Malformation in index pregnancy in women with epilepsy is not followed by recurrence in subsequent pregnancy


  • Shehanaaz Begum,

    1. Kerala Registry of Epilepsy and Pregnancy, Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum, Kerala State, India
    Search for more papers by this author
  • Sankara P. Sarma,

    1. Kerala Registry of Epilepsy and Pregnancy, Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum, Kerala State, India
    Search for more papers by this author
  • Sanjeev V. Thomas

    Corresponding author
    1. Kerala Registry of Epilepsy and Pregnancy, Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum, Kerala State, India
    • Address correspondence to Sanjeev V. Thomas, Professor of Neurology, Department of Neurology, Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum 695011, India. E-mail:

    Search for more papers by this author


Use of antiepileptic drugs (AEDs) in pregnant women with epilepsy (WWE) is associated with an increased risk of major congenital malformations (MCM). Previous studies have suggested that WWE who had a malformation in their index pregnancy were at an increased risk of recurrence in future pregnancies. We aimed to assess the risk of recurrence of MCM in 1,616 WWE from Kerala Registry of Epilepsy and Pregnancy. The pregnancy outcome of women (n = 246) with two prospective pregnancies in the registry were analyzed. They had partial seizures with or without generalization (57.3%) or generalized seizures (42.7%). Polytherapy was used in 26.4% (index pregnancy) and 23.6% (follow-up pregnancy). The mean dosage of AED for valproate was 498 mg/day and carbamazepine was 555 mg/day. The malformation rate in the index pregnancy was 8.5% (21/246) and in the follow-up pregnancy was 8.9% (22/246) with only one recurrence. There was no increased risk of MCM in follow-up pregnancy for those who had MCM in the index pregnancy (p = 0.70; OR 0.49; 95% CI 0.06–3.80). The use of any specific drug, continuation of the same drug or a change in drug therapy between two pregnancies did not alter the recurrence risk.

The use of antiepileptic drugs (AEDs) is associated with risk of major congenital malformation (MCM) of about 5–9% in comparison to a background risk of 1–3% (Meador et al., 2008; Tomson et al., 2011). There is inadequate data on risk of recurrence of malformations to provide evidence based counseling to women with epilepsy (WWE) who are considering a second pregnancy. Data from the United Kingdom Epilepsy and Pregnancy Register (UKEPR) and the Australian Pregnancy Register (APR) indicate that the risk of malformation recurrence in subsequent pregnancy is increased if the index pregnancy had a malformation (Campbell et al., 2013; Vajda et al., 2013). There is little prospective data on the recurrence risk from other parts of the world. We aimed to estimate the risk of recurrence of MCM in follow-up pregnancy for the index pregnancies that had MCM from the data from the registry in Kerala, India.

Materials and Methods

Kerala Registry of Epilepsy and Pregnancy (KREP) is prospectively following up WWE from preconception period, through pregnancy and delivery until their children are 12 years of age. The main objective is to estimate the MCM risk in this cohort. We ascertain malformation status ante natally (detailed anomaly scan and serum alpha feto protein estimation) and postnatally (clinical examination at birth, echocardiography and abdomen ultrasonography at 3 months of age followed by further verifications up to 1 year of age) (Beghi & Annegers, 2001; Thomas et al., 2001; Tomson et al., 2011) We had 2,140 registrations between April 1998 and March 2013.

In this study, we considered all WWE who had two or more pregnancies registered in KREP. For each woman we defined index pregnancy as the first pregnancy in the registry and follow-up pregnancy as the next pregnancy for which she was registered again. Pre index pregnancies, if any were not included. All live births and terminations after anomaly scans (irrespective of malformation status) were included. Pregnancy loss (abortions, intrauterine deaths or stillbirths) without anomaly scans were excluded. All cases where major malformations were detected on antenatal ultrasonography (in the case of abortions or medical terminations), neonatal physical examination, and echocardiography or abdomen ultrasonography at 3 months of age were included as positive cases. Malformation rate is the number of positive cases expressed as a percentage of the total number of pregnancies where malformation status was ascertained.

