Pregnancy, delivery, and outcome for the child in maternal epilepsy

Transcription

1 FULL-LENGTH ORIGINAL RESEARCH Pregnancy, delivery, and outcome for the child in maternal epilepsy *zgyri Veiby, yxanne K. Daltveit, *zbernt A. Engelsen, and *znils E. Gilhus *Department of Clinical Medicine, Section for Neurology, University of Bergen, Norway; ydepartment of Public Health and Primary Health Care, University of Bergen, Norway; zdepartment of Neurology, Haukeland University Hospital, Bergen, Norway; and xmedical Birth Registry of Norway, Division of Epidemiology, Norwegian Institute of Public Health, Bergen, Norway SUMMARY Purpose: To investigate pregnancy, delivery, and child outcome in an unselected population of women with both treated and untreated epilepsy. Methods: In the compulsory Medical Birth Registry of Norway, all 2,861 deliveries by women with epilepsy recorded from were compared to all 369,267 nonepilepsy deliveries in the same period. Results: The majority (66%, n = 1900) in the epilepsy group did not use antiepileptic drugs (AEDs) during pregnancy. A total of 961 epilepsy-pregnancies were exposed to AEDs. Compared to nonepilepsy controls, AED-exposed infants were more often preterm (p = 0.01), and more often had birth weight <2,500 g (), head circumference <2.5 percentile (), and low Apgar score (p = 0.03). Small-for-gestational-age (SGA) infants (<10 percentile) occurred more frequently in both AED-exposed (p = 0.05) and unexposed (p = 0.02) epilepsy-pregnancies. Frequency of major congenital malformations (MCMs) was 2.8% (n = 81) in the epilepsy group versus 2.5% in controls (p = 0.3). Increased risk for MCMs could be demonstrated only for exposure to valproate (5.6%, p = 0.005) and AED polytherapy (6.1%, p = 0.02). Neonatal spina bifida was not significantly increased, but was a major indication for elective pregnancy termination among women with epilepsy. Cesarean section was performed more often in maternal epilepsy, regardless of AED-exposure (). Discussion: Adverse pregnancy and birth outcome in women with epilepsy is mainly confined to AEDexposed pregnancies, although some risks are associated also with untreated epilepsy. The risk for congenital malformations was lower than previously reported. This could be due to a shift in AED selection, folic acid supplement, or possibly reflect the true risks in an unselected epilepsy population. KEY WORDS: Epilepsy, Pregnancy, Antiepileptic drugs, Neonatal outcome, Congenital malformations. Accepted March 10, 2009; Early View publication June 1, Address correspondence to Dr Gyri Veiby, Department of Neurology, Haukeland University Hospital, 5021 Bergen, Norway. gyri. veiby@hotmail.com Wiley Periodicals, Inc. ª 2009 International League Against Epilepsy Epilepsy is the most common maternal neurologic disorder requiring medical treatment during pregnancy (Pennell, 2006). Risks associated with medical treatment during pregnancy must be weighed against the risk for fetal or maternal complications due to epileptic seizures (Kalviainen & Tomson, 2006; Meador et al., 2006). Increased frequency of pregnancy complications (preeclampsia, vaginal bleeding) as well as adverse perinatal outcome (low birth weight, prematurity, mortality) and development delay has been reported (Crawford, 2001; Adab et al., 2004; Pennell, 2006), although results have been conflicting (Hiilesmaa et al., 1985; Viinikainen et al., 2006). Teratogenicity of antiepileptic drugs (AEDs) has been extensively debated. Infants of mothers with epilepsy have a reported two to three fold higher risk for congenital malformations, mainly associated with AED treatment (Holmes et al., 2001; Perucca, 2005). Intrauterine exposure to valproate or to multiple AEDs seems to represent the highest malformation risk (Perucca, 2005; Breen & Davenport, 2006; Pennell, 2006). Dose-dependency has 2130

2 2131 Child Outcome in Maternal Epilepsy been demonstrated for valproate (Artama et al., 2005; Vajda & Eadie, 2005; Meador et al., 2006) and lamotrigine (Morrow et al., 2006). Some AEDs are associated with specific malformations, that is, spina bifida and cleft lip-palate (Samren et al., 1999; Artama et al., 2005). Risks associated with pregnancy and AEDs will change over the years parallel to a shift regarding preferred drugs. Changes in guidelines for folic acid supplementation during pregnancy could also be important. Reports on malformation rates are based primarily on retrospective or prospective observational studies in pregnancy registries with an acceptable cohort size that permits systematic data collection. However, such studies often have several methodologic limitations that is, insufficient sample size, recruitment and assessment bias, retrospective design, limited follow-up length for malformation identification, inadequate definition of major congenital malformations (MCMs), questionable control group, not accounting for elective/spontaneous abortions and drawing final conclusions regarding causality and effect have been difficult (Perucca, 2005; Pennell, 2006). The population-based Medical Birth Registry of Norway (MBRN) offers a unique opportunity to examine a complete epilepsy cohort including both women with and without AED treatment. The compulsory, prospective registration of all deliveries, obtained from a population of 4.5 million people during a 7-year period, ensures a large cohort without selection bias. In addition, the material represents a population with homogenous health services during both pregnancy and delivery, ensuring comparable notification of all the study subjects and minimizing the effect of socioeconomic factors. The aim of this study was to investigate and evaluate pregnancy and birth outcome in an unselected and timerepresentative epilepsy population, including women with both treated and untreated epilepsy. Materials and methods Patient selection Patients were identified through MBRN, a national registry established in 1967 based on the compulsory notification of all births at 12 or more weeks of gestation. MBRN is placed under the Norwegian Institute of Public Health. Throughout the pregnancy, medical information on the mother, including compulsory information on drug exposures, is routinely ascertained by the attending physician and midwife. Such data, as well as information on delivery and on the newborn, is collected in a standardized notification form at delivery and sent to MBRN within 9 days after discharge from the hospital. Since 1999, all medication during pregnancy has been reported. Additional information on congenital malformations diagnosed within the first year of life is obtained by MBRN from all Norwegian pediatric wards. The notification form is coded at MBRN, partly according to the International Classification of Diseases, 10th revision (ICD-10), and partly according to a classification system developed by MBRN after consensus among obstetricians, neonatologists, and epidemiologists. Information on maternal epilepsy, including both treated and untreated epilepsy, AEDs, and folic acid supplementation is compulsory. The MBRN is notified of all mdeications according to the Anatomical Therapeutic Chemical (ATC) Classification System, consisting of a 5-digit ATC-number. Complete ascertainment of all births is ensured through a record linkage to the National Population Registry of Norway. Our study included all births recorded from December 1, 1998 through 2005 as registered per October 6, 2005 in MBRN. An unchanged notification form was used during this time-period. The epilepsy group included all 2,861 deliveries by mothers with epilepsy diagnosed prior to or during pregnancy by the attending physician. The control group consisted of all 369,267 other deliveries during the same time-period. Supplementary data on medically induced abortions performed after the 12th week of gestation were obtained from a separate database at MBRN. Frequency of fetal malformations and chromosomal disorders in such pregnancies are presented in the results, but could not be included in the analysis of the epilepsy cohort. The database did not contain information on AED exposures for such pregnancies. Variables Demographic data included maternal age (years), smoking during pregnancy (yes/no), maternal educational level (0 9 years, years, 13 or more years), and the child s birth order (1st, 2nd, 3rd, or more). All outcome variables were selected a priori to minimize potential effects of multiple testing on the results. Pregnancies exposed to AEDs (yes/no) were identified by including all anticonvulsants with a registered ACT-number. Supplement of folic acid and multivitamins ( 4 times/week) was registered prior to pregnancy (within 1 month to conception) and during pregnancy (within 1st trimester). The following birth outcomes were selected: Birth weight (g), gestational age (completed weeks), low birth weight (<2,500 g), prematurity (<37 weeks), small-forgestational-age (SGA) (<10th percentile), and small head circumference (<2.5 percentile) based on the Norwegian population (Skjaerven et al., 2000), low Apgar score (<7 at 5 min), neonatal transfers (to neonatal care unit at pediatric department), neonatal intracranial hemorrhage/ hematoma (subependymal, intraventricular, intracerebral, subdural, or epidural), anemia (hemoglobin <9 g/dl), hypoglycemia (b-glucose <2 mm), neonatal seizures, respiratory distress, stillbirth (fetal death 16 weeks of gestation), perinatal mortality, and congenital malformations.

3 2132 G. Veiby et al. Perinatal mortality included all deaths in live-born children 16 weeks of gestation occurring within the 6 first days, and all stillbirths 28 weeks of gestation (unknown gestational age: birth weight >1,000 g and/or birth length >35 cm). Congenital malformations included malformations diagnosed within the first year of life, and were classified into any congenital malformation (including minor) and MCMs. MBRN s definition of MCM includes congenital malformations either causing significant functional impairment or leading to operative intervention or both. The definition also includes chromosomal disorders. We further classified congenital malformations according to ICD-10 main categories. Selected specific malformations in the epilepsy group were compared to controls. Pregnancy and delivery complications included: Gestational hypertension [pregnancy-induced blood pressure (BP) 140/90 after 20 weeks of gestation], preeclampsia [gestational hypertension and proteinuria, including HELLP (hemolytic anemia, elevated liver enzymes, and low platelet count) syndrome], vaginal bleeding during pregnancy (1st, 2nd, or 3rd trimester), mean placenta weight (g), placental abruption, placenta previa, placental disorders (according to ICD-10), thromboembolic episodes (any episode during pregnancy/postpartum within 7 days leading to medical intervention), vaginal bleeding >500 ml during delivery, hypotonic uterine activity, cesarean section (emergency or elective), and maternal death. Information on epileptic seizures during pregnancy and delivery is not compulsory in MBRN, and is, therefore, better documented in hospital records. Such information was not included in this material. Statistics The analyses were based on crude and adjusted measures, and performed by SPSS 15.0 for Windows (SPSS Inc., Chicago, IL, U.S.A.). The epilepsy group was stratified into untreated and AED-treated epilepsy. AED treatment within the epilepsy group was further classified into carbamazepine, lamotrigine, valproate, and polytherapy. Outcome in all epilepsy subgroups was compared to the control group of all nonepilepsy births. Two-sided p-values <0.05 were considered statistically significant. Arithmetic mean was calculated for birth weight, gestational age, birth order, maternal age, and placenta weight, and analyzed by Independent samples t test, comparing the epilepsy and control group. Dichotomized characteristics of the epilepsy group compared to nonepilepsy controls were analyzed by Pearson s chi-square test (smoking, educational level, dietary supplementation). Cross-tabulated measures for small samples with expected cell count <5 were analyzed by Fisher s exact test. Such data are presented as unadjusted p-values with corresponding odds ratios (ORs) (specified in the text and tables for each unadjusted result). All other dichotomized outcomes were analyzed by unconditional logistic regression with potential confounders represented as categorical variables, and presented as adjusted ORs with corresponding 95% confidence intervals (CIs) and p-values. Adjusted relative risk (RR) was calculated by generalized linear models in SPSS for cesarean section, postnatal vaginal bleeding, and neonatal transfer (three most frequent outcomes in the epilepsy cohort), as well as for gestational hypertension, preeclampsia, anemia, 3rd trimester bleeding, and congenital malformations (all outcomes in Table 4). The RRs for these outcomes did not differ significantly from