Obstetric outcome in women with epilepsy: a hospital-based, retrospective study

Authors


Dr I Borthen, Department of Gynaecology and Obstetrics, Haukeland University Hospital, 5021 Bergen, Norway. Email ingrid.borthen@med.uib.no

Abstract

Please cite this paper as: Borthen I, Eide M, Daltveit A, Gilhus N. Obstetric outcome in women with epilepsy: a hospital-based, retrospective study. BJOG 2011;118:956–965.

Objective  To report the complications during pregnancy and delivery in women with epilepsy, compared with a control group without epilepsy, with special focus on potential risk factors, such as epilepsy severity and dosage of antiepileptic drugs.

Design  Hospital-based retrospective study.

Setting  Data from pregnancy notification forms and hospital case records.

Population  Women with a past or present history of epilepsy (n = 205) delivered in Bergen, Norway, in the period 1999–2006, and a matched control group of women (n = 205) without epilepsy.

Methods  Data were compared and odds ratios (ORs) with 95% confidence intervals (CIs) were calculated by multiple logistic regression models.

Main outcome measures  Pre-eclampsia (mild and severe), gestational hypertension, vaginal bleeding (early and late), caesarean section, vaginal operative delivery, postpartum haemorrhage and major malformations.

Results  Women with epilepsy using antiepileptic drugs had an increased risk of severe pre-eclampsia (OR, 5.0; 95% CI, 1.3–19.9), bleeding in early pregnancy (OR, 6.4; 95% CI, 2.7–15.2), induction (OR, 2.3; 95% CI, 1.2–4.3) and caesarean section (OR, 2.5; 95% CI, 1.4–4.7) adjusted for maternal age, parity, education, smoking, medical conditions and body mass index ≥30 kg/m2. There was also an increased risk of malformations in the offspring (OR, 7.1; 95% CI, 1.4–36.6). Women without antiepileptic drug use had increased risks of forceps delivery and preterm birth. Active epilepsy (seizures during the last 5 years) versus nonactive epilepsy did not discriminate for any of these complications; 84.5% of women with epilepsy and antiepileptic drug use were using folate.

Conclusion  Women with epilepsy using antiepileptic drugs had an increased risk of pregnancy and delivery complications, whereas women not using antiepileptic drugs had few complications. Seizures, high doses of antiepileptic drugs, obesity and lack of folate could not explain these increased risks.

Introduction

Epilepsy is a common neurological condition characterised by recurrent seizures,1 with an annual incidence varying between 40 and 80 per 100 000 worldwide.2,3 Women with epilepsy are assumed to account for 0.3–0.7% of all pregnancies in developed countries.4–7 The proportion of women using antiepileptic drugs (AEDs) in pregnancy may be even higher, considering the widespread and growing use of AEDs for pain and psychiatric conditions.8,9

Women with epilepsy have been considered to be at high risk in pregnancy,10,11 although most of these women have uneventful pregnancies.12 Older studies have indicated a two to three times increased risk of pre-eclampsia, placental bleeding and preterm birth.13,14 However, recent studies have been conflicting. Some authors have demonstrated increased risks of gestational hypertension,15,16 pre-eclampsia4,16,17 and bleeding in pregnancy4 compared with the general population, whereas others could not demonstrate such differences.5,6,18 It is unclear whether or not any increased risk of complications occurs because of epilepsy per se, the use of AEDs19 or other factors affecting women with epilepsy. If the effect of AEDs is important, the change to newer AEDs may lead to changes in these risks. A recent population-based register study from Norway reported an increased risk of pre-eclampsia during pregnancy in women with epilepsy using AEDs, but no increased risk in women with epilepsy not using AEDs.4 In addition, a register-based study from Sweden, including only women using AEDs during pregnancy, reported an increased risk of pre-eclampsia.17 This could indicate an association between exposure to AED use in pregnancy and hypertensive disorders. Pre-eclampsia and gestational hypertension are influenced by several risk factors, such as high body mass index (BMI), family history of pre-eclampsia, age, parity, chronic hypertension and pre-gestational diabetes.20–22 Population-based register studies are usually unable to evaluate all of these risk factors.

A diagnosis of epilepsy also seems to influence delivery, including the risk of induction,6,13–15,23 caesarean section,6,13,14,17,23–25 and the use of vacuum and forceps.13 Marked differences between studies may be a result of selection criteria, recording systems or population differences in complication rates for women with epilepsy, for example because of the effectiveness and type of AED treatment.5,16 It is important to clarify these increased risks and to identify individual risk factors to optimise clinical practice for women with epilepsy during pregnancy and labour.

In previous studies, we have reported an increased risk of complications in women with epilepsy using AEDs,4,23 but data on other potential risk factors were unavailable. The aim of this study was to report the complications during pregnancy and delivery in women with epilepsy, compared with a control group of women without epilepsy, focusing on potential risk factors, such as the severity of epilepsy and the type and dose of AEDs.

