eTable 1. ICD-10 Codes for Maternal, Neonatal, and Childhood Diseases
eTable 2. Incidence Rates and Hazard Ratios of Epilepsy in Children According to Maternal Characteristics
eTable 3. Incidence Rates and Hazard Ratios of Epilepsy According to Maternal and Neonatal Complications
eTable 4. Early Pregnancy Body-Mass Index and Rates of Epilepsy in Children Without Cerebral Palsy
Customize your JAMA Network experience by selecting one or more topics from the list below.
Razaz N, Tedroff K, Villamor E, Cnattingius S. Maternal Body Mass Index in Early Pregnancy and Risk of Epilepsy in Offspring. JAMA Neurol. 2017;74(6):668–676. doi:10.1001/jamaneurol.2016.6130
What is the association between body mass index during early pregnancy and the risk of childhood epilepsy?
In this population-based cohort study that included 1.4 million live single births, we found that maternal overweight and obesity are associated with increased risks of childhood epilepsy. Greater risks were noted with increasing body mass index.
Given that overweight and obesity are modifiable, prevention of obesity during early pregnancy may be an important public health strategy to reduce the incidence of childhood epilepsy.
There is growing concern about the long-term neurologic effects of prenatal exposure to maternal overweight and obesity. The causes of epilepsy are poorly understood and, in more than 60% of the patients, no definitive cause can be determined.
To investigate the association between early pregnancy body mass index (BMI) and the risk of childhood epilepsy and examine associations between obesity-related pregnancy and neonatal complications and risks of childhood epilepsy.
Design, Setting, and Participants
A population-based cohort study of 1 441 623 live single births at 22 or more completed gestational weeks in Sweden from January 1, 1997, to December 31, 2011, was conducted. The diagnosis of epilepsy as well as obesity-related pregnancy and neonatal complications were based on information from the Sweden Medical Birth Register and National Patient Register. Multivariate Cox proportional hazards regression models were used to estimate adjusted hazard ratios (HRs) and 95% CIs after adjusting for maternal age, country of origin, educational level, cohabitation with partner, height, smoking, maternal epilepsy, and year of delivery. Data analysis was conducted from June 1 to December 15, 2016.
Main Outcomes and Measures
Risk of childhood epilepsy.
Of the 1 421 551 children born between January 1, 1997, and December 31, 2011, with covariate information available, 7592 (0.5%) were diagnosed with epilepsy through December 31, 2012. Of these 3530 (46.5%) were female. The overall incidence of epilepsy in children aged 28 days to 16 years was 6.79 per 10 000 child-years. Compared with offspring of normal-weight mothers (BMI 18.5 to <25.0), adjusted HRs of epilepsy by maternal BMI categories were as follows: overweight (BMI 25.0 to <30.0), 1.11 (95% CI, 1.04-1.17); obesity grade I (BMI 30.0 to <35.0), 1.20 (95% CI, 1.10-1.31); obesity grade II (BMI 35.0 to <40.0), 1.30 (95% CI, 1.12-1.50); and obesity grade III (BMI≥40.0), 1.82 (95% CI, 1.46-2.26). The rates of epilepsy were considerably increased for children with malformations of the nervous system (adjusted HR, 46.4; 95% CI, 42.2-51.0), hypoxic ischemic encephalopathy (adjusted HR, 23.6; 95% CI, 20.6-27.1), and neonatal convulsions (adjusted HR, 33.5; 95% CI, 30.1-37.4). The rates of epilepsy were doubled among children with neonatal hypoglycemia (adjusted HR, 2.10; 95% CI, 1.90-2.33) and respiratory distress syndrome (adjusted HR, 2.43; 2.21-2.66), and neonatal jaundice was associated with more than a 50% increased risk of epilepsy (adjusted HR, 1.47; 95% CI, 1.33-1.63). The elevated risk of epilepsy in children of overweight or obese mothers was not explained by obesity-related pregnancy or neonatal complications.
Conclusions and Relevance
The rates of childhood epilepsy increased with maternal overweight or obesity in a dose-response manner. Given that overweight and obesity are modifiable, prevention of obesity may be an important public health strategy to reduce the incidence of childhood epilepsy.
Epilepsy is one of the most common neurologic disorders of childhood.1 Currently, 50 million persons are affected by epilepsy worldwide.2,3 The cause of epilepsy is poorly understood and no definitive cause can be determined in more than 60% of the cases.4,5 Preterm birth, low birth weight for gestational age, chromosomal and congenital anomalies, neonatal convulsions, and low Apgar scores are generally associated with increased risks of childhood epilepsy.5-8
Maternal overweight and obesity have increased globally during the past 3 decades,9,10 and the prevalence of severe obesity, commonly defined as obesity classes II to III (body mass index [BMI]≥35.0 [calculated as weight in kilograms divided by height in meters squared]), has rapidly increased.11 There is growing concern about the long-term neurologic effects of children exposed to maternal obesity. Quiz Ref IDObesity in pregnancy has been linked to systemic inflammation, altered endocrine responses, folic acid deficiency, and insulin resistance.12 Risks of severe birth asphyxia, birth injuries, and congenital anomalies increase with maternal obesity,13-16 which in turn may influence long-term neurologic and intellectual development.17 A few studies have found increased risks of intellectual disability, attention-deficit/hyperactivity disorder, autism, and cerebral palsy in children whose mothers were overweight or obese during pregnancy.17,18 Only 1 previous study examined the association between maternal BMI and risk of epilepsy in the offspring, and no association was found.19
We investigated the association between early pregnancy BMI and risk of childhood epilepsy in a nationwide cohort study including more than 1.4 million live singleton births in Sweden. We also examined associations between obesity-related pregnancy and neonatal complications and the risks of childhood epilepsy.
This retrospective, nationwide cohort study included all live singleton births at 22 or more completed gestational weeks in Sweden from January 1, 1997, through December 31, 2011. With the use of the person-unique national registration numbers of mothers and their offspring,20 information on births included in the Swedish Medical Birth Register was cross-linked with the National Patient Register, Cause of Death Register, Total Population Register, and Education Register. The Medical Birth Register includes data on more than 98% of all births in Sweden.21 This population-based, clinically focused database contains information on antenatal, obstetrical, and neonatal care that is prospectively recorded on standardized forms.21 The National Patient Register contains diagnostic codes on hospital inpatient care since 1987 and hospital outpatient care from 2001.22,23 The Cause of Death Register includes information on all deaths in Sweden.24 Diagnoses and causes of death since 1997 are coded using the Swedish version of the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10). Maternal educational level and country of origin were obtained from the Education Register and the Total Population Register.20,25 The study was approved by the Research Ethics Committee at Karolinska Institutet, Stockholm, Sweden. This study was based on encrypted data, for which the ethics committees do not require informed consent.
Maternal BMI in early pregnancy was calculated from self-reported height and weight measured in light indoor clothing without shoes at the first antenatal visit, which occurs within the first 14 weeks of gestation (ie, first trimester) for 90% of pregnant women in Sweden.21 We computed the median height across pregnancies for multiparous women to reduce measurement error.26 Maternal BMI was categorized according to the World Health Organization as underweight (BMI≤18.4), normal weight (18.5 to <25.0), overweight (25.0 to <30.0), obesity grade I (30.0 to <35.0), obesity grade II (35.0 to <40.0), and obesity grade III (≥40.0).27
Maternal characteristics included age at delivery, country of origin, educational level, cohabitation with a partner, parity, height, smoking during pregnancy, maternal epilepsy, and year of delivery. Maternal age at delivery was calculated as the date of delivery minus the mother’s birth date. Parity was defined as the number of births of each mother. Information on cohabitation with the mother’s partner was obtained at the first antenatal visit. Women who reported daily smoking at the first antenatal visit and/or at 30 to 32 gestational weeks were classified as smokers, whereas women who stated that they were nonsmokers were classified as nonsmokers. Maternal epilepsy before the child’s birth was defined using the same algorithm that was used for children (eTable 1 in the Supplement). Pregnancy complications included hypertension (pregestational and preeclampsia), diabetes (gestational and mellitus), and infection-related disorders.
Information about the newborn included birth weight, gestational age, sex, congenital malformations, birth trauma, neonatal asphyxia-related complications, and more common neonatal complications. Categorizations of the covariates are provided in Table 1 and eTable 2 in the Supplement; specific ICD-10 codes for diseases in the mothers and children are provided in eTable 1 in the Supplement.
In Sweden, all women are offered ultrasonography scanning at 18 gestational weeks or earlier, and gestational age (in completed weeks) was estimated using the following hierarchy: the date of early second trimester ultrasonography (87.7%), the date of the last menstrual period (7.4%), or a postnatal assessment (4.9%).
In the study population of 1 441 623 live singleton births, we excluded 2334 (0.2%) children who died before age 28 days and 17 738 (1.2%) children with missing maternal and infant personal registration numbers. We also excluded 168 645 children (11.7%) for whom information on maternal BMI was missing. There were 1 421 551 births with complete information on covariates.
The definition of epilepsy in children was based on the 2011 recommendations of the International League Against Epilepsy’s Commission on Epidemiology, using a combination of hospital and outpatient diagnoses from the National Patient Registry.28,29 A case of epilepsy was identified if it met any of the following conditions between January 1, 1997, and December 31, 2012:
an occurrence of 2 or more ICD-10 codes G40 and/or G41 (epilepsy and status epilepticus, respectively) on separate dates or
an occurrence of 1 or more ICD-10 code R56 (convulsions not elsewhere classified) and 1 or more ICD-10 codes G40 and/or G41 on separate dates; diagnosis of R56 had to precede that of G40 and/or G41.
The date associated with the first record of epilepsy was considered the date of disease onset. Diagnosis of epilepsy was restricted to the period after the child’s 27th day from birth.
Incidence rates of epilepsy in children were quantified by dividing the number of children with epilepsy by the total follow-up time in child-years. We followed up each child from the 28th day of life until the date of the first diagnosis of epilepsy, emigration, death, or end of follow-up (December 31, 2012), whichever occurred first.
We compared crude rates between categories of maternal BMI and other maternal characteristics as well as pregnancy and neonatal complications for the whole cohort and for children born at term (≥37 weeks). To account for the varying lengths of follow-up, we calculated hazard ratios (HRs) and 95% CIs with the use of Cox proportional hazards models. Multivariate Cox proportional hazards regression was used to compare the rates of epilepsy between BMI categories and by levels of maternal and neonatal risk factors. We specified the robust sandwich estimate of the covariance matrix to account for the correlations of sequential births to the same mother in the study. Quiz Ref IDThe adjusted Cox proportional hazards model included maternal age, country of origin, educational level, cohabitation with partner, height, smoking, maternal epilepsy, and year of delivery. A linear trend in the association of BMI with the offspring’s epilepsy in the child was assessed using the Wald test by introducing a variable representing ordinal categories of BMI as a continuous predictor into the models. The percentage hazard difference per BMI unit was estimated as 100% × (HR − 1) from models with the BMI as a continuous variable. Interaction analyses were performed to investigate whether the association between maternal BMI and rates of epilepsy was different in children born preterm (≤36) or term (≥37 weeks). The interaction term was tested with the use of the χ2 test.
We conducted additional analyses to ascertain potentially mediating effects of obesity-related pregnancy and/or neonatal complications by introducing them as covariates into the multivariate models. These analyses assumed a lack of interaction between maternal BMI and maternal and neonatal complications and no unmeasured confounders of the potential mediators and rates of epilepsy.30 Regression model fit was assessed using the likelihood ratio test, and a 2-sided P value <.05 was used to determine statistical significance. All analyses were carried out with the use of SAS, version 9.4 (SAS Institute). Data analysis was conducted from June 1 to December 15, 2016.
Quiz Ref IDOf 1 421 551 children born between 1997 and 2011, 7592 children (0.5%) were diagnosed with epilepsy through December 31, 2012. A total of 3530 (46.5%) of the children were female. The overall incidence of epilepsy was 6.79 per 10 000 child-years. The rates and HRs of epilepsy were slightly higher among offspring of mothers who were born outside the Nordic countries, had lower educational levels, lived without a partner, were multiparous (≥4 births), or were smokers (eTable 2 in the Supplement).
A maternal diagnosis of epilepsy was associated with a more than 4-fold increase in epilepsy rates in the offspring (Table 1). Pregnancy complications (gestational diabetes, preeclampsia, infections, and chorioamnionitis), cesarean delivery, and birth trauma were also associated with increased rates of epilepsy. Most children who developed epilepsy were born at term, but the rates of epilepsy increased by decreasing gestational age, ranging from a 1.7-fold increased rate in moderately preterm infants (32-36 weeks) to an almost 5-fold increased rate in extremely preterm (22-27 weeks) infants.
The overrepresentation of congenital malformations among children with epilepsy was mainly attributed to malformations of the central nervous system (Table 1). Quiz Ref IDNeonatal convulsions, hypoxic ischemic encephalopathy, birth asphyxia, and low Apgar scores (0-3 or 4-6) at 5 minutes were associated with considerably increased epilepsy rates. Rates of epilepsy were also increased among children with neonatal infections, hypoglycemia, respiratory distress syndrome, and jaundice. Restricting the analyses to children born at term did not change our results (eTable 3 in the Supplement).
The adjusted rate of epilepsy increased linearly by maternal BMI (Table 2) from 6.30 per 10 000 child-years among normal-weight women to 12.4 per 10 000 child-years among women with obesity grade III. Quiz Ref IDThe rates of epilepsy increased by 11% in children of overweight mothers compared with children of normal-weight mothers; obesity grades I and II were associated with 20% and 30% increased rates, respectively; and obesity grade III was associated with an 82% increase in epilepsy (Table 2). In term births, the rates of epilepsy increased consistently with maternal overweight and obesity severity; only obesity grade III was associated with epilepsy in preterm births (test for interaction P = .89) (Table 3).
In the mediation analyses of BMI and epilepsy, neither obesity-related pregnancy nor neonatal complications substantially changed the association between maternal BMI and rates of epilepsy (Table 4, adjusted models 2 and 3). When both neonatal and obesity-related pregnancy complications were added to the multivariate model, the association between maternal BMI and rates of epilepsy was attenuated, but not eliminated (Table 4, adjusted model 4). The associations also persisted after excluding children with cerebral palsy (eTable 4 in the Supplement).
In this large, population-based study, the rate of childhood epilepsy increased with maternal overweight and obesity severity in a dose-response manner. We also found strong evidence that several asphyxia-related neonatal complications are associated with substantially increased rates of epilepsy and that the less severe but more prevalent neonatal complications, including neonatal jaundice, hypoglycemia, and respiratory distress, could potentially increase the rates of childhood epilepsy. However, the elevated rates of epilepsy in offspring of overweight or obese mothers could not be explained by obesity-related pregnancy or neonatal complications.
Overall, we found a steady increase in the rate of childhood epilepsy across the entire range of maternal BMI. We are aware of only 1 previous study investigating this issue, and no association was found between maternal obesity and the risk of childhood epilepsy.19 Reliance on a non–population-based data source (ie, Medicaid records) and use of a strict case ascertainment for diagnosis of epilepsy may have induced an underestimation of the risk of childhood epilepsy and contributed to the lack of association.19 In our population-based study, the definition of epilepsy was based on international recommendations.28,29
Few studies have reported associations between prepregnancy overweight and obesity and other long-term neurodevelopmental outcomes of the children, such as intellectual disability, lower cognitive performance, attention-deficit/hyperactivity disorder, cerebral palsy, and autism.17,18,31 Given the importance and sensitivity of fetal brain development and its susceptibility to hormonal and inflammatory reactions, even at the early embryonic stage,1,6,32 it is possible that maternal overweight and obesity increase the risk of brain injury, leading to a range of neurodevelopmental disorders. A meta-analysis reported a 2-fold increased risk of major malformations in the central nervous system, specifically neural tube defects, in the children of obese mothers.16 This finding lends support to the hypothesis that epilepsy may result from the gradual accumulation of environmental insults to the central nervous system.4 Another potential mechanism by which maternal obesity might affect neurodevelopment could be through obesity-induced inflammation. There is increasing evidence that maternal inflammatory conditions are associated with fetal systemic inflammation, which increases the risk of neonatal brain injury33 leading to outcomes such as cerebral palsy and epilepsy.6,34 In our study, fetal exposure to maternal infections and chorioamnionitis were associated with overall increased risks of epilepsy, but not when the analysis was restricted to term births. Maternal obesity during pregnancy, regardless of infections, is also associated with endothelial dysfunction and inflammatory upregulation, which may have a lasting effect on fetal neurodevelopment.32,35 Evidence also indicates that high levels of leptin that are observed in obesity are associated with placental dysfunction36 potentially disrupting normal neurodevelopment in the children.
In addition, maternal overweight and obesity are related to increased risks of pregnancy-associated diseases and neonatal complications13-15 that may be contributing to the increased risk of epilepsy in children. A Swedish population-based study has shown that maternal obesity is associated with increased risks of severe asphyxia-related complications in term infants.15 This increased risk may be attributed partly to shoulder dystocia37 and otherwise traumatic labor due to fetal macrosomia.38 However, our findings did not change substantially after adjusting for obesity-related pregnancy and neonatal complications, suggesting that maternal overweight or obesity per se may play a critical role in fetal neurodevelopment. Other underlying issues, including unmeasured genetic, lifestyle, and environmental factors that interact with obesity, could also contribute to this association.
We found strong evidence that asphyxia-related neonatal complications other than neonatal convulsions or low Apgar scores, including meconium aspiration and hypoxic ischemic encephalopathy, are associated with markedly increased rates of epilepsy. Previous studies have shown that children with a history of hypoxic ischemic encephalopathy had 5 times the risk of developing epilepsy.39 Furthermore, our study showed that neonatal jaundice, hypoglycemia, and respiratory distress may increase the risk of epilepsy. These findings persisted when analyses were restricted to children born at term.
Few previous studies have explored the potential role of more-prevalent neonatal complications on the risk of epilepsy in children.6,40 Moderate hypoglycemia for 3 or more days is associated with impaired neurologic development at 18 months.41 Respiratory distress has been linked to changes in cerebral blood flow, which in turn can increase the risks of periventricular or intraventricular hemorrhage in the developing brain.42 Even moderate degrees of hyperbilirubinemia have been associated with an increase in the risk of minor neurologic dysfunction throughout the first year of life.43
Moreover, we found higher rates of childhood epilepsy among children whose mothers had epilepsy, gestational diabetes, and preeclampsia. Other factors associated with higher rates of childhood epilepsy were preterm birth, low birth weight for gestational age, birth trauma, birth asphyxia, and congenital malformations. Chromosomal abnormalities, malformation of the nervous system, and neonatal convulsions were all associated with substantially increased risks of childhood epilepsy. These findings are in line with those of previous studies demonstrating robust associations between congenital anomalies and neonatal convulsions and the development of epilepsy.5,6 Convulsions in newborns with perinatal asphyxia are independently associated with early life brain injury and later neurodevelopmental impairment.44
The strengths of our study include the use of a large, comprehensive, nationwide population-based data source, which allowed us to study the rates of childhood epilepsy in relatively small groups, such as offspring of mothers with BMI of 40 or more and in children with specific neonatal complications. We used a previously validated case definition for epilepsy.28,29 In addition, prospective data collection minimized the risks of selection and recall bias.
Even though we relied on both hospital inpatient and outpatient registries to capture maternal and neonatal complications, misclassification and underreporting of some diagnoses, such as maternal epilepsy, maternal infections, and chorioamnionitis, should be considered. Any misclassification would have been nondifferential, which generally results in an underestimation of rates. In addition, the cause of epilepsy may be multidimensional, with interaction between genetic and environmental factors. Environmental factors, including maternal smoking, alcohol use, socioeconomic factors, receipt of antiepileptic drugs, and deficiency of vitamins, could potentially increase the risk of epilepsy in offspring.45 Although we controlled for a broad range of potential confounders, unmeasured and residual confounding may have occurred due to factors not available in our data sources.
This population-based study demonstrates that the risk of epilepsy in offspring increased with maternal BMI category in a dose-response pattern. Furthermore, our study provides strong evidence that asphyxia-related neonatal complications, as well as less severe neonatal complications, including neonatal hypoglycemia, jaundice, and respiratory distress, independently increase the risk of childhood epilepsy. Given that overweight and obesity are potentially modifiable risk factors, prevention of obesity in women of reproductive age may be an important public health strategy to reduce the incidence of epilepsy.
Accepted for Publication: December 15, 2016.
Corresponding Author: Neda Razaz, PhD, Clinical Epidemiology Unit, Department of Medicine Solna, Karolinska University Hospital, Karolinska Institutet, SE-171 76 Stockholm, Sweden (email@example.com).
Published Online: April 3, 2017. doi:10.1001/jamaneurol.2016.6130
Author Contributions: Drs Razaz and Cnattingius had full access to all of the data in the study and take full responsibility for the integrity of the data and the accuracy of the data analyses. Dr Razaz is the guarantor.
Concept and design: All authors.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Razaz, Tedroff.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Razaz.
Obtained funding: Cnattingius.
Administrative, technical, or material support: Cnattingius.
Supervision: Razaz, Cnattingius.
Conflict of Interest Disclosures: None reported.
Funding/Support: The study was supported by grant 2014-0073 from the Swedish Research Council for Health, Working Life and Welfare and an unrestricted grant from Karolinska Institutet (Distinguished Professor Award to Professor Cnattingius).
Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Create a personal account or sign in to: