planning data differed by pregnancy planning category (planned vs unplanned outcomes: χ22 = 88.329; P < .001) (Table 1). Spontaneous fetal loss was the outcome for 180 of 794 pregnancies (22.7%). The risk for SFL in the 653 unaborted pregnancies was greater for unplanned (137 of 391; 35.0%) than planned (43 of 262; 16.4%) pregnancies: RR, 2.14; 95% CI, 1.59-2.90; P < .001 (Table 2). Regression analysis found that the risk for SFL in unaborted pregnancies was greater for unplanned pregnancies when planning was entered alone (OR, 2.75; 95% CI, 1.87-4.05; P < .001) and more so when adjusted for maternal age, interpregnancy interval, and AED category (OR, 3.57; 95% CI, 1.54-8.78; P = .003). Regression analysis also identified maternal age (OR, 0.957; 95% CI, 0.928-0.986 for each year from 18 to 47 years; P = .02) and interpregnancy interval (OR, 2.878; 95% CI, 1.8094-4.5801; P = .008) but not AED category as significant independent associated factors.
Quiz Ref IDRelative to the cohort younger than 18 years, which had the highest rate of SFL (n = 15 of 29; 51.7%), the risks for the other age cohorts were as follows: cohort aged 18 to 27 years (n = 118 of 400 [29.5%]; RR, 0.57; 95% CI, 0.39-0.84; P = .004), cohort aged 28 to 37 years (n = 44 of 212 [20.8%]; RR, 0.40; 95% CI, 0.26-0.62; P < .001), and the cohort aged 38 to 47 years (n = 3 of 12 [25.0%]; RR, 0.48; 95% CI, 0.17-1.37; P = .17). Among the 368 women with epilepsy who had more than 1 pregnancy, the risk for SFL was greater when the latency between pregnancies was less than 1 year (n = 56 of 122; 45.9%) compared with latencies greater than 1 year (n = 56 of 246; [22.8%]; RR, 2.02; 95% CI, 1.49-2.72; P < .001).
Quiz Ref IDAmong the 620 unaborted pregnancies for which AED data were provided, the frequencies of SFL were comparable among women with epilepsy who did (n = 142 of 505; 28.1%) and did not (n = 31 of 115; 27.0%) use AED during pregnancy. They were also similar between women with epilepsy receiving monotherapy (n = 90 of 325; 27.6%) and polytherapy (n = 52 of 180; 28.9%). A comparison of proportions among individual categories, ie, including and excluding mixed, showed no significant difference (Table 3). No individual category frequency differed significantly from that of the no AED category or the AED category with the lowest frequency, ie, glucuronidated AEDs (lamotrigine).
Spontaneous fetal loss is the outcome of approximately 15% to 25% of pregnancies in the general population.15 The Guttmacher Institute reported a rate of 17.0% for the years 1990 to 2008.16 In women with epilepsy, Annegers et al,6 in their Rochester Project and Mayo Clinic Study, found a spontaneous abortion rate (up to 20 weeks) of 14.6%, which was similar to the 13% rate generated from the pooled data of 4 previous studies. Annegers et al6 found no difference in rate from a cohort of the wives of men with epilepsy nor of nonepilepsy populations. Tomson et al,11 in the EURAP Registry, report SFL in only 9.0% (632 of 7055) of pregnancies, but participants could enter the study as late as the 16th week of gestation, by which time more than 80% of SFL may occur.11 In the prospective, rigorously monitored Women With Epilepsy: Pregnancy Outcomes and Deliveries investigation of women with epilepsy and control individuals seeking pregnancy,9 miscarriage was reported in 12.9% of women with epilepsy and 19.7% of control individuals. If, as the EBCR findings suggest, unplanned pregnancies make up 65% of pregnancies in women with epilepsy,2 and unplanned pregnancies in women with epilepsy carry a 2.14 risk ratio for SFL, the SFL rate of 12.9% in planned pregnancies in the WEPOD study would suggest an overall SFL rate of 22.5% in women with epilepsy. This is very similar to the EBCR retrospective survey finding that 22.7% of pregnancies ended in SFL. Consistent with general population reports,3,17 the EBCR found that pregnancy planning, maternal age, and interpregnancy latency were significantly associated with SFL.