the presented ORs. Maternal age (<25, 25 29, 30 34, and 35+ years), child s birth order (first, second, third, or more), and smoking during pregnancy (yes/no) were considered potential confounders. Adjustment for these potential confounders had very little effect. The results were also adjusted for educational level. This did not have significant impact on the results. Because cognitive function in women with epilepsy can be causally related to the disease, educational level might represent an intermediate variable and was not included as a confounder in the final analyses. Results Mothers with epilepsy During the recording period, 0.8% of all births were delivered by women with epilepsy. Characteristics of the epilepsy group compared to nonepilepsy controls are presented in Table 1. High educational level ( 13 years) was especially less frequent among women treated with valproate (26%) or lamotrigine (29%). Antiepileptic drug treatment The majority of the women in the epilepsy group did not use AEDs during pregnancy (66%, n = 1,900). A total of 961 pregnancies in the epilepsy group were exposed to AEDs, mainly as monotherapy (86%). The most commonly used AEDs were carbamazepine (46%, n = 439), lamotrigine (25%, n = 237), and valproate (22%, n = 215). Other frequently used AEDs during pregnancy were oxcarbazepine (n = 41), clonazepam (n = 36), topiramate (n = 26), phenytoin (n = 27), phenobarbital (n = 27), levetiracetam (n = 18), gabapentin (n = 18), and vigabatrin (n = 9). Folic acid and multivitamin supplementation Dietary supplementation is presented in Table 1. Women exposed to AEDs used more folic acid supplementation prior to conception (32%), compared to both the control group (10%) and women with untreated epilepsy (13%). Folic acid supplementation during pregnancy was also most common in AED exposure (66%), especially for AED polytherapy (75%) and valproate treatment (72%). The frequency of MCMs and any congenital malformations (ACMs) as a function of folate

4 2133 Child Outcome in Maternal Epilepsy Table 1. Characteristics of the epilepsy group compared to all nonepilepsy controls Characteristic Controls Epilepsy p-value a Odds ratio 95% CI Mean maternal age, years (SD) 29.3 (5.0) 28.9 (5.2) <0.001 Mean parity (SD) 0.94 (1.1) 0.91 (1.1) 0.2 High educational level b (%) < Smoking (%) < Preconceptual folate (%) < st trimester folate (%) < Preconceptual multivitamins (%) st trimester multivitamins (%) < CI, confidence interval; SD, standard deviation. a Means are analyzed by Independent samples t-test and unadjusted bivariate associations by Pearson s chi-square test. b 13 or more years. use were analyzed within the AED-treated subgroups (carbamazepine, valproate, and lamotrigine). For all groups there was a tendency toward a higher frequency of MCMs and ACMs in pregnancies with folate supplementation, both prior to and during pregnancy. In the valproate subgroup, the higher frequency of ACMs in pregnancies with preconceptual folate supplementation was significant (13.2%, n = 9 vs. 4.8%, n = 7, p = 0.049, OR = 2.9, 95% CI = ). Folic acid supplementation varied according to maternal educational level. In epilepsy pregnancies, 36% of women with 0 9 years education reported 1st trimester folic acid supplementation compared to 43% with years education and 48% with 13 years education. Birth outcome Infants in the epilepsy group had a lower mean birth weight compared to controls (3,464 g vs. 3,521 g, ), whereas mean gestational age was the same (39 weeks for both). Low birth weight (<2,500 g), small head circumference (<2.5 percentile), and prematurity (<37 weeks) were significantly increased in AEDexposed infants, and especially in patients with polytherapy or with carbamazepine treatment (Table 2). Children in the epilepsy group were more often transferred to a pediatric ward during the neonatal period, especially AEDexposed infants (Table 2). Two infants in the epilepsy group were diagnosed with neonatal intracranial hemorrhage (Fisher s exact test p = 0.1, unadjusted OR = 3.8, 95% CI = ), none of them exposed to AEDs. During both these deliveries fetal respiratory distress was reported, and cesarean section was performed. Seizures during the neonatal period occurred more frequently among infants in the epilepsy group (0.2%, n = 6 vs. 0.1%, Fisher s exact test p = 0.004, unadjusted OR = 4.1, 95% CI = ), regardless of AED exposure. The frequency of hypoglycemia, anemia, and respiratory distress syndrome was not increased in the neonates of the epilepsy group. Perinatal mortality (0.6%, n = 16 vs. 0.5%, p = 0.7, OR = 1.1, 95% CI = ) and stillbirth-rate (Table 2) were similar to the controls. Congenital malformations Frequencies and risks for MCMs and ACMs diagnosed within the first year of life are shown in Tables 3 and 4. The epilepsy group had risks similar to those of the nonepilepsy controls, both for ACMs and MCMs. However, infants exposed to either valproate or polytherapy had significantly higher malformation rates (Tables 3 and 4). The malformation rates in the epilepsy group were identical between AED-exposed and unexposed infants when valproate pregnancies were excluded, both for ACMs (4.7% vs. 4.7%) and MCMs (2.7% vs. 2.6%). Frequencies of congenital malformations within main categories are listed in Table 5. Cardiovascular malformations were significantly more common in AED-exposed infants versus controls (1.9%, n = 18 vs. 1.1%, p = 0.02, OR = 1.8, 95% CI = ), and especially for valproate exposure (3.7%, n = 8 vs. 1.1%, Fisher s exact test p = 0.002, unadjusted OR = 3.6, 95% CI = ). In the untreated epilepsy group there was a higher occurrence of genital malformations (1.2%, n = 22 vs. 0.6%,, OR = 2.1, 95% CI = ). There were two children with spina bifida in the epilepsy group (0.07% vs. 0.03% in controls, Fisher s exact test p = 0.3, unadjusted OR = 2.1, 95% CI = ), one exposed to valproate. The rate of cleft lip/palate was similar between AED-treated and untreated epilepsy pregnancies, and also compared to nonepilepsy controls (Table 5). Infants with Down syndrome were significantly more common in untreated epilepsy pregnancies versus controls (0.3%, n = 6 vs. 0.1%, Fisher s exact test p = 0.04, unadjusted OR = 2.5, 95% CI = ), and not due to difference in maternal age. In lamotrigine monotherapy, a total of five malformations were reported: persistent ductus arteriosus (n = 2), hernia diaphragmatica (n = 1), facial deformity (n = 1), and cardiac ventricular septal defect (n = 1).