The prescribed daily dose was recorded from the clinical records and compliance was ascertained through the daily medication- seizure – pregnancy calendar maintained by the patient. No drug levels were monitored. The highest daily dose any time in a month is taken as the representative daily dose for that month. The mean of the daily dose for the months of pregnancy is recorded as the mean daily dose for that pregnancy.

A MCM is defined as an abnormality that can interfere with the quality of life and warrants definite management. Asymptomatic cardiac malformations detected by echocardiography were rigorously defined as: atrial septal defect: a defect >5 mm size in the interatrial septum confirmed by Doppler echo at 3 months of age, patent ductus arteriosus: Persistent flow across the ductus at 3 months of age. Any defects in the ventricular septum or other complex malformations such as tetralogy of Fallot were included as MCMs. We excluded from MCM all patent foramen ovale (defect in interatrial septum <5 mm) and trivial valvular regurgitation or stenosis (detected only on Doppler echo and not demanding medical attention). Minor asymptomatic and unexplained dilatations of the pelvi-ureteric system of the kidneys detected only on ultrasonography were also excluded. Congenital malformation rates were further analyzed with respect to individual AED exposure and mono and polytherapy. Statistical significance of differences were ascertained by Fisher's exact t-test, McNemar test or chi-square test and odds ratios (ORs) with 95% confidence intervals (CIs).


Of the 1,616 WWE who had pregnancies under KREP, there were 634 repeat pregnancies. We excluded 142 pregnancies without malformation screening: spontaneous abortions prior to malformation screening (61), incomplete screening for malformation (13), medical termination of pregnancy prior to malformation screening (27), intrauterine deaths (9), stillbirths (4), lost to follow-up (4) or third pregnancy (24). Remaining 492 pregnancies in 246 women were included in this analysis. Their mean age (SD) at time of index pregnancy was 24.1 (3.7) years. The mean interval between two pregnancies was 1,233 ± 693 days. The seizures were classified as generalized seizures in 105 (42.7%) and partial with or without generalization in 141 (57.3%). Four WWE had inter district change of residence between two consecutive pregnancies. The AED usage and seizure control in index and follow-up pregnancies of WWE are given in Table 1.

Table 1. Antiepileptic drug usage and malformations in index and follow-up pregnancy
AEDIndex pregnancyFollow-up pregnancy
N (%)Mean dosage (SD) mg/dayProportion (%) with seizuresN (%)Mean dosage (SD) mg/dayProportion (%) with seizures
Any seizure typeGTCSAny seizure typeGTCS
  1. a

    Includes 7 who were off AEDs in the index pregnancy also.

  2. b

    Cardiac: ASD, PDA, TOF.

  3. c

    Nervous system: neural tube defect, hydrocephalus.

  4. d

    Skeletal: Polydactyly, Congenital Talipes Equino Varus.

  5. e

    One case with NTD also had CTEV in follow-up pregnancy.

All cases246 54.122246 36.215.4
1171 47.920.5163 34.415.9
2 or more65 78.429.258 56.920.7
No AED10 0025a 00
Phenobarbitone5483 (44)59.327.74781 (37)51.119.1
Lamotrigine6185 (105)66.7509170 (113)44.422.2
Primidone3425 (153)10033.33333 (288)33.30
Clonazepam71.2 (0.7)71.428.551.3 (0.7)6040
Phenytoin46193 (98)69.523.933251 (84)57.624.2
Clobazam207.1 (6)95351812 (6.7)66.611.1
Carbamazepine108555 (280)6116.694552 (308)41.513.8
Valproate64498 (311)42.131.366472 (277)25.815.2
Topiramate0   275 (35)1000
Oxcarbazepine5897 (258)804051,023 (377)8080
Cardiacb7 (33.3)   8 (36.4)   
Nervous systemc2 (9.5)   4 (18)d   
Skeletale5 (23.8)   0   
Genito urinary4 (19)   5 (22.7)   
Gastro intestinal1 (4.8)   4 (18)   
Multiple system2 (9.5)   1 (4.5)   

