Abstract
Background
Severe maternal morbidity (SMM) is associated with adverse perinatal outcomes. However, its association with longer-term offspring outcomes, including neurodevelopment disorders, is poorly understood. This study aims to examine the association between SMM and the risk of epilepsy in offspring.
Methods
A retrospective birth cohort study of all singleton hospital births between 2002 and 2018 in Ontario, Canada was followed up until March 2020. SMM (20 weeks’ gestation to 42 days postpartum) and epilepsy diagnosed before 18 years of age in offspring were identified by using administrative health records. The risk of epilepsy associated with SMM was estimated by using Cox proportional hazard models and the robustness of the results was examined by using quantitative bias analyses (outcome and exposure misclassification and residual or unmeasured confounding).
Results
Of 2 060 668 children with a median follow-up time of 10 years, 40 830 (2%) were exposed to severe maternal morbidity and 16 992 (0.8%) were diagnosed with epilepsy (1.2% among those exposed to SMM and 0.8% among those unexposed). After adjustment for maternal, birth, and socioeconomic characteristics, children exposed to SMM had a 45% increased risk of epilepsy diagnosis compared with those unexposed [hazard ratio (HR) 1.45, 95% confidence interval (CI) 1.33–1.59]. This increase was similar across the subtypes of SMM. Estimates accounting for potential misclassification biases and residual or unmeasured confounding (by obesity or maternal education) showed increased risks (HR 1.91, 95% CI 1.50–2.42).
Conclusion
An increased risk of epilepsy was found in children exposed to SMM, suggesting a perinatal etiology and underlining the importance of ensuring a healthy pregnancy.
Keywords: severe maternal morbidity, perinatal etiology, neurodevelopmental disorder, epilepsy, pediatric, quantitative bias analysis
Key Messages.
This study sought to discover whether severe maternal morbidity (SMM) is associated with an increased risk of epilepsy in offspring.
It was found that children exposed to SMM had a 45% higher risk of developing epilepsy compared with unexposed children, with this increased risk being consistent across various subtypes of SMM and further elevated when potential misclassification biases were accounted for.
This is one of the first large-scale studies to have examined the long-term neurodevelopmental impact of SMM and it suggests that prenatal exposure to SMM may contribute to epilepsy risk, emphasizing the importance of maternal health in preventing childhood neurological disorders.
Introduction
Severe maternal morbidity (SMM) refers to a set of health-impacting and life-threatening conditions and procedures occurring throughout the pregnancy until 42 days after the delivery [1]. SMM is considered an indicator of maternal health and a quality indicator of obstetric care [2]. Although the lack of a standardized definition introduces variability [3–5], studies have estimated the prevalence of SMM at between 2% and 13% of pregnancies [6–8], depending on the country. In Canada, the prevalence was 16.1 [95% confidence interval (CI) 15.9–16.3] per 1000 deliveries for the period of 2012–2016 [8]. The prevalence of SMM has been increasing over time in North America and Europe [7, 9, 10].
SMM is associated with not only maternal mortality [11], but also several adverse perinatal outcomes, including infant death, neonatal intensive care unit admission, preterm birth, and low birthweight [2, 12–17]. However, its longer-term impact on children is poorly understood. Prenatal exposure to adverse environments affected by SMM could impact several aspects of fetal development, including brain development [18], which manifest as neurological disorders in childhood.
Epilepsy is one of the most common neurological disorders in children and approximately half of the cases have no identified causes [19]. In Ontario, Canada, the estimated prevalence is ∼8.0 per 1000 children, with higher rates among those born preterm or with perinatal complications [20]. Most etiological studies have examined the role of gestational age and birthweight, finding a greater risk of epilepsy in children born preterm or with low birthweight [21]. Fewer studies have investigated maternal factors, which may contribute through shared inflammatory, vascular, or hypoxic pathways and, to the knowledge of the current authors, none has focused on SMM [22, 23]. Some SMM conditions, such as antenatal hemorrhage, preeclampsia, and sepsis [24], are preventable and manageable with proper monitoring and treatment [25]. Thus, if SMM increases the risk of epilepsy in offspring, then its early management could be important for preventing the onset of epilepsy in childhood. The present study aims to investigate the association between SMM (and SMM subtypes) and epilepsy in children in a large population-based birth cohort of children born in Ontario, Canada.
Methods
Study design, setting, and population
A birth cohort linking 10 population datasets from Ontario, Canada was created by using unique encoded identifiers, analysed at ICES (Supplementary Table 1). In 2018, the institute formerly known as the Institute for Clinical Evaluative Sciences formally adopted the initialism of ICES as its official name. The data were collected prospectively during routine healthcare delivery and administrative processes. These datasets included inpatient and outpatient healthcare interactions with information about chronic diseases and sociodemographic characteristics. ICES is an independent, nonprofit research institute, the legal status of which, under Ontario’s health information privacy law, allows the collection and analysis of healthcare data, without consent, for health system evaluation. All births in Ontario hospitals between 1 April 2002 and 31 December 2018 were included, with follow-up data until 31 March 2020. The Institutional Review Board of the Faculty of Medicine and Health Sciences at McGill University approved the study. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies was followed.
Outcome
Epilepsy was defined based on a algorithm validated in Ontario administrative data among adults aged 20–55 years, showing a sensitivity of 73.7% and a specificity of 99.8%, as (i) an inpatient hospitalization with a diagnosis of epilepsy or (ii) three or more outpatient diagnoses of epilepsy based on physician billing codes, separated by ≥30 days within 2 years before the age of 18 years, based on the International Statistical Classification of Diseases-Ninth and Tenth revisions (ICD-9 and 10) codes (G40, G41, or F80.3, Supplementary Table 2) [26–28].
Exposure
SMM was identified according to the criteria of the Canadian Perinatal Surveillance System [5], using diagnostic and procedure codes from the ICD-10 and the Canadian Classification of Interventions. It includes 46 conditions, grouped into 13 subtypes (Supplementary Table 3). SMM was also classified into four specific subtypes as exposure according to the frequency of occurrence in the data from this study and the literature [8]: (i) severe hemorrhage, (ii) severe hypertensive disorders of pregnancy [severe preeclampsia, hemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome, or eclampsia], (iii) sepsis, and (iv) others. In a sensitivity analysis, SMM was further refined into three categories based on the date of the first diagnosis of SMM to differentiate antenatal, perinatal, and postnatal events. Pregnancy onset was estimated by subtracting the clinically recorded gestational age at delivery from the date of birth. Gestational age was obtained from the hospital birth record and is based on accurate pregnancy dating using first- or second-trimester ultrasound for >95% of births in Ontario.