5 2134 G. Veiby et al. Table 2. Adverse birth outcome in epilepsy subgroups compared to the control group of all nonepilepsy births Adverse birth outcome Controls N = 369,267 % No AED N = 1,900 AED total N = 961 Polytherapy N = 132 VPA N = 215 LTG N = 237 Stillbirth (20) 1.0 (10) 0.8 (1) 0.5 (1) 0.4 (1) 1.6 (7) OR = 1.3 CI = p = 0.3 OR = 1.3 CI = p = 0.5 OR = 0.9 a CI = p = 1.0 OR = 0.6 a CI = p = 1.0 OR = 0.5 a CI = p = 0.7 CBZ N = 439 OR = 1.9 a CI = p = 0.1 Apgar score <7 (at 5 min) (32) 2.7 (25) 3.1 (4) 1.9 (4) 3.4 (8) 2.8 (12) OR = 1.1 CI = p = 0.6 OR = 1.6 CI = p = 0.03 OR = 1.8 a CI = p = 0.3 OR = 1.1 a CI = p = 0.8 OR = 2.0 a CI = p = 0.07 OR = 1.7 CI = p = 0.08 Low birth weight (<2,500 g) (115) 9.6 (91) 15.2 (20) 5.2 (11) 9.7 (23) 12.3 (53) OR = 1.2 CI = p = 0.1 OR = 1.7 CI = OR = 2.8 CI = OR = 1.0 CI = p = 0.9 OR = 1.8 CI = p = OR = 2.3 CI = Small for gestational age (<10 percentile) (181) 10.0 (96) 17.4 (23) 7.0 (15) 9.6 (23) 11.4 (50) OR = 1.2 CI = p = 0.02 OR = 1.2 CI = p = 0.05 OR = 2.3 CI = OR = 0.8 CI = p = 0.5 OR = 1.2 CI = p = 0.5 OR = 1.5 CI = p = 0.01 Small head circumference (<2.5 percentile) (39) 4.1 (39) 6.9 (9) 2.8 (6) 2.1 (5) 5.3 (23) OR = 1.0 CI = p = 1.0 OR = 2.0 CI = OR = 3.6 a CI = p = OR = 1.4 a CI = p = 0.3 OR = 1.0 a CI = p = 0.8 OR = 2.7 CI = Prematurity (<37 weeks) (153) 10.7 (103) 17.6 (23) 8.1 (17) 8.9 (21) 14.8 (63) OR = 1.1 CI = p = 0.4 OR = 1.3 CI = p = 0.01 OR = 2.3 CI = OR = 1.0 CI = p = 0.9 OR = 1.1 CI = p = 0.6 OR = 1.8 CI = Neonatal transfers (pediatric ward) (216) 15.4 (148) 22.0 (29) 15.3 (33) 14.5 (35) 17.5 (77) OR = 1.2 CI = p = OR = 1.7 CI = OR = 2.6 CI = OR = 1.8 CI = p = OR = 1.6 CI = p = OR = 2.0 CI = AED, antiepileptic drug; CBZ, carbamazepine; LTG, lamotrigine; OR, odds ratio; VPA, valproate. Results are analyzed by logistic regression, adjusted for maternal age, parity, and smoking, and presented as p-values with corresponding ORs and 95% confidence intervals (CIs). a Small sample with expected cell count <5: Fisher s exact test p-value with unadjusted OR and 95% CI. Additional information on medically induced abortions after the 12th pregnancy week demonstrated 14 pregnancy terminations due to congenital malformations and/or chromosomal disorders among women with epilepsy during the study period. The majority were performed in association with spina bifida (36%, n = 5 vs. 4.5% in control group) and Down syndrome (29%, n = 4 vs. 7.4% in control group). Two of the cases with spina bifida had additional malformations: (1) Arnold-Chiari syndrome and (2) anencephaly, anal atresia, polycystic kidney, and reduction defect of radius. The five remaining abortions were performed due to congenital hydrocephalus (n = 1), thanatophoric short stature (n = 1), Edwards syndrome (n = 1), and unspecified major malformation (n = 2). Information on AED use in such pregnancies could not be obtained. Pregnancy and delivery complications Women in the epilepsy group had a higher frequency of preeclampsia, primarily in pregnancies exposed to AEDs (Table 6), and especially for carbamazepine (8.9%, n = 39 vs. 4.3% in nonepilepsy controls,, OR = 2.2, 95% CI = ). Pregnancies with AEDs had a higher occurrence of third trimester vaginal bleeding (Table 6), mainly for carbamazepine (1.8%) and lamotrigine (1.7%) treatment. Postnatal vaginal bleeding >500 ml (Table 6) and hypotonic uterine activity (5.9%, n = 57 vs. 4.3%, p = 0.009, OR = 1.4, 95% CI = ) were more common during delivery in the AED-exposed group than in the controls. Risk for postnatal vaginal bleeding was highest in association with exposure to oxcarbazepine (29.5%) and valproate (25.1%). Placenta weight, as well as occurrence of placenta previa, placental abruption, and placental disorders, was similar between the epilepsy and control groups, regardless of AED exposure. Thromboembolic episodes during pregnancy and postpartum were not increased in the epilepsy group. There was a higher rate of cesarean section in the epilepsy group, especially in AED-exposed pregnancies (Table 6), where 11.4% were delivered by emergency

6 2135 Child Outcome in Maternal Epilepsy Table 3. Major congenital malformations (MCMs) in epilepsy subgroups compared to the control group of all nonepilepsy births Group MCM, n (%) p-value a Odds ratio 95% CI Controls 9309 (2.5) Epilepsy group 81 (2.8) AED total 32 (3.3) No AED 49 (2.6) Polytherapy 8 (6.1) 0.02 c Polytherapy b 4 (4.5) 0.3 c VPA 12 (5.6) CBZ 12 (2.5) LTG 8 (3.2) AED, antiepileptic drug; CBZ, carbamazepine; CI, confidence interval; LTG, lamotrigine; OR, odds ratio; VPA, valproate. a Results are analyzed by logistic regression; adjusted for maternal age, birth order and smoking. b VPA treatment excluded. c Small sample with expected cell count <5: Fisher s exact test p-value with unadjusted OR and 95% CI. section compared to 8.9% in the control group. There were no maternal deaths during delivery in the epilepsy group. Discussion Our study demonstrated a low malformation risk of only 4.9% for ACMs and 2.8% for MCMs in maternal epilepsy, similar to the rates in the general population. However, women with epilepsy experienced more pregnancy complications and adverse perinatal outcomes. This was mainly confined to pregnancies associated with AED