The overall control of convulsive seizures during pregnancy was good (22% in index pregnancy and 15.4% in follow-up pregnancy); the proportion of women who were totally seizure free was 45.9% in index pregnancy and 63.8% in follow-up pregnancy (See Table 1). AED treatment was unchanged between pregnancies for 166 WWE (67.5%) while it was changed for 73 (29.7%) (See Table 2). The malformation rate in the index pregnancy was 21/246 (8.5%; 95% CI 5.7–12.7) and in the follow-up pregnancy was 22/246 (8.9%; 95% CI 6.0–13.2). These rates were comparable to that for the group who had only one pregnancy in the registry (7.2%; 95% CI 5.9–8.8). MCM had occurred for 42 of the 246 women in index or follow-up pregnancy of which only one had MCM in both pregnancies. There was no increased risk of MCM in follow-up pregnancy for those who had MCM in the index pregnancy (See Table 2).

Table 2. Malformation (MCM) rate for the follow-up pregnancy according to the malformation status in the index pregnancy and AED usage
Index pregnancy malformation statusFollow-up pregnancyORCIp
MCM-Yes (%)MCM-No
All index pregnancies (246)     
MCM-Yes (21)1200.490.06–3.800.703
MCM-No (225)21204
VPA used (64)     
MCM – Yes (7)070.900.819–0.9781.000
MCM – No (57)651
Other AEDs (182)     
MCM – Yes (14)1130.790.10–6.421.000
MCM – No (168)15153
AED changed (73)     
MCM – Yes (9)090.890.82–0.970.300
MCM– No (64)7 (10.9)54
AED unchanged (166)     
MCM-Yes (12)1110.910.11–7.570.930
MCM-No (154)14 (9.1)140
Valproate used in both pregnancies (48)     
MCM-Yes (4)040.860.77–0.971.000
MCM-No (44)6 (13.6)38

There were 166 women (67.5%) who had same treatment in the index and follow-up pregnancy. AED treatment was changed for 73 patients (29.7%). Seven women were off AEDs in both pregnancies. The risk of recurrence of MCM was not different for those who had a change versus those who did not have a change in the AED therapy between index and follow-up pregnancy. There was no difference in the recurrence risk, for any given AED exposure.

There was no difference in the risk of malformation recurrence between those exposed to valproate (VPA) and those exposed to AEDs other than VPA in the index pregnancy (See Table 2). Out of the 64 WWE on VPA in index pregnancy, 16 stopped it before follow-up pregnancy. Among the 48 who were on VPA in index and follow-up pregnancy, MCM occurred in four index and six follow-up pregnancies but none of them were recurrence for the same person (See Table 2).


The background of a pregnancy registry had given us an opportunity to prospectively estimate the risk of recurrence of malformations in consecutive pregnancies for WWE. We did not observe any excess risk of MCM in follow-up pregnancy if the index pregnancy had a MCM. In this study of 492 pregnancies in 246 women, 43 pregnancies had MCM (42 women). Only one woman had recurrence of MCM in the follow-up pregnancy. The MCM rates (8.5% in index pregnancy and 8.9% in follow-up pregnancy) were comparable to the rate of MCM in other pregnancy registries (Campbell et al., 2013; Vajda et al., 2013). The use of any specific drug, continuation of the same drug or a change in drug therapy between two pregnancies did not alter the recurrence risk.

Our results differ from that from the registries in United Kingdom and Australia, where researchers had demonstrated an increased risk of recurrence of MCM in follow-up pregnancy, if the index pregnancy had MCM (Campbell et al., 2013; Vajda et al., 2013). All three studies had similar proportion of repeat registrations (20–23%) within the registry and the spectrum of malformations is comparable. The maternal age, seizure types and AED usage are comparable between these registries. There are differences between these studies in terms of population (Caucasian vs. Asian), methodology (United Kingdom registry had excluded women with family history of malformations and pregnancy loss without malformations), selection criteria (Australian study had included retrospective cases also) and AED exposure (higher dosage and more of newer AEDs in United Kingdom and Australia).

The malformation risk for valproate was higher than that of other AEDs in this study, as observed elsewhere. Nevertheless, we did not observe any excess risk of recurrence of MCM associated with any AED exposure particularly valproate. It is possible that the low recurrence risk in our study is related to the low dosage of AED exposure in contrast to the dosage in the other registries. The mean dosage of VPA monotherapy in the United Kingdom registry was 948.1 mg/day for the normal group and 1,065 mg/day for the MCM group. The comparable values for carbamazepine were 632 and 682 mg/day in the United Kingdom registry. In the Australian registry, the mean daily dose of VPA was 882 mg/day (first pregnancy) and 803 mg/day (second pregnancy) for the normal group and 1,148 or 1,057 mg/day for the MCM group. The mean dose in KREP for valproate was 474 mg/day (index pregnancy), 452 mg/day (follow-up pregnancy) for normal outcome and 693 mg/day (index pregnancy) 665 mg/day (follow-up pregnancy) for the MCM group. These exposures are in the range of 50% (for normal outcome) to 65% (for the MCM group) that of United Kingdom registry. The risk of recurrence appears to be low in the setting of low dose of AEDs. The teratogenic impact of AEDs had been shown to be dose dependent (Tomson et al., 2011). Although the dosage of AEDs were on lower side in this study, there was good control of convulsive seizures (15–22%). The control of seizures with valproate was superior to other AEDs although the dosage is rather low (Table 1) (Thomas et al., 2012). The overall seizure control during pregnancy in this study is comparable to that reported from EURAP Study Group (2006) and Australia (Vajda et al., 2008). One should not directly extrapolate seizure control during pregnancy to seizure control in non-pregnant state as the treatment objectives and strategies in these situations are not identical. It appears that the recurrence risk for malformation is low when the AED exposure is at low dosage. Another reason could be the ethnic difference between populations as this study is exclusively on Asian Indians while the other two studies are mostly on Caucasian population.

Birth defects in a population have a polygenic etiology. About 10% of MCM is attributed to environmental factors and only <1% are related to chemicals, prescription drugs and high ionizing radiations (Brent, 2004). Population based studies from different geographical areas have pointed towards a genetic basis for malformation recurrence (Stoltenberg et al., 1999). Studies from Norway and Denmark have shown that malformation recurrence risk remained low, when women had moved to new geographical location (Lie et al., 1994) or when partners had changed (Christensen et al., 1995). Caution should be exercised while extrapolating from large epidemiological studies to higher risk groups such as WWE who constitute only 0.1–0.2% of the population. There was little geographical movement between pregnancies in this study. The absence of increased risk in this study is unlikely to be due to geographic movement or changes in partners. Drugs probably add to the polygenic mechanism and induce malformations if the fetus is exposed to a suprathreshold dosage at a critical point in organogenesis (Brent, 2004).

In this study, normal outcome in index pregnancy was not associated with reduced risk of MCM in follow-up pregnancy, even when the AEDs are same in both pregnancies. In our series, 14 of the 154 (9.1%) who had normal index pregnancy outcome had MCM in follow-up pregnancy, even while the AEDs were unchanged. Similar trends were observed with and without valproate (see Table 2).

Our data suggests that the malformation recurrence risk may be dose dependent and at low dose there may not be increased risk of recurrence. We believe that there is an opportunity to try low dose of AEDs in well controlled women planning pregnancy. An occasional myoclonus or mild partial seizure may be more acceptable to some women who are considering pregnancy than the potential risk of major malformation associated with standard doses of antiepileptic drugs. A limitation of this study is the rather small number of cases (n = 246) available to analyse the impact of various antiepileptic drugs on malformation recurrence. The comparable numbers for the United Kingdom registry and Australian registries were 646 and 228 respectively. The outcome of 1,006 follow-up pregnancies (503 with malformation in index pregnancy and 503 without malformation in index pregnancy) need to be ascertained in order to achieve a power (1 – beta) of 80%. Hence great caution should be applied while interpreting these data particularly with AED specific risk of recurrence of MCM. The population genetic contribution and socioeconomic and environmental factors also need to be addressed in future studies.


None of the authors have any conflicts of interest to disclose. 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.


  • Image of creator

    S. Begum is a Senior Research Fellow in Kerala Registry of Epilepsy and Pregnancy, Department of Neurology.