Materials and methods

Study population

A sample of 222 consecutive deliveries at Haukeland University Hospital with a diagnosis of maternal epilepsy in the Medical Birth Registry of Norway (MBRN),26 in the period 1 January 1999 to 31 December 2006, was identified. MBRN is based on the compulsory notification of all births in Norway after 12 weeks of gestation, including abortions induced for medical indications.27 Complete ascertainment of MBRN is accomplished through linkage to Statistic Norway. In Norway, 99% of pregnant women receive antenatal care, mainly provided by their general practitioners.28 A standardised notification form is used in pregnancy, in which social status, diseases of the mother and pregnancy outcomes are recorded by doctors.1 The data are transferred to the MBRN notification form by the midwives and doctors attending the delivery.27 Epilepsy is recorded as a checkbox or by a code according to the International Classification of Diseases, 10th revision (ICD-10) in the pregnancy notification form. Epilepsy is defined as a neurological condition characterised by recurrent epileptic seizures unprovoked by any immediately identifiable cause.

Haukeland University Hospital is the regional hospital for western Norway. It is the only local hospital of Bergen (population, 250 000) and delivers approximately 5000 babies per year. During the study period, 38 483 deliveries took place at Haukeland University Hospital, 222 of which were to women recorded with epilepsy in MBRN. All deliveries were to women residing in the county of Hordaland, except for two deliveries to one mother with epilepsy living in a neighbouring county who gave birth at Haukeland University Hospital because of the long distance to the local hospital. All women in the epilepsy group gave their consent to participate in the study after a written enquiry.

The hospital was asked for the following information: (i) the pregnancy notification form of the index pregnancy; (ii) the hospital case records from all consultations by obstetricians and neurologists before and during pregnancy; (iii) the case records from labour and delivery; (iv) the case records for the newborn; and (v) the MBRN notification form. This information was identified for all 222 deliveries. Altogether, 17 deliveries in 15 women were excluded because a diagnosis of epilepsy could not be identified in either the hospital records or pregnancy notification forms. Thus, the study cohort consisted of 205 deliveries to 170 women.

The epilepsy deliveries were divided into two groups according to the mother’s epilepsy: (i) those with epileptic seizures occurring <5 years before conceiving according to the neurological case record (termed ‘active epilepsy’); and (ii) those with seizures 5 years or more before conceiving (termed ‘nonactive epilepsy’). We identified the type of epilepsy according to the hospital case records made by a neurologist. We classified epilepsy as either generalised or focal, and as unspecified if the woman could not be assigned to either of the two groups. The control group of 205 women without epilepsy was recruited and identified from MBRN. For each delivery in the epilepsy group, one control was randomly selected among the deliveries in the same week at the same hospital as the case with epilepsy, and matched for age and parity. These control women received a written enquiry for participation in the study. If a woman did not consent, a new delivery control was selected.

Variables

The pregnancy outcomes examined in this study were as follows: previous miscarriages; pre-eclampsia (mild or severe); gestational hypertension; eclampsia; preterm birth; bleeding in pregnancy (early or late). The criteria for the outcomes were set in advance of the study and the diagnoses were abstracted from the pregnancy notification forms and/or the obstetrical hospital case records by a specialist in obstetrics (the first author). The term ‘all pre-eclampsia’ includes mild and severe pre-eclampsia. The diagnostic criteria of pre-eclampsia were as follows: an increase in blood pressure to a value higher than 140/90 mmHg (one or both values exceeded) in two measurements taken at least 6 hours apart and after the 20th week of gestation.29 Alternatively, either the diastolic blood pressure had to be at least 15 mmHg higher or the systolic blood pressure had to be at least 30 mmHg higher than measures before the 20th week. Proteinuria (protein excretion of at least 0.3 g per 24 hours, usually equivalent to ≥1+  on two different urine reagent strips) also had to be present.22 Mild pre-eclampsia was defined as a systolic blood pressure in the range 140–159 mmHg and a diastolic blood pressure in the range 90–109 mmHg on two measurements taken 6 hours apart after the 20th week of gestation, combined with proteinuria. Severe pre-eclampsia was defined as a blood pressure of 160/110 mmHg or more and proteinuria of 0.5 g per 24 hours or more (or ≥2+  on two different strips).30 Pre-eclampsia diagnosed before the 34th week of gestation was always included in the severe group, as was pre-eclampsia in women with a diagnosis of HELLP syndrome (H, haemolysis; EL, elevated liver enzymes; LP, low platelet count).

Gestational hypertension or pregnancy-induced hypertension was defined as a blood pressure of 140/90 mmHg or more with no proteinuria on two measurements taken at least 6 hours apart after the 20th week of gestation. Eclampsia (before, during and until 7 days after delivery) was defined as hypertension and generalised tonic–clonic convulsions. Any vaginal bleeding notified at 12 weeks of gestation or less was classified as early, and 28 weeks of gestation or more as late. Bleeding between 12 and 28 weeks of gestation was not included in the analyses because these women also had bleeding either before 12 weeks or after 28 weeks. In the term ‘medical conditions’ (yes/no), we included hyperprolactinaemia, asthma, hypothyroidism, congenital heart disease and psychiatric disease including depression. Diabetes (yes/no) was recorded as a categorical variable and included type 1, type 2 and gestational diabetes (i.e. positive glucose tolerance test).