Potential confounders
Confounder selection was guided by using a directed acyclic graph (Supplementary Figure 1), informed by prior literature and clinical knowledge (details of their identification appear in the datasets in Supplementary Tables 1 and 2). At the individual level, maternal age was examined at the time of birth in years [22, 29], parity (number of previous live births classified as 0, one, or at least two) [30], and maternal neurodevelopmental disorders diagnosed before childbirth [epilepsy, other neurodevelopmental disorder (NDD) histories, and none] [22]. Pre-gestational conditions of diabetes, hypertension, and clinically documented obesity were defined by using diagnoses recorded prior to pregnancy onset [23]. Clinically documented maternal substance use including smoking, alcohol, and drug use was identified based on diagnoses recorded before or during pregnancy [29–31]. Additionally, the mothers taking part in the Ontario Drug Benefit (ODB) program were women who were in financial need and met the program criteria, and maternal immigration status was categorized into refugee, non-refugee immigrant, and Canadian-born mothers [32, 33]. At the area level, this analysis included rurality, two dissemination area-specific Ontario Marginalization (ON-Marg) indices—deprivation index and ethnic diversity index (based on the proportion of ethnic minorities or recent immigrants)—and the neighborhood income, with each indexed by quintiles to assess socioeconomic factors [31, 34]. All those characteristics were assessed at or before delivery.
Statistical analysis
Maternal and child characteristics were first compared according to SMM status and major SMM subtypes. The cumulative incidence of epilepsy was then estimated according to SMM by using the Kaplan–Meier estimator.
The crude and adjusted associations between SMM and epilepsy were estimated by using a Cox proportional hazard model. The time axis was the time elapsed from birth to epilepsy diagnosis, death, or end of follow-up data on 31 March 2020. All covariates included in the adjustment set were measured prior to or at delivery and treated as fixed (time-invariant) variables. Maternal age at birth and year of birth were modeled as continuous variables by using penalized splines, while all other covariates were included as categorical variables (‘pspline’ function in the ‘Coxph’ function of R) [35]. The correlation between children born to the same mother was accounted for via a sandwich estimator [36]. The proportional hazard assumption was verified by using Schoenfeld residuals to evaluate whether the hazard ratio (HR) could reasonably be interpreted as constant over the follow-up. In the absence of meaningful violations, standard Cox models were retained. This analysis was repeated on the refined SMM categories considering the time of first diagnosis. The analyses were carried out with R-3.6.1.
Sensitivity analyses
To assess the extent to which the association between antenatal SMM and childhood epilepsy is mediated by preterm birth, a sensitivity analysis that estimated the controlled direct effect (CDE) of antenatal SMM on epilepsy risk was conducted by using marginal structural models. This approach uses inverse-probability weighting to account for potential mediator–outcome confounding, including confounders that may be affected by the exposure. Stabilized weights were calculated based on predicted probabilities from logistic regression models for both the exposure and the mediator, including relevant covariates such as gestational diabetes and the number of prenatal visits. Extreme weights were truncated at the 99th percentile to improve model stability. In 2012, the Canadian ICD codes for hypertensive disorders during pregnancy were updated, distinguishing between severe preeclampsia and mild preeclampsia. To test the robustness of these results against this change, analyses were restricted to children born between 2013 and 2018, for the composite SMM and hypertensive disorders of pregnancy exposure only. Finally, the primary models were repeated whilst restricting the cohort to primiparous women (first recorded pregnancy).
Quantitative bias analysis
To address the impact of potential exposure and/or outcome misclassifications and residual confounding, probabilistic bias analyses were conducted, separately and combined, on record-level data [37] by using sensitivity and specificity values of validation studies based on administrative databases [27, 28]. Obesity is a potential confounder in the relationship between epilepsy in children and SMM [38, 39]. Although there was adjustment for pre-pregnancy obesity in the primary analysis, it only captured clinically documented obesity based on ICD-10 codes, known as being inaccurate for obesity [40]. Hence, this potential residual confounding was accounted for by adjustment for a simulated binary variable representing obesity. Beta-distributed parameters were used, reflecting literature-based estimates for the prevalence of obesity among women with and without SMM (e.g. mode 17.6% vs 13.2%) and its association with epilepsy in offspring [e.g. relative risk (RR) 1.32]. Each analysis was repeated 500 times to generate bias-adjusted hazard ratios and 95% CIs that account for both systematic and random error. Similar analyses were conducted to adjust for individual-level socioeconomic characteristics, using maternal education as a proxy [30, 41]. Details of the parameter selection, assumptions, distributions, and references are provided in the Supplementary Methods.
Results
Of the 2 188 175 children identified, 1215 babies born before 21 and after 43 weeks of gestation, 1494 stillbirths, 6030 children with missing or implausible records (e.g. date of epilepsy diagnosis that occurred before birth), and 73 883 multiple births were excluded (Supplementary Figure 2). Hence, this cohort had 2 060 668 children born between 2002 and 2018, with a median follow-up period of 10.1 years (interquartile range, 6.0–14.2). Of these, 40 826 (2%) were born to mothers who had experienced SMM, among whom 10 950 had a severe hemorrhage, 6343 had a severe hypertensive disorder of pregnancy, and 9221 developed sepsis. Mothers with SMM were older at delivery, had more comorbidities (14.1% vs 11.1% of pre-pregnancy obesity and 3.1% vs 1.5% of pre-pregnancy diabetes), were twice as likely to have a history of epilepsy, and were more likely to be nulliparous (50.9% vs 43.9%). Similar results were found across subtypes, except for certain characteristics. For instance, women with a severe hypertensive disorder of pregnancy were more likely to have pre-gestational hypertension and were more likely to be nulliparous (Table 1). Mothers with SMM were more likely to be participants in the provincial drug benefit program (21.8% recipients vs 18.1%, Table 1) and to live in rural and more socially deprived areas (Table 1). This difference was more pronounced for women who had a severe hypertensive disorder of pregnancy.
Table 1.