Spontaneous fetal loss was 2 times more common in unplanned (n = 137 of 391 [35.0%]) than planned (n = 43 of 262 [16.4%]) pregnancies. This is particularly notable because most EBCR pregnancies (n = 530 of 794 [66.8%]) and unaborted pregnancies (n = 391 of 653 [59.9%]) were unplanned. As in the general population, the EBCR found that SFL rates were highest for pregnancies younger than 18 years (51.9%) and declined with age (20.8% for pregnancies between ages 28-37 years) except for the oldest cohort, aged 38 to 47 years.17Quiz Ref ID As in the general population, interpregnancy latencies shorter than 1 year were associated with a 2-fold greater frequency of SFL (45.9% vs 22.8%).18 Antiepileptic drug use was not significantly associated with SFL. This is consistent with the finding of Annegers et al,6 who found rates of 14.6% with use of AED and 15.7% with no use of AED. Bech et al12 found that 16.0% of pregnant women who took AEDs had SFL. This constituted a statistically significantly greater risk of SFL by 13% in comparison with women who did not take AEDs.12 However, they found no such association when restricting the analysis to women with an epilepsy diagnosis.12 The EBCR survey found the highest frequency of SFL to occur on valproate (n = 21 of 65 [32.3%]), but the frequency was not significantly greater than for no AED (n = 31 of 115 [27.0%]) or other AED categories (n = 20 of 88 [25.3%] to n = 64 of 214 [29.9%]).
There are potential limitations to the reliability of survey data that are not verified by medical records.19 Stigmatized health information tends to be underreported regardless of the method of investigation.19 A comparison of the medical records of 104 US women aged 27 to 30 years in 1990 to 1991 with their self-reported abortion histories revealed that 19% of these women failed to report 1 or more abortions.19 The EBCR induced abortion rate of 26.2% for unplanned pregnancies was substantially lower than the 42% estimated by the Guttmacher Institute for unplanned pregnancies in the general population.17 This raises the concern that induced abortions might have been underreported by participants in the EBCR survey in favor of SFL even if participants remained anonymous and had the option to decline to answer the question of pregnancy outcome. Ten women with epilepsy (1.2%) did decline. There are also demographic concerns regarding the EBCR survey, especially the underrepresentation of minority individuals. Hispanic and black women in the general population of the United States have higher induced abortion rates than white women. The higher rate in Hispanic women is associated with a higher rate of unintended pregnancy whereas the higher rate in black women is associated not only with a higher rate of unintended pregnancy but also with a higher rate of induced abortions in these unintended pregnancies.17 The application of a correction to the EBCR unplanned pregnancy abortion rate so that the 10 declined responses are categorized as induced abortions and a correction for the underrepresentation of black women (3.9% vs 13.6%) and the finding that they have a 38.9% higher rate of induced abortions for unintended pregnancies17 would yield an abortion rate of 33.0%.19 The correction would still result in a substantially and significantly greater rate of SFL in unplanned (n = 101 of 355; 28.4%) than planned (n = 43 of 262; 16.4%) unaborted pregnancies in women with epilepsy (RR, 1.72; 95% CI, 1.26-2.38; P = .001). The younger age and higher education of participants would not be expected to underestimate the abortion rate in comparison with the general population.17
Epilepsy practice guidelines suggest that women with epilepsy achieve optimal seizure control on the minimum effective dosage of AED and take recommended dosages of folic acid supplement before conception to achieve optimal maternal and fetal outcomes.20-24 The timing of conception by women with epilepsy, therefore, often requires planning. However, previously reported EBCR findings, have suggested that most pregnancies in women with epilepsy (65%) are unplanned.2 In terms of possible causality, EBCR findings have suggested that 30% of women with epilepsy may be using contraceptive methods that are not considered to be highly effective in the general population (withdrawal, barrier, or combinations thereof), and there are an additional 20% who use systemic hormonal contraception in combination with enzyme-inducing AEDs, which may double their risk.2 Moreover, 34.7% of unintended pregnancies occur in women with epilepsy who use no contraception.2 This analysis adds the finding that unplanned pregnancy may increase the risk of SFL in women with epilepsy and identifies pregnancy planning, maternal age, and interpregnancy interval as significant modifiable variables. In view of the finding of increased risk for SFL in unplanned pregnancies in women with epilepsy, and because a history of SFL in women with epilepsy may increase the risk that subsequent live-born offspring will develop epilepsy, the finding warrants prospective investigation with medical record verification of pregnancy outcomes.
Corresponding Author: Andrew G. Herzog, MD, MSc, Harvard Neuroendocrine Unit, Beth Israel Deaconess Medical Center, 422 Worcester St, Ste 303, Wellesley, MA 02481 (firstname.lastname@example.org).
Accepted for Publication: August 10, 2018.
Published Online: October 15, 2018. doi:10.1001/jamaneurol.2018.3089
Author Contributions: Dr Herzog had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Dr Herzog and Mss Mandle and MacEachern all contributed equally to the research study and manuscript.
Concept and design: Herzog, Mandle.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Herzog, MacEachern.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Herzog, MacEachern.
Obtained funding: Herzog.
Administrative, technical, or material support: All authors.
Conflict of Interest Disclosures: Dr Herzog is the principal investigator for the Epilepsy Birth Control Registry and this research, which was supported by grants from the Epilepsy Foundation and Lundbeck. Mss Mandle and MacEachern are coinvestigators who received salary support from grants from the Epilepsy Foundation and Lundbeck.
Funding/Support:Epilepsy Foundation and Lundbeck.
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Data Sharing Statement: See Supplement.