We recorded data on the induction of labour, caesarean section and vaginal operative delivery (forceps and vacuum), and also examined the indication for the use of these interventions. The amount of haemorrhage postpartum, Apgar score, birthweight, major congenital malformations, late miscarriages (weeks 12–24) and perinatal death were recorded. Major congenital malformations were defined as structural defects present from birth causing significant functional impairment and/or leading to surgical intervention.31 There were no terminations/abortions as a result of congenital malformations in this study population.

Medication was extracted from the pregnancy notification form written by the general practitioners, as well as from the neurological case records, and noted 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 ATC number (yes/no). The highest dosage before and during pregnancy was noted.

Data on the mother’s age at delivery and date of delivery were obtained from MBRN. From the pregnancy notification form, we obtained information on smoking during pregnancy (yes/no), parity (0, 1+ ) and highest attained maternal educational level (<10 years, 10–12 years, >12 years). Gestational age was calculated from the ultrasonographic measurements performed at 18–19 weeks of gestation. Gestational age at birth was categorised as <32 weeks, 32–36 weeks and ≥37 weeks. Post-term pregnancies were defined as pregnancies >42 weeks. Supplements of folic acid and multivitamins daily prior to pregnancy and during pregnancy were recorded (yes/no). The dosage of folate was also recorded.

Pre-pregnancy BMI was calculated for all women and analysed in two groups: BMI < 30 kg/m2 and BMI ≥ 30 kg/m2.

Statistical analysis

Crude and adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were calculated by multiple logistic regression in spss 18.0 for Windows (SPSS Inc., Chicago, IL, USA). Potential confounding variables analysed as categorical variables in the multivariable regression models were smoking during pregnancy, mother’s age, highest maternal education, parity, BMI ≥ 30 kg/m2, diabetes and medical conditions. Information on these variables was complete, except for one missing BMI. Diabetes as a confounding variable was not included in the analysis of bleeding because bleeding is not associated with diabetes. Previous caesarean section was included as a confounder in the analyses of delivery outcomes. Pre-eclampsia and bleeding were used as confounding factors in the regression analysis of delivery outcome, as we would expect these to interfere with the results.

As crude and adjusted measures were very similar, only adjusted values are reported.

Cross-tabulated measures for small samples with an expected cell count of <5 were analysed by Fisher’s exact test and presented as unadjusted P values with corresponding ORs. Two-sided P values of <0.05 were considered to be statistically significant. Interactions were evaluated in stratified analyses and with interaction terms in the logistic models.

Results

In the epilepsy cohort of 205 deliveries, 104 had active epilepsy (i.e. seizures within the last 5 years) and 101 had nonactive epilepsy (i.e. seizures 5 years ago or more) (Table 1). We observed 48.3% (n = 99) women with generalised epilepsy, 28.8% (n = 59) with focal epilepsy and 22.9% (n = 47) with unspecified epilepsy. The mean age was 28 years in both the epilepsy and control groups. Smoking did not differ significantly between the groups. The active epilepsy group more often had a preterm birth before week 32 relative to the control group (P = 0.03), but there was no significant difference in educational level or previous miscarriages (P = 0.07 and P = 0.06, respectively). The mean birthweight in the epilepsy group was 3447 g [standard deviation (SD), 886 g] compared with 3524 g (SD, 638 g) for the control group (P = 0.3), with no statistically significant difference in birthweight for AED use or active epilepsy compared with controls. Of all women with epilepsy, 23.9% had seizures during pregnancy.

Table 1.   Characteristics of mother, pregnancy and the newborn in 205 births to mothers with epilepsy and in a matched control group
 No epilepsy (reference) (n = 205)Epilepsy, all (n = 205)Active epilepsy (n = 104)Nonactive epilepsy (n = 101)Epilepsy, no AED (n = 89)Epilepsy, AED (n = 116)
  1. *Mean (standard deviation) is reported for maternal age, BMI, gestational age, birthweight and head circumference.