Individual and area-level characteristics at birth [n (%)] according to maternal pregnancy illness status of 2 060 668 children born between 2002 and 2018 in Ontario, Canada.
| Variables | No SMM (n = 2 019 838) | SMM (n = 40 830) | Severe hemorrhage (n = 10 950) | Hypertensive disorders of pregnancya (n = 6343) | Sepsis (n = 9221) | Other types (n = 15 534) |
|---|---|---|---|---|---|---|
| Epilepsy | ||||||
| Yes | 16 850 (0.8) | 493 (1.2) | 124 (1.1) | 79 (1.2) | 105 (1.1) | 196 (1.3) |
| No | 2 002 988 (99.2) | 40 337 (98.8) | 10 826 (98.9) | 6264 (98.8) | 9116 (98.9) | 15 338 (98.7) |
| Maternal age at delivery (years) | ||||||
| <20 | 61 776 (3.1) | 1421 (3.5) | 406 (3.7) | 245 (3.9) | 455 (4.9) | 370 (2.4) |
| 20–24 | 250 201 (12.4) | 4869 (11.9) | 1278 (11.7) | 794 (12.5) | 1307 (14.2) | 1655 (10.7) |
| 25–29 | 563 213 (27.9) | 10 207 (25.0) | 2793 (25.5) | 1625 (25.6) | 2555 (27.7) | 3556 (22.9) |
| 30–34 | 707 745 (35.0) | 13 394 (32.8) | 3603 (32.9) | 1983 (31.3) | 2964 (32.1) | 5242 (33.7) |
| 35–39 | 360 982 (17.9) | 8463 (20.7) | 2222 (20.3) | 1294 (20.4) | 1584 (17.2) | 3587 (23.1) |
| ≥40 | 75 574 (3.7) | 2476 (6.1) | 648 (5.9) | 402 (6.3) | 356 (3.9) | 1124 (7.2) |
| Recipient of ODB | ||||||
| Yes | 366 347 (18.1) | 8906 (21.8) | 2164 (19.8) | 1286 (20.3) | 1974 (21.4) | 3725 (24.0) |
| No | 1 653 391 (81.9) | 31 924 (78.2) | 8786 (80.2) | 5057 (79.7) | 7247 (78.6) | 11 809 (76.0) |
| Maternal immigration status | ||||||
| Refugees | 74 842 (3.7) | 1803 (4.5) | 417 (3.8) | 244 (3.8) | 386 (4.2) | 805 (5.2) |
| Non-refugees | 480 865 (23.8) | 9246 (22.6) | 2301 (21.0) | 1376 (21.7) | 2057 (22.3) | 3731 (24.0) |
| Canadian-born | 1 464 130 (72.5) | 29 781 (72.9) | 8232 (75.2) | 4723 (74.5) | 6778 (73.5) | 10 998 (70.8) |
| Maternal substance use (clinically documented) | ||||||
| Yes | 172 596 (8.5) | 4189 (10.3) | 1168 (10.7) | 550 (8.7) | 1027 (11.1) | 1592 (10.2) |
| No | 1 847 242 (91.5) | 36 641 (89.7) | 9782 (89.3) | 5793 (91.3) | 8194 (88.9) | 13 942 (89.8) |
| Pre-pregnancy obesity (clinically documented) | ||||||
| Yes | 215 709 (10.7) | 5359 (13.1) | 1072 (9.8) | 992 (15.6) | 1120 (12.1) | 2295 (14.8) |
| No | 1 804 129 (89.3) | 35 471 (88.9) | 9878 (90.2) | 5351 (84.4) | 8101 (87.9) | 13 239 (85.2) |
| Pre-pregnancy hypertension | ||||||
| Yes | 43 946 (2.2) | 1924 (4.7) | 284 (2.6) | 558 (8.8) | 260 (2.8) | 864 (5.6) |
| No | 1 975 892 (97.8) | 38 906 (95.3) | 10 666 (97.4) | 5785 (91.2) | 8961 (97.2) | 14 670 (94.4) |
| Pre-pregnancy diabetes | ||||||
| Yes | 29 983 (1.5) | 1271 (3.1) | 210 (2.0) | 285 (4.5) | 204 (2.2) | 599 (3.9) |
| No | 1 989 855 (98.5) | 39 559 (96.9) | 10 740 (98.0) | 6058 (95.5) | 9017 (97.8) | 14 935 (96.1) |
| Maternal history of NDD | ||||||
| Epilepsy | 12 150 (0.6) | 426 (1.0) | 93 (0.8) | 64 (1.0) | 75 (0.8) | 207 (1.3) |
| Other NDD | 3785 (0.2) | 106 (0.3) | 23 (0.3) | 215 (0.2) | 19 (0.2) | 52 (0.3) |
| No | 2 003 903 (99.2) | 40 298 (98.7) | 10 834 (98.9) | 6264 (98.8) | 9127 (99.0) | 15 275 (98.3) |
| Parity | ||||||
| 0 | 886 950 (43.9) | 20 794 (50.9) | 5580 (51.0) | 3988 (62.9) | 4785 (51.9) | 7071 (45.5) |
| 1 | 734 263 (36.4) | 12 084 (29.6) | 3108 (28.4) | 1441 (22.7) | 2716 (29.5) | 5175 (33.3) |
| ≥2 | 398 244 (19.7) | 7943 (19.5) | 2258 (20.6) | 911 (14.4) | 1718 (18.6) | 3288 (21.2) |
| Baby’s sex | ||||||
| Male | 1 036 292 (51.3) | 20 799 (50.9) | 5293 (51.7) | 3238 (51.0) | 4773 (51.8) | 8076 (52.0) |
| Female | 983 463 (48.7) | 20 028 (49.1) | 5656 (48.3) | 3103 (49.0) | 4447 (48.2) | 7457 (48.0) |
| ON-Marg material deprivation quintile | ||||||
| Q1 (least deprived) | 405 642 (20.3) | 7619 (19.0) | 2082 (19.6) | 1139 (18.3) | 1746 (19.4) | 2881 (18.8) |
| Q2 | 375 074 (18.8) | 7137 (17.8) | 1902 (17.9) | 1099 (17.6) | 1644 (18.3) | 2701 (17.6) |
| Q3 | 371 464 (18.6) | 7257 (18.5) | 1949 (18.4) | 1140 (18.2) | 1669 (18.6) | 2713 (17.7) |
| Q4 | 377 905 (19.0) | 7668 (19.0) | 2012 (19.0) | 1232 (19.8) | 1720 (19.1) | 2924 (19.1) |
| Q5 (most deprived) | 463 520 (23.3) | 10 292 (25.7) | 2663 (25.1) | 1618 (26.1) | 2203 (24.5) | 4109 (26.8) |
| Missing | 26 233 (1.3) | 857 (2.1) | 342 (3.1) | 115 (1.8) | 239 (2.6) | 206 (1.3) |
| ON-Marg ethnic diversity quintile | ||||||
| Q1 (least diverse) | 256 920 (12.9) | 5848 (14.6) | 1696 (16.0) | 1003 (16.1) | 1319 (14.7) | 2021 (13.2) |
| Q2 | 292 156 (14.7) | 6125 (15.3) | 1818 (17.1) | 982 (15.8) | 1419 (15.8) | 2157 (14.1) |
| Q3 | 332 898 (16.7) | 6182 (16.9) | 1841 (16.7) | 1088 (17.5) | 1499 (16.7) | 2494 (16.3) |
| Q4 | 425 672 (21.3) | 8300 (20.8) | 2137 (20.1) | 1281 (20.6) | 1838 (20.5) | 3276 (21.4) |
| Q5 (most diverse) | 685 959 (34.4) | 12 947 (32.4) | 3116 (29.4) | 1874 (30.1) | 2907 (32.4) | 5380 (35.1) |
| Missing | 26 233 (1.3) | 857 (2.1) | 342 (3.1) | 115 (1.8) | 239 (2.6) | 206 (1.3) |
| Rural | ||||||
| Yes | 124 616 (6.3) | 3043 (7.6) | 939 (9.0) | 557 (9.0) | 723 (8.1) | 927 (6.0) |
| No | 1 870 720 (93.7) | 36 946 (92.4) | 9662 (91.0) | 5657 (91.0) | 8258 (91.9) | 14 438 (94.0) |
| Missing | 26 602 (1.2) | 841 (2.1) | 349 (3.2) | 129 (2.0) | 240 (2.6) | 169 (1.1) |
| Neighborhood income quintile | ||||||
