Complications during pregnancy in women with epilepsy: population-based cohort study
Dr I Borthen, Department of Gynecology and Obstetrics, Haukeland University Hospital, N-5021 Bergen, Norway. Email firstname.lastname@example.org
Objective To investigate whether women with epilepsy have an increased risk of complications during pregnancy and to explore the impact of antiepileptic drug (AED) use.
Design Population-based cohort study.
Setting Data from Medical Birth Registry of Norway based on all births in Norway 1999–2005.
Population All births (n = 372 128) delivered in Norway, ensured through linkage with the National Population Registry run by Statistics Norway. All singleton births and the first child in multiple pregnancies were included, leaving 365 107 pregnancies for analyses.
Main outcome measures Pre-eclampsia (mild and severe), gestational hypertension, eclampsia, vaginal bleeding (early and late) and preterm birth.
Results We compared 2805 pregnancies in women with a current or past history of epilepsy (0.8%) and 362 302 pregnancies in women without a history of epilepsy. Women with epilepsy had an increased risk of mild pre-eclampsia, [odds ratio 1.3: 95% confidence interval (1.1–1.5)] and delivery before week 34 [1.2: (1.0–1.5)].
Antiepileptic drugs were used in 33.6% (n = 942) of the pregnant women with epilepsy. Compared to women without epilepsy, women with epilepsy and AED use had an increased risk of mild pre-eclampsia [1.8: (1.3–2.4)], gestational hypertension [1.5: (1.0–2.2)], vaginal bleeding late in pregnancy [1.9: (1.1–3.2)], and delivery before 34 weeks of gestation [1.5: (1.1–2.0)]. No significant increase in the risk of these complications was observed in women with epilepsy not using AED. These results remained unchanged after exclusion of multiple pregnancies.
Conclusion Women with epilepsy have a low complication rate, but special attention should be paid to those using AED during pregnancy.
Epilepsy is a common neurological disorder in obstetrical practice.1 In pregnant women, the prevalence is reported to be 0.3–0.7%.2–4 Women with epilepsy are often considered at high risk in pregnancy,5 although over 90% of pregnancies in such women proceed without any apparent problems.4,6
The risk of pre-eclampsia, placental bleeding and preterm birth has been reported to be two to three times increased in women with epilepsy,7–11 but this has not been reported in other studies.1,4,12–14 It is unclear whether the risk of complications is because of epilepsy per se or the use of antiepileptic drugs (AED),13,15,16 or the combination of these factors. If the effect of AEDs is important, the change to newer AEDs may lead to a change in these risks. Understanding the influence of AEDs on pregnancy outcome is important for developing initiatives aimed at preventing adverse outcomes. This area remains understudied, with practitioners lacking important information useful for optimal clinical practice.15
Our aim of the present study was to investigate whether women with epilepsy have an increased risk of complications in pregnancy, and to explore the impact of AED use.
Our data included all births (n = 372 128) from December 1st 1998 to October 6th 2005 in the population-based Medical Birth Registry of Norway (MBRN). We included all singleton births and the first child in multiple pregnancies leaving 365 107 pregnancies for analyses. The study population was divided into pregnancies in women who gave a past or present history of epilepsy (henceforth designated as ‘women with epilepsy’), n = 2805 (0.8%) and those who did not, n = 362 302 (reference group). The MBRN is based on the compulsory notification of all births in Norway after 12 weeks of gestation, including abortions induced on medical indications.17 Complete ascertainment of the births is ensured through linkage with the National Population Registry run by Statistics Norway. Miscarriages and stillbirths before week 21 were included in our study. In Norway, more than 99% of pregnant women receive antenatal care, mainly provided by general practitioners.18 A standardised antenatal form is filled in for each pregnancy, starting at the first antenatal visit, and brought to the obstetrical unit at delivery. These data are transferred to the MBRN notification form by the midwives attending the delivery.17 Furthermore, all data related to the delivery are recorded by the midwife and follow-up data are added until discharge.17 An unchanged MBRN notification form was used in the study period. The maternal epilepsy diagnosis is notified in a checkbox in the notification form, or by a code according to the International Classification of Diseases, 10th revision (ICD-10). Data on maternal education were obtained from Statistics Norway. Data on all other variables were obtained from the MBRN.