  2. Independent sample t-tests: **P < 0.05; ***P < 0.001.

Maternal age (years) (SD)*28.2 (5.0)28.3 (5.0)28.0 (5.0)28.6 (5.0)28.1 (5.0)28.8 (4.9)
Education ≥12 years, % (n)43.4 (89)38.0 (78)32.7 (34)43.6 (44)36.0 (32)39.7 (46)
Smoking, % (n)24.4 (50)29.3 (60)33.7 (35)24.8 (25)30.0 (27)28.7 (33)
Folate, % (n)17.6 (36)58.5 (120)**70.2 (73)***46.5 (47)***24.7 (22)84.5 (98)***
Seizure in pregnancy, % (n)023.9 (49)44.2 (46)3.0 (3)5.6 (5)37.9 (44)
Present multiple pregnancy, % (n)0.5 (1)2.4 (5)1.0 (1)4.0 (4)**3.4 (3)1.7 (2)
Diabetes, type 1 and 2, % (n)1.5 (3)3.9 (8)5.8 (6)**2.0 (2)2.2 (2)5.2 (6)
Previous miscarriages, % (n)18.7 (38)25.5 (52)27.9 (29)23.0 (23)22.7 (20)27.6 (32)
Body mass index (BMI) (kg/cm2) (SD)*23.4 (4.0)25.1 (5.6)***25.4 (6.0)***24.8 (5.2)**24.4 (4.5)25.7 (6.2)***
Gestational age (weeks) (SD)*39.5 (2.1)38.8 (3.7)**38.8 (3.8)38.7 (3.6)**38.9 (3.8)38.6 (3.6)**
Preterm delivery <37 weeks, % (n)6.8 (14)12.7 (26)**12.5 (13)12.9 (13)10.1 (9)14.7 (17)**
Birthweight (g) (SD)*3524 (638)3447 (886)3476 (859)3417 (916)3460 (939)3437 (847)
Head circumference (cm) (SD)*35.0 (1.9)34.7 (2.4)34.8 (2.6)34.6 (2.3)34.7 (2.3)34.8 (2.5)

Among women with epilepsy, 12.7% developed pre-eclampsia, compared with 5.4% among the controls (OR, 2.3; 95% CI, 1.1–5.0) (Table 2). In addition, severe pre-eclampsia was more frequent in the epilepsy group: 5.9% versus 1.5% (OR, 4.0; 95% CI, 1.1–14.7). There were no cases of eclampsia or chronic hypertension in the epilepsy or control group, and only one woman in the epilepsy group had reported pre-eclampsia in a previous pregnancy. Generalised and nonactive epilepsy had an increased risk of pre-eclampsia (OR, 3.0; 95% CI, 1.1–8.4 and OR, 2.7; 95% CI, 1.1–6.9, respectively) and an even higher risk for severe pre-eclampsia (OR, 5.4; 95% CI, 1.1–27.0 and OR, 4.9; 95% CI, 1.1–21.9, respectively). Seizures in pregnancy were observed in 30.8% of women with pre-eclampsia. Folate was used by 76.9% of women with epilepsy and pre-eclampsia. The epilepsy group had more bleeding in early pregnancy relative to controls (OR, 3.8; 95% CI, 1.7–8.8). Generalised, focal and active epilepsy had an increased risk for this complication (OR, 4.0; 95% CI, 1.5–10.2; OR, 6.9; 95% CI, 2.4–19.5 and OR, 5.3; 95% CI, 2.1–13.3, respectively). Folate was used by 80% of women with epilepsy and early bleeding. The inclusion of folate in the regression model for bleeding in early pregnancy reduced the excess risk (OR, 2.8; 95% CI, 1.1–6.8).

Table 2.   Pregnancy and delivery outcome in 205 women with epilepsy compared with a matched control group
   No epilepsy (n = 205) (reference)Epilepsy, all (n = 205)Epilepsy, no AED use (n = 89)Epilepsy, AED use (n = 116)
% (n)% (n)Adjusted OR (95% CI)% (n)Adjusted OR (95%CI)% (n)Adjusted OR (95% CI)
  1. AED, antiepileptic drug.

  2. *Adjusted for the following categorised factors: maternal education (<10 years, 10–12 years, >12 years); smoking (no/yes); body mass index (BMI) (<30 kg/m2, ≥30 kg/m2); medical conditions (no/yes); diabetes (no/yes).

  3. **Adjusted for the following categorised factors: maternal education (<10 years, 10–12 years, >12 years); smoking (no/yes); BMI (<30 kg/m2, ≥30 kg/m2); medical conditions (no/yes).

  4. ***Adjusted for the following categorised factors: maternal education (<10 years, 10–12 years, >12 years); smoking (no/yes); BMI (<30 kg/m2, ≥30 kg/m2); medical conditions (no/yes); diabetes (no/yes); previous caesarean section (CS); pre-eclampsia and vaginal bleeding.

  5. ****Adjusted for the following categorised factors: maternal education (<10 years, 10–12 years, >12 years); smoking (no/yes); BMI (<30 kg/m2, ≥30 kg/m2); medical conditions (no/yes); diabetes (no/yes); parity (0, 1+ ); maternal age (<25, 25–29, 30–34, >35 years).

  6. *****Adjusted for the following categorised factors: maternal education (<10 years, 10–12 years, >12 years); smoking (no/yes); BMI (<30 kg/m2, ≥30 kg/m2); medical conditions (no/yes), parity (0, 1+ ); maternal age (<25, 25–29, 30–34, >35 years).