| Q1 (lowest) | 449 585 (22.4) | 10 268 (25.4) | 2713 (25.0) | 1602 (25.4) | 2231 (24.3) | 4030 (26.0) |
| Q2 | 402 770 (20.0) | 8048 (19.8) | 2131 (19.6) | 1231 (19.5) | 1873 (20.4) | 3051 (19.7) |
| Q3 | 412 441 (20.5) | 8191 (20.1) | 2183 (20.1) | 1305 (20.7) | 1844 (20.1) | 3090 (20.0) |
| Q4 | 416 668 (20.7) | 7932 (19.5) | 2116 (19.5) | 1231 (19.5) | 1829 (20.0) | 2993 (19.3) |
| Q5 (highest) | 330 829 (16.4) | 6137 (15.2) | 1706 (15.8) | 929 (14.8) | 1388 (15.1) | 2310 (14.9) |
| Missing | 7545 (0.4) | 254 (0.6) | 101 (0.9) | 45 (0.7) | 56 (0.6) | 60 (0.4) |
Severe preeclampsia, HELLP syndrome, and eclampsia.
A total of 16 992 (0.8%) cases of epilepsy were diagnosed in this population, with a median age of 3 years at diagnosis (Supplementary Table 4). Figures 1 and 2 present the crude cumulative incidence of epilepsy according to SMM and its different types, respectively. The cumulative incidence of epilepsy was consistently higher among children born to mothers who had any SMM or a specific type compared with those who were unexposed.
Figure 1.
Cumulative incidence of epilepsy (per 1000 children) according to exposure to SMM estimated using the Kaplan–Meier estimator among children born between 2002 and 2018 in Ontario, Canada (N = 2 060 668).
Figure 2.
Cumulative incidence of epilepsy (per 1000 children) according to SMM subtypes estimated using the Kaplan–Meier estimator among children born between 2002 and 2018 in Ontario, Canada (N = 2 060 668).
Results from the adjusted Cox models showed positive associations between SMM and the diagnosis of epilepsy in children, overall and for each SMM subtype considered (Table 2). After adjustment for available potential confounders, children born to mothers who had SMM had a 45% higher risk of being diagnosed with epilepsy compared with those whose mothers did not have SMM (HR 1.45, 95% CI 1.33–1.59). Similar results were found for each subtype, with overlapping CIs with each other (Table 2). When estimates according to the timing of the diagnosis of the first SMM were looked at (before, during, or after birth), the risk of epilepsy was elevated in all groups, including SMM diagnosed after birth, and there were no differences by the timing of the first condition diagnosed (Supplementary Table 5). After the mediating role of preterm birth was accounted for, the CDE was HR 1.29 (95% CI 1.07–1.55), suggesting that part, but not all, of the association may be explained by preterm delivery (Supplementary Table 6).
Table 2.
Association between SMM or subtypes of SMM and epilepsy in children born in Ontario, Canada between 2002 and 2018.
| Epilepsy cases/n | Crude HR (95% CI)a n = 2 060 668 | Epilepsy cases/n | Adjusted HR (95% CI)b n = 2 022 275 | |
|---|---|---|---|---|
| SMM (overall) | ||||
| No | 16 850/2 019 838 | Ref. | 16 532/1 982 591 | Ref. |
| Yes | 493/40 830 | 1.52 (1.39–1.66) | 481/39 684 | 1.45 (1.33–1.59) |
| Severe hemorrhage | ||||
| No SMM | 16 850/2 019 838 | Ref. | 16 532/1 982 591 | Ref. |
| Yes | 124/10 950 | 1.41 (1.19–1.69) | 122/10 519 | 1.42 (1.19–1.70) |
| Severe preeclampsia, HELLP syndrome, and eclampsia | ||||
| No SMM | 16 850/2 019 838 | Ref. | 16 532/1 982 591 | Ref. |
| Yes | 79/6343 | 1.80 (1.44–2.24) | 77/6170 | 1.66 (1.33–2.08) |
| Sepsis | ||||
| No SMM | 16 850/2 019 838 | Ref. | 16 532/1 982 591 | Ref. |
| Yes | 105/9221 | 1.38 (1.14–1.67) | 102/8904 | 1.36 (1.12–1.66) |
| Other SMM subtypes | ||||
| No SMM | 16 850/2 019 838 | Ref. | 16 532/1 982 591 | Ref. |
| Yes | 196/15 534 | 1.54 (1.34–1.78) | 191/15 249 | 1.45 (1.25–1.67) |
Crude estimates are shown for descriptive purposes only; adjusted models are used for primary inference.
Adjusted for maternal age at birth (spline), parity (zero, one, or at least two pregnancies), maternal history of the neurodevelopmental disorder (binary), pre-gestational obesity (binary), pre-gestational diabetes (binary), pre-gestational hypertension (binary), participation in the ODB program (binary), maternal immigration status(refugee, non-refugee, Canadian-born mothers), maternal substance use (including smoking, alcohol, and drug/binary), ON-Marg deprivation index (quintiles), ON-Marg ethnic diversity (quintiles) index, neighborhood income index (quintile), rurality (binary), baby’s sex (male, female, other), and year of birth (spline).
In the sensitivity analysis on children born between 2013 and 2018, the estimates were similar to those of the main results (Supplementary Table 7). Results were also consistent when analyses were restricted to primiparous women (RR 1.41, 95% CI 1.24–1.61; Supplementary Table 8).
After the association accounting for potential for exposure and/or outcome misclassification was estimated (Table 3), it was observed that they were slightly higher, particularly after accounting for outcome misclassification, showing a 2-fold increased risk of epilepsy for children exposed to SMM (HR 2.00, 95% CI 1.77–2.29). When residual confounding by maternal obesity or education was accounted for, the results were not substantially different from the primary results. Results accounting for all three biases simultaneously showed increased risks (Table 3).