The outcomes of interest in this study were mild or severe pre-eclampsia, gestational hypertension, eclampsia, preterm birth, premature rupture of membranes, placenta previa and early or late bleeding in pregnancy. The term ‘all pre-eclampsia’ includes mild, severe or unspecified pre-eclampsia. Pre-eclampsia is reported to the MBRN as a specific diagnosis abstracted from the medical chart. Diagnostic criteria are: an increase in blood pressure to a value higher than 140/90 mmHg (one or both values exceeded) on two measurements taken at least 6 hours apart and after 20th week of gestation. Alternatively, either the diastolic blood pressure has to be at least 15 mmHg higher or the systolic blood pressure has to be at least 30 mmHg higher, compared with measures before the 20th week. Proteinuria (protein excretion at least 0.3 g per 24 hours, usually equivalent to ≥1+ on two different urine reagent strips), also had to be present.19 Mild pre-eclampsia is defined as systolic blood pressure in the range 140–159 mmHg and diastolic blood pressure 90–109 mmHg on two measurements taken 6 hours apart after 20th week of gestation, combined with proteinuria. Severe pre-eclampsia was defined as blood pressure ≥160/110 and proteinuria of ≥0.5 g per 24 hours (or ≥2+ on two different strips).20 Pre-eclampsia diagnosed before 34th week of gestation is always included in the severe group, as are also pregnancies with a diagnosis of HELLP syndrome. A small proportion of pre-eclampsia cases were classified as unspecified (n = 601, 0.16% of the total population).
Pregnancy-induced hypertension (defined as blood pressure ≥140/90 with no proteinuria on two measurements taken at least 6 hours apart after 20th week of gestation.); eclampsia (before, during and until 7 days after delivery); placenta previa; premature rupture of membranes (>24 hours before delivery) were notified either by checkboxes in the MBRN notification form or by an ICD-10 diagnosis.
Any vaginal bleeding notified by the doctors during pregnancy (first trimester, early i.e. ≤12 weeks; and third trimester, late, i.e. ≥28 weeks), and pregestational and gestational diabetes were notified by checkboxes in the MBRN notification form.
Medication is notified according to the Anatomical Therapeutic Chemical (ATC)-Classification System, consisting of a five-digit ATC-number. Pregnancies exposed to AED were identified by including all AEDs, as registered by a relevant ACT-number (yes/no). The most commonly used AEDs were carbamazepine (46%), lamotrigine (25%) and valproate (22%). AED was mainly used as monotherapy (86%).21
Socio-demographic data included maternal age (<24, 25–29, 30–34, 35+ years); smoking during pregnancy (yes/no); parity (0, 1+) and highest attained maternal educational level (≤10 years, 10–14 years, ≥15 years). Gestational age was calculated from the ultrasonographic measurements performed at 18–19 weeks of gestation. In Norway, 97.8% of pregnant women are routinely examined with ultrasound before week 20.18 When ultrasound data were unavailable, gestational age was estimated on basis of the first day of the last menstrual period. Gestational age was categorised as <34 weeks, 34–36 weeks and ≥37 weeks. Supplement of folic acid and multivitamins (≥4 times/week) prior to pregnancy and during pregnancy was recorded by a checkbox in the MBRN form (yes/no).
Crude and adjusted odds ratios (OR) with 95% confidence limits (CI) were calculated by multiple logistic regression in SPSS 15.0 for Windows (SPSS Inc., Chicago, IL, USA). Potential confounding variables analysed as categorical variables included in the multivariable regression models were maternal age at delivery, smoking during pregnancy, maternal educational level at delivery and diabetes. Diabetes was either gestational diabetes or diabetes type 1 or 2. Diabetes as a confounding variable was not included in the analysis of bleeding. Women with pre-pregnancy hypertension were excluded in the analysis of pre-eclampsia and gestational hypertension. Folic acid was evaluated as a potential confounder during analysis.
Cross tabulated measures for small samples with expected cell-count <5, were analysed by Fisher’s Exact test and presented as unadjusted P-values with corresponding OR. Two-sided P-values <0.05 were considered statistically significant. Interactions were evaluated in stratified analysis and with interaction terms in the logistic models.
In our study population, 2805 pregnancies (0.8%) occurred in women with a current or past history of epilepsy. They were younger, had shorter pregnancies (Table 1) and lower education. Among women with past or present epilepsy, 24.7% with AED use and 28.1% without AED use had more than 14 years of education compared with 34.6% in the reference group (P < 0.001). Smoking was reported in pregnancy by 25.5% (n = 240) of women with epilepsy using AED and by 23.9% (n = 445) of those not using AED compared to 19.0% (n = 68 728) in the reference group (P < 0.001).