  7. ******Adjusted for the following categorised factors: maternal education (<10 years, 10–12 years, >12 years); smoking (no/yes); BMI (<30 kg/m2, ≥30 kg/m2); medical conditions (no/yes); diabetes (no/yes); parity (0, 1+ ); maternal age (<25, 25–29, 30–34, >35 years); previous CS; pre-eclampsia and vaginal bleeding.

Pregnancy outcomePre-eclampsiaAll5.4 (11)12.7 (26)2.3 (1.1–5.0)*12.4 (11)2.3 (0.9–6.0)****12.9 (15)2.2 (0.9–5.2)****
Mild3.9 (8)6.8 (14)1.6 (0.6–4.1)*9.0 (8)2.4 (0.8–7.7)****5.2 (6)1.1 (0.4–3.5)****
Severe1.5 (3)5.9 (12)4.0 (1.1–14.7)*3.4 (3)2.8 (0.5–15.7)****7.8 (9)5.0 (1.3–19.9)****
Gestational hypertension 3.9 (8)6.3 (13)1.7 (0.7–4.2)*4.5 (4)1.4 (0.4–4.9)****7.8 (9)1.8 (0.6–5.0)****
Vaginal bleeding≤12 weeks3.9 (8)12.2 (25)3.8 (1.7–8.8)**4.5 (4)1.0 (0.3–3.5)*****18.1 (21)6.4 (2.7–15.2)*****
Delivery outcomeInduction 11.2 (23)21.0 (43)1.8 (1.0–3.2)***15.7 (14)1.3 (0.6–2.9)******25.0 (29)2.3 (1.2–4.3)******
caesarean sectionAll12.7 (26)24.9 (51)1.8 (1.0–3.1)***14.6 (13)1.0 (0.4–2.2)******32.8 (38)2.5 (1.4–4.7)******
Acute8.3 (17)17.1 (35)1.9 (1.0–3.6)***11.2 (10)1.1 (0.5–2.8)******21.6 (25)2.4 (1.2–5.0)******
Planned4.4 (9)7.8 (16)1.4 (0.6–3.4)***3.4 (3)0.7 (0.2–3.5)******11.2 (13)2.1 (0.8–5.5)******
Cephalic vaginalForceps2.9 (6)7.3 (15)3.0 (1.1–8.0)***7.9 (7)4.5 (1.2–16.5)******6.9 (8)2.5 (0.8–8.1)******
Apgar score<7 after 5 min1.0 (2)2.9 (6)3.4 (0.7–17.3)***2.2 (2)2.8 (0.3–22.5)******3.4 (4)4.1 (0.7–25.5)******
Preterm birth<32 weeks1.0 (2)5.4 (11)5.9 (1.3–27.8)***5.6 (5)16.8 (2.1–133.3)******5.2 (6)5.9 (1.0–36.1)******
Malformations 1.0 (2)5.9 (12)6.5 (1.4–29.9)***4.5 (4)6.0 (1.0–34.9)******6.9 (8)7.1 (1.4–36.6)******

The induction of labour was performed in 21.0% of deliveries with epilepsy compared with 11.2% in the control group (OR, 1.8; 95% CI, 1.0–3.2) (Table 2). Pre-eclampsia represented the indication for induction in 22.0% of women with epilepsy (n = 9) and in 4.5% (n = 1) of the controls. Premature rupture of membranes led to induction in 26.8% (n = 11) of women with epilepsy and in 27.3% (n = 6) of the controls. Post-term pregnancy was the indication for induction in 18.6% (n = 8) of women with epilepsy and in 34.8% (n = 8) of the controls. None of the induced women with epilepsy had seizures immediately prior to induction. However, 41% of women with epilepsy and induction experienced seizures during pregnancy.

Caesarean section was performed in 24.9% of deliveries with epilepsy compared with 12.7% of controls (OR, 1.8; 95% CI, 1.0–3.1). As women with epilepsy had a high rate of preterm birth, we included preterm births in the regression analysis of caesarean section. Then, women with epilepsy had no increased risk of overall caesarean section (OR, 1.7; 95% CI, 1.0–3.0) (P = 0.065). The indications for caesarean section in the epilepsy deliveries were asphyxia (19.5%), epilepsy (11.8%) and failed induction (9.8%). We observed three women (1.5%) with seizures during labour; two were delivered by caesarean section. For 7.3% of deliveries with epilepsy, forceps was used, compared with 2.9% for controls (OR, 3.0; 95% CI, 1.1–8.0). The exclusion of women with previous interventional delivery (n = 27) gave a minor change in operative vaginal delivery (OR, 2.9; 95% CI, 1.1–8.0).

No increased risks of vacuum deliveries, bleeding postpartum or low Apgar score were present in the epilepsy group. In the active epilepsy group, 33.7% (n = 35) of women were delivered by caesarean section, compared with 15.8% (n = 16) in the nonactive group (P = 0.003). Planned caesarean section was undertaken more frequently in active (11.5%) than in nonactive (4%) epilepsy (P = 0.043).