Table 3.
Association between epilepsy in children and SMM corrected for exposure and outcome misclassifications and residual confounding
| Crude HR (95% CI) | 1.52 (1.39–1.66) |
| Crude HR with correction for exposure misclassification (95% CI) | 1.68 (1.49–1.88) |
| Crude HR with correction for outcome misclassification (95% CI) | 1.97 (1.76–2.22) |
| Crude HR with correction for exposure and outcome misclassifications (95% CI) | 2.00 (1.77–2.29) |
| Adjusted HR (95% CI) | 1.42 (1.17–1.71) |
| Adjusted HR with correction for uncontrolled confounding by maternal education (95% CI) | 1.47 (1.23–1.78) |
| Adjusted HR with correction for uncontrolled confounding by obesity (95% CI) | 1.47 (1.24–1.78) |
| Adjusted HR with correction for exposure and outcome misclassifications and for uncontrolled confounding by maternal education (95% CI) | 1.91 (1.50–2.41) |
| Adjusted HR with correction for exposure and outcome misclassifications and for uncontrolled confounding by obesity (95% CI) | 1.91 (1.50–2.42) |
Adjusted for maternal age at birth, parity, maternal history of the NDD, pre-gestational obesity, pre-gestational diabetes, pre-gestational hypertension, participation in the ODB program, maternal immigration status, maternal substance use, ON-Marg deprivation index, ON-Marg ethnic diversity index, neighborhood income index, rurality, baby’s sex, and year of birth.
Discussion
In this large population-based cohort study of >2 million births, children born to mothers with SMM had an increased risk of epilepsy. The elevated risk remained across subtypes and different time windows of the onset of SMM.
Although SMM is a composite measure of severe maternal–perinatal adversity that substantially increases the risk of maternal mortality, the heterogeneity in underlying conditions makes causal interpretation complex. Therefore, studies focused on the effects of specific conditions of SMM are informative for understanding its impact on epilepsy. For instance, preeclampsia and eclampsia have mechanisms known to lead to abnormal fetal brain development, impaired placental perfusion, fetal hypoxia, or conditions such as encephalopathy. Although some of the previous studies have shown no associations [42, 43], the results of this study are consistent with those of most of the existing studies reporting an increased risk [44–46]. For example, Whitehead et al., in a Canadian study of >120 000 children, showed that eclampsia increased the risk of epilepsy (adjusted relative risk 14.2, 95% CI 3.5–57.3) [47]. The large difference in the risk estimate from the results of this current study could be explained by the fact that eclampsia was extremely rare in their population (<34 among 124 207 live births), that they used a different method to identify epilepsy cases (registry and database cases and electroencephalography records reviewed by three pediatric neurologists), or that the adjustment factors chosen were different and did not include important confounders such as maternal age, previous hypertension, obesity, or maternal history of epilepsy. However, similar magnitudes to those of this study were observed in Denmark [48], South Carolina [49], Norway [50], and Taiwan [51], outlining a consistent association across diverse populations, suggesting that the pathophysiological mechanisms connecting severe preeclampsia and eclampsia to pediatric epilepsy may be generalizable across different healthcare settings and population demographics.
Although the literature on the other major types of SMM is much more scarce, it does tend to support the results of this study. A higher risk of epilepsy was observed among children born to women who had a placental abruption (condition of severe antenatal hemorrhage) [52], leading to fetal hypoxia or preterm birth, as well as those born to mothers who had infections during pregnancy [53, 54]. Preterm birth is a well-established risk factor for childhood epilepsy and may lie on the causal pathway between antenatal SMM and neurodevelopmental outcomes. It may act as a collider if conditioned upon; however, the primary analyses of this study avoided such conditioning and mediation analyses were conducted within a causal inference framework. However, in a sensitivity analysis exploring the potential mediation by preterm birth, the association was attenuated but remained positive, suggesting that preterm birth does not fully explain it. The findings of this study regarding sepsis are consistent with this as well. This may be explained by the fact that sepsis is a severe systemic response to infection, leading to an increased maternal inflammatory response. This may cause maternal cytokine production, resulting in fetal neurological injury that can contribute to the risk of epilepsy [53]. Maternal inflammation could also lead to the dysfunction of cortical GABAergic interneurons [54], which has been identified in the microcircuits in epilepsy [55].
Strengths and limitations
This study is based on a large population-based cohort of children including all hospital births from Ontario, covering >15 years and with a median follow-up of >10 years. Because 44% of epilepsy cases in this population were diagnosed between the ages of 0 and 2 years, the length of follow-up in the data used in this study enabled the capture of most cases. This dataset also provided considerable statistical power and precision to estimate the risk of epilepsy associated with SMM and reduced potential selection bias, as all live births were included.
However, both epilepsy and SMM cases were identified by using diagnostic or procedure codes, and therefore potential misclassification in the exposure and outcome cannot be ruled out. The definition applied in this study to identify epilepsy cases has a sensitivity of 73.7% and a specificity of 99.8% [28]. Although validated in adults only, the epilepsy case definition was applied within the same provincial administrative data. Many adult cases represent childhood-onset epilepsy, supporting their applicability. However, it is acknowledged that sensitivity may vary by epilepsy type and age of onset, and it was not possible to differentiate between early- and late-onset pediatric epilepsy or epilepsy subtypes, which may be differentially associated with prenatal exposures. Results of the quantitative bias analyses of this study suggest that the misclassification of epilepsy might have attenuated the association, strengthening the conclusion of an increased risk for exposed children. Regarding SMM, a published definition based on a combination of “a priori clinical knowledge, prolonged length of stay, high case fatality, and expert consensus” was used [5]. It has not been validated and Dzakpasu et al. have acknowledged the difficulty in capturing all conditions with sufficient precision by using only ICD-10 codes. For instance, it is challenging to distinguish between preexisting conditions and acute complications due to pregnancy. However, in the quantitative bias analyses of this study, as for the outcome misclassification bias, it seems that exposure misclassification may have attenuated the observed association. The probabilistic bias analyses of this study assumed non-differential misclassification; differential misclassification, if present, could have biased the effect estimates in either direction and it is difficult to quantify by using the available data.