Table 1. Characteristics of the study population, Medical Birth Registry of Norway 1999–2005
|Age in years, mean (SD)||28.8 (5.2)|
P < 0.001*
P < 0.05*
P < 0.001*
| 29.3 (5.0)|
|Nulliparous % (n)||42.2 (1185)|
| 40.7 (147 310)|
|Length of pregnancy in days||275 (19.4)|
P < 0.001*
P < 0.001*
P < 0.05*
| 277 (17.2)|
|Weight of baby in grams||3464 (723)|
P < 0.001*
P < 0.001*
P < 0.01*
|Multiple pregnancies % (n)||4.0 (114)|
| 3.9 (14 270)|
|Chronic hypertension % (n)||4.0 (111)|
P < 0.001
P < 0.001
| 0.5 (1825)|
|Diabetes type 1 and 2% (n)||(32)|
P < 0.001
P < 0.001
| 0.6 (2197)|
|Periconsep. folat use % (n)||18.8 (526)|
P < 0.001
P < 0.001
P < 0.001
| 9.6 (34 794)|
|Folat use in preg. % (n)||43.6 (1223)|
P < 0.001
P < 0.001
P < 0.001
| 28.5 (103 380)|
Among pregnant women with epilepsy, 5.7% developed pre-eclampsia compared to 4.3% in the reference group, OR 1.3 (CI 1.1–1.6) (Table 2). For mild pre-eclampsia the frequencies were 3.6% versus 2.5%, respectively, OR 1.4 (1.2–1.7). There was no increased risk of placenta previa, OR 0.7 (0.3–1.7), premature rupture of membranes, OR 1.1 (1.0–1.3) or eclampsia for pregnant women with epilepsy.
Table 2. Pregnancy outcomes in epileptic women compared with women with no epilepsy. Crude and adjusted measurements. Medical Birth Registry of Norway 1999–2005
|All||4.3 (15 568)||5.7 (153)||1.3 (1.1–1.6)||0.001||1.3 (1.1–1.5)||0.001|
|Mild||2.5 (9152)||3.6 (96)||1.4 (1.2–1.7)||0.001||1.4 (1.1–1.7)||0.002|
|Severe||1.6 (5823)||2.0 (53)||1.2 (0.9–1.6)||0.149||1.2 (0.9–1.6)||0.175|
|Gestational hypertension||1.9 (6795)||2.2 (58)||1.2 (0.9–1.5)||0.309||1.2 (0.9–1.5)||0.262|
|Early ≤ 12 weeks||2.1 (7439)||1.8 (50)||0.9 (0.7–1.2)||0.314||0.9 (0.7–1.1)||0.281|
|Late ≥ 28 weeks||0.8 (2794)||1.0 (28)||1.3 (0.9–1.9)||0.171||1.3 (0.9–1.9)||0.206|
|<34 weeks||3.2 (11 733)||4.0 (112)||1.2 (1.0–1.5)||0.025||1.2 (1.0–1.5)||0.039|
|34–36 weeks||4.8 (17 229)||5.5 (154)||1.2 (1.0–1.4)||0.069||1.1 (1.0–1.3)||0.165|
About 33.6% (n = 942) of the women with epilepsy were exposed to AED. We repeated the analyses for pregnant women with epilepsy and AED use and women with epilepsy and no AED use, separately. The risks of both all pre-eclampsia and mild pre-eclampsia were significantly increased in pregnant women with epilepsy using AED compared to the reference group, OR 1.5 (1.2–1.2.0) and 1.7 (1.2–2.3) respectively (Table 3). Pregnant women with epilepsy using AED also had increased risks of gestational hypertension, OR 1.5 (1.0–2.2), vaginal bleeding late in pregnancy 1.9 (1.1–3.2) and delivery before week 34, OR 1.6 (1.2–2.1), when compared to the reference group. Pregnant women with epilepsy with AED use and pre-eclampsia or late vaginal bleeding did not deliver before week 34 more often than women in the reference group, OR 1.0 (0.5–2.2) and OR 1.2 (0.3–5.6) respectively.
Table 3. Pregnancy outcomes in epileptic women without and with AED, compared with women with no epilepsy
|All||5.2 (92)||1.2 (1.0–1.5)||6.5 (61)||1.5 (1.2–2.0)||4.3 (15 559)|
|Mild||3.2 (56)||1.2 (0.9–1.6)||4.3 (40)||1.7 (1.2–2.3)||2.5 (9147)|
|Severe||2.0 (35)||1.2 (0.9–1.7)||1.9 (18)||1.2 (0.8–1.9)||1.6 (5819)|
|Gestational hypertension||1.8 (32)||1.0 (0.7–1.4)||2.8 (26)||1.5 (1.0–2.2)||1.9 (6792)|
|Early ≤ 12 weeks||1.5 (28)||0.7 (0.5–1.1)||2.3 (22)||1.1 (0.7–1.7)||2.1 (7432)|
|Late ≥ 28 weeks||0.8 (14)||1.0 (0.6–1.6)||1.5 (14)||1.9 (1.1–3.2)||0.8 (2791)|
|<34 weeks||3.5 (66)||1.1 (0.9–1.4)||4.9 (46)||1.6 (1.2–2.1)||3.2 (11 726)|
|34–36 weeks||5.3 (99)||1.1 (0.9–1.3)||5.8 (55)||1.2 (0.9–1.6)||4.8 (17 214)|
Among pregnant women with epilepsy not using AED, no increased risks were observed for any of the investigated complications. A borderline difference in pre-eclampsia rate was observed when comparing epilepsy women with and without AED, OR 1.4 (1.0–1.9).
We investigated whether the effects varied by parity or gestational age. No statistically significant interactions between these factors and occurrence of pre-eclampsia, eclampsia or vaginal bleeding were found in the group with epilepsy. A statistically significant interaction occurred between parity and gestational hypertension, OR 1.8 (1.0–3.0) for nulliparous and OR 0.6 (0.3–1.0) for parous, P for interaction 0.034. An interaction also occurred between AED use and parity (nonusers as reference) with respect to gestational hypertension; OR 2.4 (1.0–5.4) for nulliparous and OR 0.4 (0.2–1.0) for parous women, P for interaction 0.043.
As crude and adjusted measures were nearly similar, only the adjusted values are reported. Excluding multiple pregnancies (4.0% in women with epilepsy and 3.9% in the reference group), or miscarriages and abortions below 21 weeks of gestation (0.5% in women with epilepsy and 0.3% in the reference group), did not influence the results (data not shown). Only one (0.04%) of the pregnancies among women with epilepsy was induced due to congenital malformations below 21 weeks of gestation, compared with 110 (0.03%) in the reference group. Folic acid use was evaluated as a potential confounder and had no impact on any of the results above.
Principal findings and interpretation
This study showed that in general, pregnant women with a past or present history of epilepsy had a satisfactory antenatal course with few complications compared with the general obstetric population. Pregnant women with epilepsy using AED had a 1.5-fold increased risk of mild pre-eclampsia, a nearly two-fold increased risk of late vaginal bleeding, a 1.5-fold increased risk of gestational hypertension and delivery before week 34. Pregnant women with epilepsy without use of AED had no increased risk of any of the complications studied.
Our cohort data were based on the mandatory reporting to a population-based registry over a 6-year period. The cohort design, comprising the whole population, excludes the possibility of selection bias. The large study size provides high precision and the standardised collection of data provides high validity in the estimates of effect. Reported maternal epilepsy in 0.8% of pregnancies indicates high diagnostic sensitivity, the prevalence in other epidemiologic studies being 0.5–0.7.1,2 Earlier reports on MBRN data have demonstrated high diagnostic validity regarding maternal disease and outcome.22,23 As a large proportion of the epilepsy group was untreated, potential under-reporting of AED-treatment is a concern and any misclassification may lead to underestimating of the risks. However, information on both epilepsy and use of drugs is compulsory in the MBRN and the drug registration is made at the first visit in pregnancy. Recall bias is therefore avoided. A new study from Norway also confirms our findings of AED use in 0.3% of women during pregnancy.24 The large proportion with untreated epilepsy may reflect that pregnancy is often planned in a phase where AED is not regarded as necessary, and also that some women with more severe epilepsy may avoid pregnancy. Patients on AED may differ from untreated patients according to the type or aetiology of the epilepsy or in terms of seizure characteristics. In the MBRN, neither the type of epilepsy nor occurrence of epileptic seizures during pregnancy is notified.
Our study had the opportunity to subdivide hypertensive disorders into gestational hypertension, mild and severe pre-eclampsia. We found that women with epilepsy had an increased risk of mild pre-eclampsia but not so for severe pre-eclampsia. In previous studies where women with epilepsy with and without AED were combined, no pre-eclampsia increase in women with epilepsy was observed.1,4,14 However, one of these studies observed an increased risk for gestational hypertension.1 These studies report risks in hospital-based populations and have methodological limitations due to recruitment bias and small sample sizes. In our study, we cannot totally exclude potential confounding by other risk factors for pre-eclampsia, such as obesity.25 In a register-based study from Sweden,16 only AED-using women were included, and an increased risk of pre-eclampsia in such women was observed. This indicates that exposure to AED during pregnancy increases the risk for mild pre-eclampsia. A recent study shows that carbamazepine in particular was associated with a higher risk for pre-eclampsia.21 Our data give no clues to the underlying biological mechanisms, but we might speculate that AED have a primary effect on placenta.
The incidence of pre-eclampsia is higher in nulliparous than in parous women.25 In our epilepsy cohort, we observed a two-fold increased risk of pre-eclampsia in nulliparous compared to parous women, the same as in the reference group. With respect to gestational hypertension, we found a stronger effect of epilepsy and of AED use in nulliparous than in parous women. Nulliparous with gestational hypertension have higher maternal and fetal morbidity when compared to normotensive nulliparas.26 Women with severe hypertension are at highest risk for adverse maternal or perinatal outcomes such as placental abruption, acute renal failure, induction of labour, low birth weight, preterm birth and admission to a neonatal unit.26 However, we found no increased risk of preterm birth in pre-eclamptic or hypertensive pregnancies among women with epilepsy.
In the present study, a two-fold risk of late vaginal bleeding among women with epilepsy using AED was found. This is in contrast to recent studies where no increased risk was found, 1,4 but in accordance with a previous study from Norway.7 AED-induced folate deficiency and alteration in the metabolism of vitamin K-dependent blood clotting factors have been suggested as possible causes of vaginal bleeding in late pregnancy.27 Neonatal bleeding disorders due to a deficiency of vitamin K-dependent clotting factors have been reported with use of enzyme-inducing AED, and may be prevented by the use of prophylactic vitamin K supplements,28 recommended in Norway during the last 4 weeks of pregnancy,29 and received by 80% of the pregnant women with epilepsy treated with enzyme-inducing drugs.30 A lower use in the general population of pregnant women with epilepsy has been reported,31 and vitamin K use in our population with epilepsy was probably well below 80%, however, there is no information regarding vitamin K prescription in the MBRN. A recent study debates the use of vitamin K in late pregnancy for women taking enzyme-inducing AEDs,32 and the UK NICE guidelines now recommend 1 mg of vitamin K to the neonates at delivery, and does not recommend giving vitamin K to the women in pregnancy.33
Conclusions and implications for clinicians
Our large cohort study has demonstrated that women with epilepsy have few complications during pregnancy. This may reflect the satisfactory antenatal care in Norway, including 99% of the obstetric population.18 We have found that women with epilepsy not using AED have no increased risk of pre-eclampsia, hypertension or preterm labour. In contrast, women with epilepsy using AED have an increased risk of mild pre-eclampsia, gestational hypertension and delivery before week 34 as well as late vaginal bleeding. Special attention should be paid to nulliparous with epilepsy and AED use during pregnancy, as their risk for gestational hypertension is increased more than two-fold. Future studies that evaluate perinatal outcome in women with epilepsy with hypertensive disorders of pregnancy should focus on those with hypertension with and without proteinuria and their use of specific medication. Careful documentation of seizures, type of epilepsy and environmental factors will contribute to an even better knowledge on the course of pregnancy in women with epilepsy.
Disclosure of interest
I hereby declare on behalf of all authors that there are no conflicts of interest as mentioned in the author guidelines.
Contribution to authorship
IB conceived and designed the study, analysed and interpreted the data and drafted the article. She is the guarantor. MGE designed the study, interpreted the data and revised the article. AKD and NEG conceived and designed the study, interpreted the data and revised the article. GV interpreted the data and revised the article. All authors approved the final version.
Details of ethics approval
Research on anonymous registry data are routinely exempted from ethical review and informed consent requirements by the ethics research committees in Norway. This study was considered exempt by The Norwegian Institute of Public Health institutional review board.
This study was supported by the Norwegian Research Council through the NevroNor research programme.