Among women with epilepsy, 57% (n = 116) were using AEDs. Of these, 71.6% (n = 83) had active epilepsy. Women with epilepsy and AED use had increased risks of severe pre-eclampsia and bleeding in early pregnancy compared with the controls (OR, 5.0; 95% CI, 1.3–19.9 and OR, 6.4; 95% CI, 2.7–15.2, respectively) (Table 2). Generalised, active and nonactive epilepsy with AED use had increased risk of severe pre-eclampsia (OR, 11.7; 95% CI, 2.0–70.1; OR, 5.4; 95% CI, 1.0–6.4 and OR, 14.9; 95% CI, 1.3–165.3, respectively). All subgroups of epilepsy using AED had increased risk of bleeding in early pregnancy. Women with epilepsy and AED use also had increased risks for induction and acute caesarean section (Table 2). In the epilepsy group, nine women with AED use had severe pre-eclampsia and four of these had one or more seizures during pregnancy. Two women with epilepsy had BMI ≥ 30 kg/m2. Seven women were using folate supplementation; one used 4 mg/day. There were no differences between the active and nonactive epilepsy groups using AEDs with respect to pre-eclampsia, early bleeding, induction and caesarean section (data not shown).

No increased risks of pregnancy complications were observed for the group with epilepsy and no AED use compared with the controls. This group did not show an increased risk of induction or caesarean section; however, an increased risk for vaginal forceps delivery (OR, 4.5; 95% CI, 1.2–16.5) and preterm birth (OR, 16.8; 95% CI, 2.1–133.3) was found. The indications for forceps delivery were asphyxia (50%), prolonged second stage (25%) and hypertension (12.5%); one indication was missing. Women with active epilepsy and no AED use (n = 21) had no increased risk of any pregnancy and delivery complications. Among women with AED use, lamotrigine was used as monotherapy in 25.8% (n = 30), carbamazepine in 24.1% (n = 28) and valproate in 16.4% (n = 19) (Table 3). Various AEDs were used by 10% (n = 12) of women with AED use. Polytherapy was used by 23.3% (n = 27) of women, lamotrigine being the most common drug in combination therapy. Women with lamotrigine monotherapy had an unadjusted OR of 7.5 (95% CI, 1.4–39.0) for severe pre-eclampsia. Women with polytherapy and with lamotrigine monotherapy had an increased risk of early bleeding (unadjusted OR, 8.6; 95% CI, 2.8–26.3 and OR, 6.2; 95% CI, 2.0–19.3, respectively). The dosage of AEDs did not affect the risk in women with epilepsy, as high and low doses were equally represented in women with complications (data not shown).

Table 3.   Pregnancy and delivery outcome in 104 women with epilepsy using antiepileptic drugs as monotherapy or polytherapy compared with a control group
OutcomeNo epilepsy (reference) (n = 205)Epilepsy with lamotrigine (n = 30)Epilepsy with carbamazepine (n = 28)Epilepsy with valproate (n = 19)Epilepsy with polytherapy (n = 27)
% (n)% (n)Unadj. OR (95% CI)% (n)Unadj. OR (95% CI)% (n)Unadj. OR (95% CI)% (n)Unadj. OR (95% CI)
Pre-eclampsia
All5.4 (11)20.0 (6)4.4 (1.5–13.0)7.1 (2)1.4 (0.3–6.5)10.5 (2)2.1 (0.4–10.1)11.1 (3)2.2 (0.6–8.5)
Severe1.5 (3)10.0 (3)7.5 (1.4–39.0)8.0 (2)5.2 (0.8–32.5)5.3 (1)3.7 (0.4–37.8)7.4 (2)5.4 (0.9–33.8)
Vaginal bleeding
<12 weeks3.9 (8)20.0 (6)6.2 (2.0–19.3)14.3 (4)4.1 (1.1–14.7)15.8 (3)4.6 (1.1–19.1)25.9 (7)8.6 (2.8–26.3)
Induction11.2 (23)36.7 (11)4.6 (1.9–10.8)10.7 (3)1.0 (0.3–3.4)26.3 (5)2.8 (0.9–8.6)25.9 (7)2.8 (1.1–7.3)
caesarean section
All12.7 (26)40.0 (12)4.6 (2.0–10.6)25.0 (7)2.3 (0.9–5.9)42.1 (8)5.0 (1.8–13.6)29.6 (8)2.9 (1.2–7.3)
Acute8.3 (17)26.7 (8)4.0 (1.6–10.4)14.3 (4)1.8 (0.6–5.9)31.6 (6)5.1 (1.7–15.1)14.8 (4)1.9 (0.6–6.2)
Postpartum haemorrhage20.5 (42)25.0 (8)1.4 (0.6–3.4)21.4 (6)1.1 (0.4–2.8)47.4 (9)3.5 (1.3–9.1)36.0 (9)1.6 (0.7–4.0)
Major congenital malformation1.0 (2)10.0 (3)11.3 (1.8–70.6)8.0 (2)7.8 (1.1–57.8)5.3 (1)5.6 (0.5–65.2)7.4 (2)8.1 (1.1–60.2)

In women with epilepsy, 12 babies (5.9%) had major congenital malformations, compared with two babies (1%) in the control group (P = 0.007). Among cases with malformations, eight mothers had used AEDs during pregnancy and six mothers had not used any AEDs (P = 0.015) (Table 3). The major congenital malformations were ventricular septal defect, atrial septal defect, myelomeningocoele, diaphragm hernia, hypospadia, hip dislocation and multiple malformations. Babies of women with epilepsy and no AED use had no increased risk of malformation compared with babies of control mothers (P = 0.05).

There were two stillbirths (1.0%) in the epilepsy group at weeks 24 and 42, and one (0.5%) in the control group (P = 0.177). One woman with epilepsy had a generalised tonic–clonic seizure before stillbirth in week 24 and one was post-term. There were two late miscarriages: one chorioamnionitis in week 18, and one after abruptio placentae in week 17.

We investigated whether the association of epilepsy varied by parity. We observed no interactions with respect to pre-eclampsia, bleeding in pregnancy, induction, caesarean section, operative vaginal deliveries or preterm birth before 32 weeks. A statistically significant interaction occurred between epilepsy and parity with respect to induction (OR, 1.0; 95% CI, 0.4–2.3 for nulliparous and OR, 3.6; 95% CI, 1.6–8.1 for multiparous) (P = 0.035 for interaction).

Discussion

Principal findings and interpretation

This study demonstrates that women with epilepsy had increased risks of pre-eclampsia, early bleeding in pregnancy, induction and caesarean section. These risks were associated with the use of AEDs in pregnancy. We were unable to demonstrate any other factors influencing these risks. The epilepsy itself had less influence, as women with active epilepsy without AED use had no excessive risks. Generalised epilepsy was associated with most complications, in both pregnancy and delivery.

The identification of cases was based on mandatory reporting in a population-based registry over a 7-year period. The diagnosis of epilepsy was secured by access to the hospital case records. The reported maternal epilepsy in 0.53% is slightly lower than expected, the prevalence in other epidemiological studies being 0.5–0.8%.4,5,15,24 Earlier reports on MBRN data have demonstrated high diagnostic validity with regard to maternal disease and outcome.32–34 Selection bias might be present as Haukeland University Hospital is a primary, but also a secondary, referral hospital. However, the proportion of epilepsy deliveries in the local hospitals in the region was similar to that at Haukeland University Hospital. The standardised collection of data provides high validity in the estimates of effect, and a controlled study like ours should be optimal to assess the influence of potential risk factors on adverse pregnancy outcomes in women with epilepsy.

One limitation of our study is the lack of information on seizure type and severity. We were able to state the time of the last seizure and hence divide the epilepsy population into active and nonactive groups. Women with nonactive epilepsy and AED use had the same risks as women with active epilepsy and AED use. The incidence of AED use in pregnancy in our study was equal to that in a previous study of the total Norwegian cohort, where 0.3% of all women used AEDs during pregnancy,35 and also similar to recent MBRN studies, where 0.26% of women with epilepsy were using AEDs.4,23

Women with epilepsy had an increased risk of severe pre-eclampsia, in contrast with most previous studies.5,6,15 However, one study observed an increased risk for gestational hypertension,15 and our recent study of a large national cohort reported an increased risk of mild pre-eclampsia.4 In the present study, the classification of pre-eclampsia was made on the basis of detailed clinical information. Potential confounding by obesity36 for pre-eclampsia was excluded by using BMI > 30 kg/m2 in the regression analyses. As nearly 80% of women were using folate before and during pregnancy, folate deficiency37 could not explain the increased risk observed. Seizures prior to pregnancy did not seem to affect the risks, as women with nonactive epilepsy also had an increased risk of pre-eclampsia. The dosage of AEDs did not seem to affect the risk, as high and low doses were equally represented in women with pre-eclampsia. In a register-based study from Sweden,17 only women using AEDs were included, and an increased risk of pre-eclampsia was observed. This suggests that exposure to AEDs during pregnancy increases the risk of pre-eclampsia. A recent study has claimed that carbamazepine, in particular, is associated with pre-eclampsia.31 Our study observed an increased risk with lamotrigine use. Recent studies have demonstrated an association between high serum folate and a reduced risk of pre-eclampsia.37,38 In our study, 80% of women with epilepsy and AED use were using folate, but the amount taken may not have been optimal and not as high as recommended.39 A recent study has demonstrated an increased risk of pre-eclampsia in women using folic acid antagonists, such as carbamazepine and phenytoin, as well as dihydrofolate-reductase inhibitors, such as lamotrigine.40 The mechanism could be a maternal folate–homocysteine metabolic defect starting before and continuing throughout the pregnancy, resulting in placental microvascular disease.40

A five-fold increased risk of early vaginal bleeding among women with epilepsy using AEDs was found in this study, in contrast with some studies,5,15 but in accordance with others.13,41 Conflicting results may be caused by methodological limitations with recruitment bias and small sample sizes, but could also be a result of changes to new medications giving better seizure prevention with less side-effects. AED-induced folate deficiency and alteration in the metabolism of vitamin K-dependent blood clotting factors have been suggested as causes of vaginal bleeding in late pregnancy.42 In our study, 80% of women with epilepsy and early vaginal bleeding were using folate. Hence, this is an unlikely explanation. A recent study has explored the risk of miscarriage with the use of AEDs.16 This study observed a higher risk of miscarriages with the use of valproate and lamotrigine, and may support our high rates of early bleeding in pregnancy with lamotrigine use.