Residual confounding is a well-known limitation of administrative data. Important individual-level maternal characteristics, such as individual socioeconomic factors (e.g. maternal education, family income, or employment), paternal history of epilepsy, and potential unmeasured genetic confounders or body mass index, were missing. The diagnostic code of clinically documented obesity was adjusted for via a variable look-back window, but this approach underestimates obesity compared with population-based estimates. As obesity is a chronic condition that is inconsistently coded in administrative data, a fixed look-back window would not necessarily improve classification. Data on maternal co-medications were also unavailable but confounding by indication was partially addressed by adjusting for pre-gestational maternal conditions. Similarly, other pre-pregnancy characteristics, including maternal substance use and chronic conditions, may have been misclassified in the administrative data. Such underestimation of confounders, particularly obesity, would likely have biased the observed associations away from the null, meaning that the true associations could be stronger than those estimated. However, the quantitative bias analyses used in this study to examine the extent of potential residual confounding did not show any important difference from the primary results, providing some reassurance that residual confounding is unlikely to fully account for the observed associations or reverse their direction.
Additionally, due to the composite nature of SMM by definition, understanding the specific mechanisms and identifying the factors responsible for the observed association with the risk of offspring epilepsy are challenging. Although antenatal complications may be more biologically plausible drivers of neurodevelopmental risk, the sensitivity analysis of this study showed that the risk of epilepsy was similarly elevated regardless of whether the first SMM diagnosis occurred before, during, or after delivery. This lack of variation by timing may reflect shared maternal and perinatal risk factors across the pregnancy continuum, including chronic maternal conditions, placental abnormalities, systemic inflammation, or delivery complications. In particular, postpartum SMM, such as hemorrhagic events, is unlikely to directly affect fetal brain development but may act as an epidemiological marker of complicated pregnancies or cumulative physiological stressors that were already present during gestation. These findings support the interpretation that maternal SMM could reflect an underlying multifactorial vulnerability shaped by individual, clinical, social, and systemic factors (e.g. maternal age, nutrition, comorbidities, access to care) rather than a direct postnatal causal mechanism. Although SMM precedes the diagnosis of epilepsy in offspring, reverse causation cannot be fully excluded, as fetal abnormalities may increase maternal physiological stress and some SMM events may reflect preexisting maternal health problems. Thus, maternal SMM is best interpreted as a marker of cumulative maternal–perinatal risk rather than direct biological exposure acting after birth. Finally, the effect of SMM may vary according to the type of epilepsy diagnosed but it was impossible to explore this due to the absence of information on epilepsy subtypes in the dataset of this study.
Conclusion
This study investigated the association between severe maternal morbidity in a large population cohort and the onset of epilepsy in children, highlighting an increased risk of epilepsy in children exposed to SMM, regardless of the morbidity subtype. These results support the exploration of a perinatal etiology of epilepsy in children and underscore the importance of targeted postnatal surveillance for children exposed to SMM to enable early detection and intervention.
Ethics approval
The Institutional Review Board of the Faculty of Medicine and Health Sciences at McGill University approved the study.
Supplementary Material
Acknowledgements
This document used data adapted from the Statistics Canada Postal CodeOM Conversion File, which is based on data licensed from Canada Post Corporation and/or data adapted from the Ontario Ministry of Health Postal Code Conversion File, which contains data copied under license from Canada Post Corporation and Statistics Canada.
Contributor Information
Bénédicte Driollet, Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Quebec, Canada.
Emmalin Buajitti, Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Quebec, Canada; ICES, Toronto, ON, Canada; Dalla Lana School of Public Health, Division of Epidemiology, University of Toronto, Toronto, Ontario, Canada.
Asma M Ahmed, Department of Epidemiology and Prevention, Division of Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, North Carolina, USA.
Jennifer A Hutcheon, Department of Obstetrics and Gynaecology, The University of British Columbia, Vancouver, BC, Canada.
Laura Rosella, ICES, Toronto, ON, Canada; Dalla Lana School of Public Health, Division of Epidemiology, University of Toronto, Toronto, Ontario, Canada; Institute for Better Health, Trillium Health Partners, Mississauga, Ontario, Canada; Temerty Faculty of Medicine, Department of Laboratory Medicine and Pathology, University of Toronto, Toronto, Ontario, Canada.
Seungmi Yang, Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Quebec, Canada.
Author contributions
B.D. conceptualized and designed the study, carried out the analyses, drafted the initial manuscript, and reviewed and revised the manuscript. S.Y. conceptualized and designed the study, and drafted, reviewed, and revised the manuscript. A.A. participated in the conceptualization of the study, contributed to coding, and reviewed the manuscript. E.B. coordinated and supervised data collection, and participated in data creation and data management. L.R. coordinated and supervised data collection, participated in the conceptualization of the study, and reviewed the manuscript. J.H. participated in the conceptualization of the study and reviewed the manuscript. All authors critically reviewed the manuscript, approved the final version as submitted, and agreed to be accountable for all aspects of the work.
Supplementary material
Supplementary material is available at IJE online.
Conflicts of interest
None declared.
Funding
This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health (MOH) and the Ministry of Long-Term Care (MLTC). This study also received funding from the Canadian Institutes of Health Research (PJT-175250). Parts of this material are based on data and information compiled and provided by MOH, CIHI, and Immigration, Refugees and Citizenship Canada (IRCC). The analyses, conclusions, opinions, and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred.
Data availability
Parts of this material are based on data and/or information compiled and provided by IRCC to 2020. However, the analyses, conclusions, opinions, and statements expressed in the material are those of the authors and not necessarily those of IRCC. The dataset from this study is held securely in coded form at ICES. ICES maintains a secure and comprehensive data repository containing individual-level, coded, and linkable health data for the population of Ontario, Canada. These data are routinely collected through administrative sources such as hospital-discharge abstracts, physician billing claims, drug-dispensing records, and vital statistics. Each Ontario resident has a unique encoded health identifier (based on the Ontario Health Insurance Plan, or OHIP number), which allows deterministic linkage across datasets within the ICES environment. This linkage is performed by using validated algorithm unique identifiers and is subject to rigorous data-quality assessments, including duplicate checks, consistency validations, and routine updates. In addition to Ontario’s Personal Health Information Protection Act (PHIPA) authorization (section 45), ICES operates under a stringent privacy and security governance framework: all policies are IPC-approved, regularly audited, and include measures such as data de-identification, controlled server access, encryption, audit logs, breach response protocols, and mandatory privacy training. While legal data-sharing agreements between ICES and data providers (e.g. healthcare organizations and government) prohibit ICES from making the dataset publicly available, access may be granted to those who meet pre-specified criteria for confidential access, available at www.ices.on.ca/DAS (e-mail: das@ices.on.ca).
Use of artificial intelligence (AI) tools
No AI tools were used in the preparation of this paper.