treatment, although cesarean section and SGA infants were also associated with untreated epilepsy. Occurrence of genital malformations and Down syndrome were specifically increased in the untreated epilepsy group. Interactive effects between genetic factors, socioeconomic factors, maternal disease effect (seizures, underlying etiology), and exposure to AEDs probably explain the higher rate of adverse outcomes in maternal epilepsy. The overall risk for congenital malformations was not significantly increased in the epilepsy group, and the malformation rates in different AED exposures were low compared to most earlier reports (Artama et al., 2005; Perucca, 2005; Vajda & Eadie, 2005; Wyszynski et al., 2005; Morrow et al., 2006). Valproate- and polytherapytreated patients were the only groups with an increased risk. In addition, previous studies have shown the highest risks in such pregnancies (Rosa, 1991; Samren et al., 1999; Wide et al., 2004; Morrow et al., 2006). Infants of women with untreated epilepsy had malformation rates very similar to the control group, and this was also true in AED-exposed pregnancies when valproate-treated patients were excluded. Lamotrigine exposure had malformation rates similar to the control group, in line with Table 4. Any congenital malformations (ACMs) in epilepsy subgroups compared to the control group of all nonepilepsy births Group ACM n (%) p-value a Odds ratio 95% CI Controls 15,564 (4.2) Epilepsy group 141 (4.9) AED total 51 (5.3) No AED 90 (4.7) Polytherapy 11 (8.3) Polytherapy b 6 (6.7) 0.3 c VPA 16 (7.4) CBZ 25 (5.7) LTG 9 (3.8) AED, antiepileptic drug; CBZ, carbamazepine; CI, confidence interval; LTG, lamotrigine; OR, odds ratio; VPA, valproate. a Results are analyzed by logistic regression; adjusted for maternal age, birth order, and smoking. b VPA treatment excluded. c Small sample with expected cell count <5: Fisher s exact test p-value with unadjusted OR and 95% CI. previous reports (Sabers et al., 2004; Cunnington & Tennis, 2005; Vajda et al., 2006). We report the malformation rate in an unselected and complete epilepsy population, in contrast to most previous studies, which have reported risks in hospital-based populations or registries with voluntary recruitment. Our cohort included also the women with uncomplicated and inactive epilepsy, and, therefore, at lower risk. The AED-exposed group was treated with a wide range of different AEDs, but the majority of these patients used carbamazepine, valproate, or lamotrigine, and mainly as monotherapy. A shift toward a higher rate of monotherapy and a change toward a less teratogenic AED selection may have contributed to the low malformation risk. The effect of epileptic seizures could not be addressed. A similar study on the Norwegian cohort from 1973 reported a malformation rate twice as high in women with epilepsy compared to the control group (Bjerkedal & Bahna, 1973). Earlier scientific reports have shown that folic acid reduces malformations in the general population, but this has not been demonstrated for AED-exposed pregnancies. In the present study, folic acid supplementation was higher in AED-treated pregnancies, especially when treated with valproate or polytherapy. Such pregnancies most likely also have a higher malformation risk. Confounding by indication due to potential treatment-selection bias, therefore, made it difficult to assess possible protective effects of folic acid. Folic acid supplementation in highrisk pregnancies may have added to the low malformation rate in our study. However, more than 60% of AEDexposed women did not use folic acid prior to conception, in line with another report on folic acid supplementation during pregnancy in Norway (Nilsen et al., 2006). To prevent neural tube defects, folic acid must be initiated

7 2136 G. Veiby et al. Table 5. Congenital malformations within first year of life in epilepsy subgroups and the control group of all nonepilepsy births. Main categories based on ICD-10 Congenital malformation Controls N = % No AED N = 1900 AED total N = 961 Polytherapy N = 132 VPA N = 215 CBZ N = 439 LTG N = 237 Cleft lip/palate (1) 0.1 (1) (1) 0 0 Nervous system (1) 0.2 (2) (1) 0.2 (1) 0 Face/neck/eye/ear (6) 0.1 (1) 0.8 (1) (1) Cardiovascular system (19) 1.9 (18) 3.0 (4) 3.7 (8) 1.4 (6) 2.1 (5) Respiratory system (1) 0.1 (1) (1) 0 Alimentary tract (6) 0.4 (4) (2) 0.5 (2) 0 Genital organs (22) 0.3 (3) (1) 0.5 (2) 0 Urinary system (2) 0.2 (2) 0.8 (1) (2) 0 Musculoskeletal system (31) 1.8 (17) 3.0 (4) 1.9 (4) 2.5 (11) 1.3 (3) Chromosomal defect (6) 0.1 (1) 0.8 (1) Other (3) 0.1 (1) AED, Antiepileptic drug; CBZ, carbamazepine; LTG, lamotrigine; VPA, valproate. Table 6. Complications during pregnancy and delivery in epilepsy subgroups and the control group of all nonepilepsy births Complication Controls N = % Epilepsy N = 2,861 Gestational hypertension (59) OR = 1.1 CI = p = 0.4 Preeclampsia (160) OR = 1.3 CI = p = Anemia (Hb <9) (15) OR = 1.1 CI = p = 0.8 3rd trimester vaginal bleeding (28) OR = 1.3 CI = p = 0.1 Cesarean section (562) OR = 1.4 CI = Vaginal bleeding >500 ml a (495) OR = 1.3 CI = AED Epilepsy N = (27) OR = 1.5 CI = p = (65) OR = 1.6 CI = (3) OR = 0.6 b CI = p = (14) OR = 2.0 CI = p = (207) OR = 1.6 CI = (193) OR = 1.5 CI = No AED