The induction rate in women with epilepsy was increased, mainly in those using AEDs. This is in accordance with some previous studies,6,15,23 but contrary to the findings in others, where no difference was seen.18,43 According to guidelines,1,44 epilepsy is not an indication for induction in uncomplicated pregnancies. Pre-eclampsia explained some of the increased induction rate in the present study. The European Registry of Antiepileptic Drugs and Pregnancy (EURAP) showed that 58.3% of women with epilepsy had no seizures during pregnancy, 15.9% had a decrease in seizures and 17.3% had an increase in seizures.45 In our study, 37% of women with epilepsy and induction had seizures during pregnancy. However, in less than one-sixth of these women was epilepsy per se used as the reason for the induction. High seizure activity may still have influenced the decision to start induction in women with epilepsy and mild pregnancy complications.

In our study, the rate of caesarean section in women with epilepsy and AED use was markedly increased in accordance with three population-based cohort studies.6,17,23 The reason why this is not seen worldwide is probably related to the very high caesarean section rate in some healthy obstetrical populations.5,15,16 Most chronic disorders increase the likelihood of caesarean section46 and, in our study, 12% of women were delivered by caesarean section as a result of epilepsy. Epilepsy represents a significant disorder, but is not an indication for caesarean section unless a seizure occurs during labour or the patient cannot co-operate during a vaginal delivery.42 The risk of seizures increases around the time of delivery.47 However, <2% of women with epilepsy experienced a seizure during labour.48 In our study, three women (1.5%) experienced seizures during labour; two of these were delivered by caesarean section. Our study demonstrated a high rate of induction in women with epilepsy, and 10% had a caesarean section with failed induction as an indication. Reducing the induction rate may lower the rate of failed induction and prolonged labour, and thereby reduce the rate of acute caesarean section in women with epilepsy.

Our study demonstrated an increased risk of major malformations in the newborn of women with epilepsy, with the highest risk in women with epilepsy and AED use. The rate (3.9%) was higher than in a recent Norwegian study,31 but lower than in a recent meta-analysis, estimating the incidence of malformations in women with epilepsy and AED use to be 7.08% (95% CI, 5.62–8.54).49 Polytherapy gave an even higher incidence. Low selection bias and a low rate of polytherapy may explain the lower malformation rate in our cohort.

Conclusions and implications for clinicians

Our study has demonstrated that women with epilepsy and AED use have an increased risk of pre-eclampsia, bleeding in pregnancy, induction, caesarean section and major malformations of the newborn. AEDs seem to play a role in the development of these complications. Active epilepsy with seizures during the last 5 years versus nonactive epilepsy did not discriminate for any of these complications. A change to newer medications, such as lamotrigine, may have increased the complication rates in the mother. Further studies, including genetic aspects and placental circulation, are needed to characterise epilepsy and AED-related pre-eclampsia.

Disclosure of interest

There are no conflicts of interest.

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. All authors approved the final version of the article.

Details of ethics approval

This study received consent from the local ethical research committee, The Norwegian Social Science Data Services, The Norwegian Directorate of Health and The Data Inspectorate of Norway. The study also received informed consent from all the women included in the study.

Funding

This study was supported by the Norwegian Research Council through the NevroNor research programme.

Acknowledgement

We thank Professor Trond Riise for important statistical advice.

Journal club

Discussion points

1. Background: Describe the antenatal management of women with epilepsy in your area and compare with published reviews.1,2 Are women routinely referred to a neurologist?2

How is this different for women on antiepileptic drugs?

2. Methods: Evaluate the involvement of a single researcher in data abstractions.

The authors have matched cases and controls for parity. Describe other analytical methods for adjusting for parity when comparing adverse pregnancy outcomes.

Discuss the variables used for adjusted multivariate analysis when comparing women with epilepsy with controls (Table 1). Would you have adjusted for any other factors? Discuss the possible interactions between the ‘delivery outcome’ variables.

3. Results and implications: Describe the possible reasons for induction of labour, and its influence on birth mode and outcomes, in women with epilepsy.

Discuss how increased early pregnancy bleeding in this study, and early pregnancy deaths in a confidential enquiry,2 would affect your management of women with epilepsy in the first trimester or pre-pregnancy (Data S1).

D Siassakos
University of Bristol and Southmead Hospital, Bristol, UK
Email jsiasakos@gmail.com

Ancillary