References
- 1. Firoz T, Chou D, von Dadelszen P et al. ; Maternal Morbidity Working Group. Measuring maternal health: focus on maternal morbidity. Bull World Health Organ 2013;91:794–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Mengistu TS, Turner JM, Flatley C et al. The impact of severe maternal morbidity on perinatal outcomes in high-income countries: systematic review and meta-analysis. J Clin Med 2020;9:2035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Schaap T, Bloemenkamp K, Deneux-Tharaux C, INOSS et al. Defining definitions: a Delphi study to develop a core outcome set for conditions of severe maternal morbidity. Bjog 2019;126:394–401. [DOI] [PubMed] [Google Scholar]
- 4. Snowden JM, Lyndon A, Kan P et al. Severe maternal morbidity: a comparison of definitions and data sources. Am J Epidemiol 2021;190:1890–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Dzakpasu S, Deb-Rinker P, Arbour L et al. Severe maternal morbidity surveillance: monitoring pregnant women at high risk for prolonged hospitalisation and death. Paediatr Perinat Epidemiol 2020;34:427–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Callaghan WM, Creanga AA, Kuklina EV. Severe maternal morbidity among delivery and postpartum hospitalizations in the United States. Obstet Gynecol 2012;120:1029–36. [DOI] [PubMed] [Google Scholar]
- 7. Aoyama K, Ray JG, Pinto R et al. Temporal variations in incidence and outcomes of critical illness among pregnant and postpartum women in Canada: a population-based observational study. J Obstet Gynaecol Can 2019;41:631–40. [DOI] [PubMed] [Google Scholar]
- 8. Dzakpasu S, Deb-Rinker P, Arbour L, et al. ; Canadian Perinatal Surveillance System (Public Health Agency of Canada). Severe maternal morbidity in Canada: temporal trends and regional variations, 2003-2016. J Obstet Gynaecol Can 2019;41:1589–98.e16. [DOI] [PubMed] [Google Scholar]
- 9. Marr L, Lennox C, McFadyen AK. Quantifying severe maternal morbidity in Scotland: a continuous audit since 2003. Curr Opin Anaesthesiol 2014;27:275–81. [DOI] [PubMed] [Google Scholar]
- 10. Creanga AA, Berg CJ, Ko JY et al. Maternal mortality and morbidity in the United States: where are we now? J Womens Health (Larchmt) 2014;23:3–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Ray JG, Park AL, Dzakpasu S et al. Prevalence of severe maternal morbidity and factors associated with maternal mortality in Ontario, Canada. JAMA Netw Open 2018;1:e184571. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Hirai AH, Owens PL, Reid LD et al. Associations between state-level severe maternal morbidity and other perinatal indicators. JAMA Netw Open 2022;5:e2224621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Zanardi DM, Parpinelli MA, Haddad SM et al. ; Brazilian Network for Surveillance of Severe Maternal Morbidity Study Group. Adverse perinatal outcomes are associated with severe maternal morbidity and mortality: evidence from a national multicentre cross-sectional study. Arch Gynecol Obstet 2019;299:645–54. [DOI] [PubMed] [Google Scholar]
- 14. Aoyama K, Park AL, Davidson AJF, Ray JG. Severe maternal morbidity and infant mortality in Canada. Pediatrics 2020;146:e20193870. [DOI] [PubMed] [Google Scholar]
- 15. Geller SE, Koch AR, Garland CE et al. A global view of severe maternal morbidity: moving beyond maternal mortality. Reprod Health 2018;15:98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Santana DS, Cecatti JG, Haddad SM et al. ; Brazilian Network for the Surveillance of Severe Maternal Morbidity Study Group. Severe maternal morbidity and perinatal outcomes of multiple pregnancy in the Brazilian network for the surveillance of severe maternal morbidity. Int J Gynaecol Obstet 2017;139:230–8. [DOI] [PubMed] [Google Scholar]
- 17. Cham M, Sundby J, Vangen S. Fetal outcome in severe maternal morbidity: too many stillbirths. Acta Obstet Gynecol Scand 2009;88:343–9. [DOI] [PubMed] [Google Scholar]
- 18. Fitzgerald E, Hor K, Drake AJ. Maternal influences on fetal brain development: the role of nutrition, infection and stress, and the potential for intergenerational consequences. Early Hum Dev 2020;150:105190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Beghi E. The epidemiology of epilepsy. Neuroepidemiology 2020;54:185–91. [DOI] [PubMed] [Google Scholar]
- 20. Driollet B, Buajitti E, Ahmed AM et al. Prevalence and temporal trends of epilepsy in children: a retrospective birth cohort study. Neuroepidemiology 2025;59:323–33. 10.1159/000540528 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Ketata I, Ellouz E, Mizouri R. Impact of prenatal, neonatal, and postnatal factors on epilepsy risk in children and adolescents: a systematic review and meta-analysis. Acta Epileptologica 2024;6:1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Walsh S, Donnan J, Fortin Y et al. A systematic review of the risk factors associated with the onset and natural progression of epilepsy. Neurotoxicology 2017;61:64–77. [DOI] [PubMed] [Google Scholar]
- 23. Razaz N, Tedroff K, Villamor E et al. Maternal body mass index in early pregnancy and risk of epilepsy in offspring. JAMA Neurol 2017;74:668–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Lawton BA, Jane MacDonald E, Stanley J et al. Preventability review of severe maternal morbidity. Acta Obstet Gynecol Scand 2019;98:515–22. [DOI] [PubMed] [Google Scholar]
- 25. van Dillen J, Mesman JA, Zwart JJ et al. Introducing maternal morbidity audit in the Netherlands. Bjog 2010;117:416–21. [DOI] [PubMed] [Google Scholar]
- 26. Jette N, Reid AY, Quan H et al. How accurate is ICD coding for epilepsy? Epilepsia 2010;51:62–9. [DOI] [PubMed] [Google Scholar]
- 27. Reid AY, St Germaine-Smith C, Liu M et al. Development and validation of a case definition for epilepsy for use with administrative health data. Epilepsy Res 2012;102:173–9. [DOI] [PubMed] [Google Scholar]
- 28. Tu K, Wang M, Jaakkimainen RL et al. Assessing the validity of using administrative data to identify patients with epilepsy. Epilepsia 2014;55:335–43. [DOI] [PubMed] [Google Scholar]
- 29. Aoyama K, Pinto R, Ray JG et al. Association of maternal age with severe maternal morbidity and mortality in Canada. JAMA Netw Open 2019;2:e199875. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Gray KE, Wallace ER, Nelson KR et al. Population-based study of risk factors for severe maternal morbidity. Paediatr Perinat Epidemiol 2012;26:506–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Asadi-Pooya AA, Hojabri K. Risk factors for childhood epilepsy: a case-control study. Epilepsy Behav 2005;6:203–6. [DOI] [PubMed] [Google Scholar]
- 32. Jairam JA, Vigod SN, Siddiqi A et al. Severe maternal morbidity and mortality among immigrant and Canadian-born women residing within low-income neighbourhoods in Ontario, Canada. JAMA Netw Open 2023;6:e2256203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Wandell P, Fredrikson S, Carlsson AC et al. Epilepsy in second-generation immigrants: a cohort study of all children up to 18 years of age in Sweden. Eur J Neurol 2020;27:152–9. [DOI] [PubMed] [Google Scholar]
- 34. Hansen AC, Slavova S, O’Brien JM. Rural residency as a risk factor for severe maternal morbidity. J Rural Health 2022;38:161–70. [DOI] [PubMed] [Google Scholar]
- 35. Eilers PHC, Marx BD. Flexible smoothing with B-splines and penalties. Stat Sci 1996;11:89–121. [Google Scholar]
- 36. Paul SR, Zaihra T. Interval estimation of risk difference for data sampled from clusters. Stat Med 2008;27:4207–20. [DOI] [PubMed] [Google Scholar]
- 37. Fox MP, MacLehose RF, Lash TL. Applying Quantitative Bias Analysis to Epidemiologic Data. Springer; 2021. [Google Scholar]
- 38. Siddiqui A, Azria E, Howell EA et al. ; EPIMOMS Study Group. Associations between maternal obesity and severe maternal morbidity: findings from the French EPIMOMS population-based study. Paediatr Perinat Epidemiol 2019;33:7–16. [DOI] [PubMed] [Google Scholar]
- 39. Pan C, Deroche CB, Mann JR et al. Is prepregnancy obesity associated with risk of cerebral palsy and epilepsy in children? J Child Neurol 2014;29:NP196–201. [DOI] [PubMed] [Google Scholar]
- 40. Wall-Wieler E, Abrams B, Snowden JM et al. Defining maternal obesity in studies of birth outcomes: comparing ICD-9 codes at delivery and measures on the birth certificate. Paediatr Perinat Epidemiol 2020;34:618–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Xinjun L, Jan S, Bengt Z et al. Neighborhood, family, and childhood and adolescent epilepsy: a nationwide epidemiological study from Sweden. Seizure 2014;23:62–8. [DOI] [PubMed] [Google Scholar]
- 42. Maher GM, O’Keeffe GW, Kearney PM et al. Association of hypertensive disorders of pregnancy with risk of neurodevelopmental disorders in offspring: a systematic review and meta-analysis. JAMA Psychiatry 2018;75:809–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Wu CS, Sun Y, Vestergaard M et al. Preeclampsia and risk for epilepsy in offspring. Pediatrics 2008;122:1072–8. [DOI] [PubMed] [Google Scholar]
- 44. Impey L, Greenwood C, Sheil O et al. The relation between pre-eclampsia at term and neonatal encephalopathy. Arch Dis Child Fetal Neonatal Ed 2001;85:F170–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Zwart JJ, Richters JM, Ory F et al. Severe maternal morbidity during pregnancy, delivery and puerperium in the Netherlands: a nationwide population-based study of 371,000 pregnancies. Bjog 2008;115:842–50. [DOI] [PubMed] [Google Scholar]
- 46. Pisani F, Orsini M, Braibanti S et al. Development of epilepsy in newborns with moderate hypoxic-ischemic encephalopathy and neonatal seizures. Brain Dev 2009;31:64–8. [DOI] [PubMed] [Google Scholar]
- 47. Whitehead E, Dodds L, Joseph KS et al. Relation of pregnancy and neonatal factors to subsequent development of childhood epilepsy: a population-based cohort study. Pediatrics 2006;117:1298–306. [DOI] [PubMed] [Google Scholar]
- 48. Mann JR, McDermott S. Maternal pre-eclampsia is associated with childhood epilepsy in South Carolina children insured by Medicaid. Epilepsy Behav 2011;20:506–11. [DOI] [PubMed] [Google Scholar]
- 49. Sun BZ, Moster D, Harmon QE et al. Association of preeclampsia in term births with neurodevelopmental disorders in offspring. JAMA Psychiatry 2020;77:823–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50. Chen KR, Yu T, Kang L et al. Childhood neurodevelopmental disorders and maternal hypertensive disorder of pregnancy. Dev Med Child Neurol 2021;63:1107–13. [DOI] [PubMed] [Google Scholar]
- 51. Lilienfeld AM, Pasamanick B. The association of maternal and fetal factors with the development of cerebral palsy and epilepsy. Am J Obstet Gynecol 1955;70:93–101. [DOI] [PubMed] [Google Scholar]
- 52. Norgaard M, Ehrenstein V, Nielsen RB et al. Maternal use of antibiotics, hospitalisation for infection during pregnancy, and risk of childhood epilepsy: a population-based cohort study. PLoS One 2012;7:e30850. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53. Sun Y, Vestergaard M, Christensen J et al. Prenatal exposure to maternal infections and epilepsy in childhood: a population-based cohort study. Pediatrics 2008;121:e1100-7–e1107. [DOI] [PubMed] [Google Scholar]
- 54. Vasistha NA, Pardo-Navarro M, Gasthaus J et al. Maternal inflammation has a profound effect on cortical interneuron development in a stage and subtype-specific manner. Mol Psychiatry 2020;25:2313–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55. Righes Marafiga J, Vendramin Pasquetti M, Calcagnotto ME. GABAergic interneurons in epilepsy: more than a simple change in inhibition. Epilepsy Behav 2021;121:106935. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Parts of this material are based on data and/or information compiled and provided by IRCC to 2020. However, the analyses, conclusions, opinions, and statements expressed in the material are those of the authors and not necessarily those of IRCC. The dataset from this study is held securely in coded form at ICES. ICES maintains a secure and comprehensive data repository containing individual-level, coded, and linkable health data for the population of Ontario, Canada. These data are routinely collected through administrative sources such as hospital-discharge abstracts, physician billing claims, drug-dispensing records, and vital statistics. Each Ontario resident has a unique encoded health identifier (based on the Ontario Health Insurance Plan, or OHIP number), which allows deterministic linkage across datasets within the ICES environment. This linkage is performed by using validated algorithm unique identifiers and is subject to rigorous data-quality assessments, including duplicate checks, consistency validations, and routine updates. In addition to Ontario’s Personal Health Information Protection Act (PHIPA) authorization (section 45), ICES operates under a stringent privacy and security governance framework: all policies are IPC-approved, regularly audited, and include measures such as data de-identification, controlled server access, encryption, audit logs, breach response protocols, and mandatory privacy training. While legal data-sharing agreements between ICES and data providers (e.g. healthcare organizations and government) prohibit ICES from making the dataset publicly available, access may be granted to those who meet pre-specified criteria for confidential access, available at www.ices.on.ca/DAS (e-mail: das@ices.on.ca).