Epilepsy N = 1, (32) OR = 0.9 CI = p = (95) OR = 1.2 CI = p = (12) OR = 1.3 CI = p = (14) OR = 1.0 CI = p = (355) OR = 1.4 CI = (302) OR = 1.1 CI = p = 0.04 AED, antiepileptic drug treatment; CI, confidence interval, Hb, hemoglobin; OR, odds ratio. All results are analyzed by logistic regression, adjusted for maternal age, parity, and smoking, and presented as p-values with corresponding ORs and 95% CIs. a During delivery b Small sample with expected cell number <5: Fisher s exact test p-value with unadjusted OR and 95% CI. periconceptionally (Czeizel & Dudas, 1992; Botto et al., 2005; Kampman, 2007), and additional efforts to improve folic acid supplementation in women with epilepsy should be undertaken. Only two newborn infants with spina bifida were identified in the epilepsy cohort, but the majority of elective pregnancy terminations among women with epilepsy were performed because of this condition. If spina bifida is diagnosed primarily by prenatal screening procedures, and the majority aborted, it is essential that such information is accounted for when malformations are investigated in pregnancy registries. Statistical power to demonstrate risk for specific malformations can be difficult to obtain due to small sample size (Pennell, 2006). The risk for cleft lip/palate malformations was not increased in the epilepsy group, regardless of AED exposure. In particular, there were no cleft lip/palate malformations in pregnancies exposed to lamotrigine (Dolk et al., 2008; Holmes et al., 2008). Down

8 2137 Child Outcome in Maternal Epilepsy syndrome was a common cause for elective abortions in epilepsy pregnancies, and was found at a significantly increased rate in deliveries by women with untreated epilepsy. Infants with untreated mothers also had significantly more genital malformations. To our knowledge, a predisposition for such chromosomal and genital defects has not earlier been reported in women with epilepsy. Advanced maternal age is by far the most important identified risk factor associated with Down syndrome (Gaulden, 1992). However, additional environmental and genetic risk factors are involved. Toxic effects such as exposure to AEDs, ovarian dysfunction linked to hormonal imbalance, and genetic alterations related to low folate status, may predispose women with epilepsy to chromosomal defects (Sherman et al., 2007). Such predisposition could also be associated with epileptic syndrome or etiology, seizure characteristics, or other genetic vulnerability. There was a higher frequency of SGA infants in both AED-exposed and unexposed epilepsy pregnancies, indicating a suboptimal intrauterine growth environment in both groups (Pallotto & Kilbride, 2006). However, occurrence of small head circumference, low birth weight, and prematurity were most strongly related to use of AEDs, suggesting an interaction between maternal disease and drug treatment. Diagnosed placental disorders were not increased in the epilepsy group. Earlier reports indicate both growth restriction and shorter pregnancy length in maternal epilepsy (Bjerkedal & Bahna, 1973; Yerby et al., 1985; Mastroiacovo et al., 1988; Meador et al., 2006). Underlying causes have not been identified, and AED effects on placenta function have not been shown (Eeg-Olofsson et al., 1990). The mechanisms affecting fetal growth and malformation risk most probably are different, as valproate treatment had the highest malformation risk, but was similar to the control group regarding risk for all other adverse perinatal outcomes. In contrast, lamotrigine and carbamazepine were safe regarding risk for fetal malformations, but were associated with an increased frequency of other adverse birth outcomes. Infants in the epilepsy group had a higher frequency of low Apgar score and were more often transferred to pediatric wards after delivery, especially in AED-treated women. This, as well as the higher rate of emergency cesarean section, indicates more fetal stress during delivery. Low Apgar score, growth restriction, and low birth weight are indicators for childhood and adult morbidity (Moster et al., 2002; O Keeffe et al., 2003; Jelliffe- Pawlowski & Hansen, 2004; Pallotto & Kilbride, 2006; Vrachnis et al., 2006). Attention toward such neonatal markers and possible long-term effects for children born to mothers with epilepsy, is an important challenge. Perinatal mortality and stillbirths were not increased in the epilepsy group, regardless of AED exposure. Some earlier reports have suggested higher mortality in infants born to mothers with epilepsy (Bjerkedal & Bahna, 1973; Hiilesmaa et al., 1985; Meador et al., 2006). Two infants were diagnosed with neonatal intracranial hemorrhage in the epilepsy group, both in pregnancies without AED treatment. Inadequate maternal seizure control has been associated with this serious complication (Sherer et al., 1998; LaJoie & Moshe, 2004). Because MBRN contains incomplete information on epileptic seizures during pregnancy and delivery, effects of maternal seizures on pregnancy and birth outcome could not be investigated. AEDs were associated with more vaginal bleeding during late pregnancy and delivery, in line with previous studies (Bjerkedal & Bahna, 1973; Pennell, 2004; Pilo et al., 2006), and may partly be explained by the reported higher rate of hypotonic uterine activity during labor. Alterations in vitamin K metabolism may also be a causal factor (Crawford, 2001; Pilo et al., 2006), possibly associated with use of enzyme-inducing AEDs (Kaaja et al., 2002). Exposure to AEDs was also associated with an increased risk for preeclampsia (Bjerkedal & Bahna, 1973; Yerby et al., 1985; Pilo et al., 2006). Whether this risk is caused by use of AEDs, or is related to genetic or socioeconomic factors, remains unclear. This registry-based study has the major strength of including all deliveries in Norway (Irgens, 2002). Notification of maternal epilepsy is based on compulsory information from the attending physician and midwife, and includes treated as well as untreated epilepsy. The reported maternal epilepsy in 0.8% of deliveries indicates good diagnostic sensitivity. Earlier reports on MBRN have also demonstrated high diagnostic validity regarding maternal diseases (Skomsvoll et al., 2002; Hoff et al., 2007). Because a large proportion of the epilepsy group was untreated, potential underreporting of AED treatment is a concern. However, risk for adverse effects of AEDs is well known to Norwegian healthcare workers, and information on both epilepsy and use of AEDs is compulsory in the registry. A recent study in which data from MBRN were linked to the Norwegian Prescription Database, reports a similar proportion of AED treatment during pregnancy in women with epilepsy, strongly supporting the validity of our data (Engeland A, Bjørge T, Daltveit AK, Vollset SE, Furu K, unpublished manuscript). Folic acid supplementation in the untreated group was reported to be similar to controls, supporting the real absence of AEDs. Most likely, the large proportion of untreated epilepsy reflects an unselected population cohort and indicates that pregnancy is more common among women with untreated and probably less active epilepsy, representing the healthiest individuals within the epilepsy population. Distribution of maternal epilepsy within geographic regions in Norway demonstrated acceptable variation of treated versus untreated pregnancies and a consistently low proportion of AED treatment during pregnancy

9 2138 G. Veiby et al. (results not shown). Unfortunately, this study lacks information on socioeconomic status, and AED dose and blood levels. In addition, the lack of data on maternal head circumference might represent a limitation in the interpretation of child head circumference. In conclusion, the low malformation rate reported in this study may reflect a large group with untreated and less active epilepsy, a shift in AED selection, and folic acid supplementation, and represents the true risk in an unselected epilepsy population. However, a diagnosis of epilepsy still implies a significantly increased maternal risk during pregnancy and delivery, and for the child. Acknowledgments This work has been supported by the Norwegian Association for Epilepsy, and the NevroNor programme of the Norwegian Research Council. We confirm that we have read the Journal s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. Disclosure of conflicts of interest: Gyri Veiby: None. Anne Kjersti Daltveit: None. Bernt Engelsen: None. Nils Erik Gilhus: None. References Adab N, Kini U, Vinten J, Ayres J, Baker G, Clayton-Smith J, Coyle H, Fryer A, Gorry J, Gregg J, Mawer G, Nicolaides P, Pickering L, Tunnicliffe L, Chadwick DW. (2004) The longer term outcome of children born to mothers with epilepsy. J Neurol Neurosurg Psychiatry 75: Artama M, Auvinen A, Raudaskoski T, Isojarvi I, Isojarvi J. (2005) Antiepileptic drug use of women with epilepsy and congenital malformations in offspring. Neurology 64: Bjerkedal T, Bahna SL. (1973) The course and outcome of pregnancy in women with epilepsy. Acta Obstet Gynecol Scand 52: Botto LD, Lisi A, Robert-Gnansia E, Erickson JD, Vollset SE, Mastroiacovo P, Botting B, Cocchi G, de Vigan C, de Walle H, Feijoo M, Irgens LM, McDonnell B, Merlob P, Ritvanen A, Scarano G, Siffel C, Metneki J, Stoll C, Smithells R, Goujard J. (2005) International retrospective cohort study of neural tube defects in relation to folic acid recommendations: are the recommendations working? BMJ 330:571. Breen DP, Davenport RJ. (2006) Teratogenicity of antiepileptic drugs. BMJ 333: Crawford P. (2001) CPD-Education and self-assessment: epilepsy and pregnancy. Seizure 10: Cunnington M, Tennis P. (2005) Lamotrigine and the risk of malformations in pregnancy. Neurology 64: Czeizel AE, Dudas I. (1992) Prevention of the first occurrence of neuraltube defects by periconceptional vitamin supplementation. N Engl J Med 327: Dolk H, Jentink J, Loane M, Morris J, de Jong-van den Berg LT. (2008) Does lamotrigine use in pregnancy increase orofacial cleft risk relative to other malformations? Neurology 71: Eeg-Olofsson O, Chen MF, Andermann E, Dansky L, Guyda HJ, Kinch RA. (1990) Evaluation of placental morphology and growth factor receptors in women receiving antiepileptic drugs: a pilot study. Epilepsia 31: Gaulden ME. (1992) Maternal age effect: the enigma of down syndrome and other trisomic conditions. Mutat Res 296: Hiilesmaa VK, Bardy A, Teramo K. (1985) Obstetric outcome in women with epilepsy. Am J Obstet Gynecol 152: Hoff JM, Daltveit AK, Gilhus NE. (2007) Myasthenia gravis in pregnancy and birth: identifying risk factors, optimising care. Eur J Neurol 14: Holmes LB, Harvey EA, Coull BA, Huntington KB, Khoshbin S, Hayes AM, Ryan LM. (2001) The teratogenicity of anticonvulsant drugs. N Engl J Med 344: Holmes LB, Baldwin EJ, Smith CR, Habecker E, Glassman L, Wong SL, Wyszynski DF. (2008) Increased frequency of isolated cleft palate in infants exposed to lamotrigine during pregnancy. Neurology 70: Irgens LM. (2002) Medical birth registry an essential resource in perinatal medical research. Tidsskr Nor Laegeforen 122: Jelliffe-Pawlowski LL, Hansen RL. (2004) Neurodevelopmental outcome at 8 months and 4 years among infants born full-term smallfor-gestational-age. J Perinatol 24: Kaaja E, Kaaja R, Matila R, Hiilesmaa V. (2002) Enzyme-inducing antiepileptic drugs in pregnancy and the risk of bleeding in the neonate. Neurology 58: Kalviainen R, Tomson T. (2006) Optimizing treatment of epilepsy during pregnancy. Neurology 67:S59 S63. Kampman MT. (2007) Folate status in women of childbearing age with epilepsy. Epilepsy Res 75: LaJoie J, Moshe SL. (2004) Effects of seizures and their treatment on fetal brain. Epilepsia 45(Suppl 8): Mastroiacovo P, Bertollini R, Licata D. (1988) Fetal growth in the offspring of epileptic women: results of an Italian multicentric cohort study. Acta Neurol Scand 78: Meador KJ, Baker GA, Finnell RH, Kalayjian LA, Liporace JD, Loring DW, Mawer G, Pennell PB, Smith JC, Wolff MC. (2006) In utero antiepileptic drug exposure: fetal death and malformations. Neurology 67: Morrow J, Russell A, Guthrie E, Parsons L, Robertson I, Waddell R, Irwin B, McGivern RC, Morrison PJ, Craig J. (2006) Malformation risks of antiepileptic drugs in pregnancy: a prospective study from the UK Epilepsy and Pregnancy Register. J Neurol Neurosurg Psychiatry 77: Moster D, Lie RT, Markestad T. (2002) Joint association of Apgar scores and early neonatal symptoms with minor disabilities at school age. Arch Dis Child Fetal Neonatal Ed 86:F16 F21. Nilsen RM, Vollset SE, Gjessing HK, Magnus P, Meltzer HM, Haugen M, Ueland PM. (2006) Patterns and predictors of folic acid supplement use among pregnant women: the Norwegian Mother and Child Cohort Study. Am J Clin Nutr 84: O Keeffe MJ, O Callaghan M, Williams GM, Najman JM, Bor W. (2003) Learning, cognitive, and attentional problems in adolescents born small for gestational age. Pediatrics 112: Pallotto EK, Kilbride HW. (2006) Perinatal outcome and later implications of intrauterine growth restriction. Clin Obstet Gynecol 49: Pennell PB. (2004) Pregnancy in women who have epilepsy. Neurol Clin 22: Pennell PB. (2006) 2005 AES annual course: evidence used to treat women with epilepsy. Epilepsia 47(Suppl 1): Perucca E. (2005) Birth defects after prenatal exposure to antiepileptic drugs. Lancet Neurol 4: Pilo C, Wide K, Winbladh B. (2006) Pregnancy, delivery, and neonatal complications after treatment with antiepileptic drugs. Acta Obstet Gynecol Scand 85: Rosa FW. (1991) Spina bifida in infants of women treated with carbamazepine during pregnancy. N Engl J Med 324: Sabers A, Dam M, Rogvi-Hansen B, Boas J, Sidenius P, Laue FM, Alving J, Dahl M, Ankerhus J, Mouritzen DA. (2004) Epilepsy and pregnancy: lamotrigine as main drug used. Acta Neurol Scand 109:9 13. Samren EB, van Duijn CM, Christiaens GC, Hofman A, Lindhout D. (1999) Antiepileptic drug regimens and major congenital abnormalities in the offspring. Ann Neurol 46: Sherer DM, Anyaegbunam A, Onyeije C. (1998) Antepartum fetal intracranial hemorrhage, predisposing factors and prenatal sonography: a review. Am J Perinatol 15: Sherman SL, Allen EG, Bean LH, Freeman SB. (2007) Epidemiology of Down syndrome. Ment Retard Dev Disabil Res Rev 13: Skjaerven R, Gjessing HK, Bakketeig LS. (2000) Birthweight by gestational age in Norway. Acta Obstet Gynecol Scand 79: Skomsvoll J, Ostensen M, Baste V, Irgens L. (2002) Validity of a rheumatic disease diagnosis in the Medical Birth Registry of Norway. Acta Obstet Gynecol Scand 81:

10 2139 Child Outcome in Maternal Epilepsy Vajda FJ, Eadie MJ. (2005) Maternal valproate dosage and foetal malformations. Acta Neurol Scand 112: Vajda FJ, Hitchcock A, Graham J, Solinas C, O Brien TJ, Lander CM, Eadie MJ. (2006) Foetal malformations and seizure control: 52 months data of the Australian Pregnancy Registry. Eur J Neurol 13: Viinikainen K, Heinonen S, Eriksson K, Kalviainen R. (2006) Community-based, prospective, controlled study of obstetric and neonatal outcome of 179 pregnancies in women with epilepsy. Epilepsia 47: Vrachnis N, Botsis D, Iliodromiti Z. (2006) The fetus that is small for gestational age. Ann N Y Acad Sci 1092: Wide K, Winbladh B, Kallen B. (2004) Major malformations in infants exposed to antiepileptic drugs in utero, with emphasis on carbamazepine and valproic acid: a nation-wide, population-based register study. Acta Paediatr 93: Wyszynski DF, Nambisan M, Surve T, Alsdorf RM, Smith CR, Holmes LB. (2005) Increased rate of major malformations in offspring exposed to valproate during pregnancy. Neurology 64: Yerby M, Koepsell T, Daling J. (1985) Pregnancy complications and outcomes in a cohort of women with epilepsy. Epilepsia 26: