Eclampsia—new onset of seizure after 20 weeks of gestation or the early postpartum period without an alternate etiology1,2—is a severe manifestation of the hypertensive disorders of pregnancy, occurring in about 8 of 10,000 deliveries.1,3,4 The imbalance in angiogenic and antiangiogenic circulating factors seen in hypertensive disorders of pregnancy5–7 may play a key role in the vascular endothelial dysfunction that leads to cerebral perfusion dysregulation and an eclamptic seizure.7–10 Until recently, eclampsia was thought to represent a monophasic episode of neurologic injury without long-term neurologic sequelae.11 Emerging evidence demonstrates that more than half of women with eclampsia have persistent neurocognitive dysfunction at 6 months postpartum12; some have white matter lesions on magnetic resonance imaging several years postpartum,11 and others demonstrate poor vision up to a decade after pregnancy.13 Women with eclampsia and the other hypertensive disorders of pregnancy are also at increased risk of premature stroke and transient ischemic attack.14
Several small studies suggest that women with eclampsia may be at elevated risk of a future seizure disorder, presumably as a result of cortical injury at the time of the eclamptic event.15–18 A clear understanding of the long-term risk of seizure disorder after eclampsia is important to women and their caregivers.
Accordingly, the Maternal Eclampsia and Long-Term Seizure study was designed with the purpose of determining the risk of a seizure disorder after eclampsia in addition to the other hypertensive disorders of pregnancy. The study hypothesis was that women with eclampsia would be at increased risk of a future seizure disorder and that risk would be higher than for the other hypertensive disorders of pregnancy.
MATERIALS AND METHODS
We completed a retrospective population-based cohort study using administrative data sets linked by unique encoded identifiers and analyzed at the Institute for Clinical Evaluative Sciences. We included all hospital births in the province of Ontario, Canada, between April 1, 2002, and March 31, 2014, accounting for more than 99% of births in Ontario during the study period. Females aged younger than 16 years or older than 50 years old and non-Ontario residents were excluded as were those with pre-existing epilepsy or a condition that could predispose to seizures such as a cerebral tumor, aneurysm or arteriovenous malformation, traumatic brain injury, stroke, or multiple sclerosis (Appendix 1, available online at http://links.lww.com/AOG/B34). Additionally, we excluded any patient who died during her index birth hospitalization or within 30 days of her delivery discharge date.
For each pregnancy, we determined whether the patient had a diagnosis of eclampsia, preeclampsia, or gestational hypertension diagnosed between 20 completed weeks of gestation and 30 days after the birth discharge date (ie, the primary exposure). Women with no hypertensive disorders of pregnancy served as the unexposed group (ie, the referent). If a given pregnancy had more than one hypertensive disorder of pregnancy diagnosis, a diagnosis of eclampsia superseded preeclampsia, which, in turn, superseded diagnosed gestational hypertension.
The primary study outcome was a hospitalization or emergency department visit for new-onset seizure disorder, defined as a seizure, seizure disorder, or epilepsy diagnosed more than 30 days after the index birth discharge date and not more than 20 weeks of gestation in a subsequent pregnancy (Appendix 1, http://links.lww.com/AOG/B34).
We used existing linked administrative health databases at the Institute for Clinical Evaluative Sciences, which capture all health encounters in the entire province of Ontario (http://www.ices.on.ca/Data-and-Privacy/ICES-data/Types-of-ICES-Data). All residents of Ontario are enrolled in the provincial Ontario Health Insurance Plan, which covers all aspects of antenatal care, delivery, and all postpartum care. The Canadian Institute for Health Information’s Discharge Abstract Database captures all hospitalizations using International Classification of Diseases, 9th Revision diagnostic codes before 2002 and International Classification of Diseases, 10th Revision, Canada diagnostic codes thereafter (Appendix 1, http://links.lww.com/AOG/B34). Emergency department visits were captured within the National Ambulatory Care Reporting System Database. The Canadian Institute for Health Information’s Discharge Abstract Database and National Ambulatory Care Reporting System data are entered by trained coders and abstractors who are considered highly trained coders and abstractors. The Canadian Institute for Health Information conducts regular data quality checks and reabstraction studies (eg, https://www.ices.on.ca/Publications/Atlases-and-Reports/2006/Canadian-Institute-for-Health-Information). The Ontario Health Insurance Plan Database was used for information on physician billing for outpatient, and some inpatient, services. The Registered Persons Database contains demographic information, including timing of death. Finally, Statistics Canada census data were used to classify income quintile and rural status.
Time-to-event analyses were conducted using Cox proportional hazards regression models with “time zero” starting at 31 days after the index birth discharge date. A participant was censored at the time at which she had an outcome event, reached the end of the study period, died while being event-free, or emigrated from the province.
The risk of a seizure disorder was expressed as an incidence rate as well as a hazard ratio (HR) and 95% CI comparing pregnancies affected by 1) eclampsia, 2) preeclampsia, or 3) gestational hypertension to 4) all unaffected pregnancies, which served as the referent. Because a given patient could contribute more than one pregnancy during the study period, any of eclampsia, preeclampsia, or gestational hypertension was treated as a time-varying exposure in the Cox proportional hazards models.19 We included the exposure in the model as an immediate (ie, on or off) exposure effect. In sensitivity analyses, we also considered a positive cumulative exposure effect and a first-time exposure effect.
The HRs were adjusted within three multivariable models. In the first, we adjusted for age, parity, income quintile, and multigestation birth at the time of birth (Appendix 1, http://links.lww.com/AOG/B34). The second model was further adjusted for diabetes mellitus, chronic hypertension, obesity as well alcohol, drug, and tobacco use disorders—any at the time of each index birth hospitalization—each treated as a time-varying covariate. In the third model, we further adjusted for a cerebral tumor, aneurysm or arteriovenous malformation, traumatic brain injury, stroke, and multiple sclerosis—occurring any time from greater than 30 days after any index birth date up to any censoring date, because these variables may be associated with the study outcome (seizure disorder). We completed an additional analysis, using the same aforementioned models, restricted to a first birth during the period of study.
All P values were two-sided at a significance level of .05. All statistical analyses were performed using SAS for Unix 9.4. This study was approved by the institutional review board at Sunnybrook Health Sciences Centre, Toronto, Canada.
Overall, 1,027,093 females delivered in an Ontario hospital during the study period. After the exclusion of 48,461 patients, 978,632 patients remained in the final cohort for the main analysis with a total of 1,565,733 births (Fig. 1). Of all 1,565,733 births, 1,615 (0.10%) were affected by eclampsia, 17,264 (1.1%) with preeclampsia, 60,863 (3.9%) with gestational hypertension, and 1,485,991 (94.9%) with none of these (Table 1). Pregnancies affected by hypertensive disorders of pregnancy had higher rates of preterm birth, poor fetal growth and stillbirth, maternal diabetes mellitus and obesity as well as a longer length of stay in the hospital at birth (Table 1).
The median duration of follow-up varied between 7.2 years for patients who had preeclampsia and 9.0 years for those who had eclampsia (Table 1) with a maximum follow-up of 12.9 years. A future seizure disorder was significantly more likely after a pregnancy with eclampsia (4.58/10,000 person-years) than a pregnancy without a hypertensive disorder of pregnancy (0.72/10,000 person-years), equivalent to a HR of 6.09 (95% CI 2.73–13.60) (Fig. 2; Table 2; Appendix 2 [Appendix 2 is available online at http://links.lww.com/AOG/B34]). The HR was unchanged after adjusting for age at baseline (HR 6.11, 95% CI 2.74–13.65). Full adjustment for all potential confounders resulted in slight attenuation (adjusted HR 5.42, 95% CI 2.42–12.12). The risk of a seizure disorder was approximately doubled in pregnancies affected by preeclampsia (adjusted HR 1.96, 95% CI 1.21–3.17), but not gestational hypertension (adjusted HR 1.01, 95% CI 0.71–1.43) (Fig. 2; Table 2). In sensitivity analyses, HRs were attenuated for the cumulative effect and first-time exposure effect (data not shown).
In the additional analysis limited to the first birth during the study period for each of the 978,632 patients, the rate of eclampsia (0.14%), preeclampsia (1.4%), and gestational hypertension (4.6%) (Appendix 3, available online at http://links.lww.com/AOG/B34) was higher than noted previously for all births. In this subset of patients, the adjusted HR risk of having a future seizure disorder was 2.77 (95% CI 1.04–7.40) after eclampsia, 1.73 (95% CI 1.09–2.73) after preeclampsia, and 0.94 (95% CI 0.67–1.33) after gestational hypertension (Appendix 4, available online at http://links.lww.com/AOG/B34).
We observed a six times higher risk of a future seizure disorder among pregnancies affected by eclampsia, but at an absolute risk of only 4.58 events per 10,000 person-years. After preeclampsia, the relative risk was approximately doubled, and for gestational hypertension, the risk was not higher than in an unaffected pregnancy.
The current findings support those noted from a limited number case reports and case series.15–18 One small case series contacted 13 patients with prior eclampsia by phone up to 6 months postpartum, of which one patient had recurrent seizures.15 In a U.S. prospective cohort study of 223 patients with eclampsia who were followed for at least 2 years postpartum, none were reported to have a seizure disorder.20 One possible explanation for the discrepancy between the latter findings and ours was their exclusion of patients who died or who could not be followed for at least 2 years postpartum. Second, given that the number of cases of eclampsia in their study (n=223) was one seventh that in ours (n=1,615), and we only had six seizure disorder outcome events, it is doubtful that they had sufficient statistical power to reveal this rare outcome.
The specific mechanism underlying why patients with eclampsia immediately seize is unknown as is the reason for why they may be prone to a seizure disorder in the long term.16,21 It is proposed that cerebral endothelial dysfunction and changes in vascular permeability result in cerebral edema at the time of severe preeclampsia, which may result in permanent cortical injury.16,21–23 Our study lends support to this theory, because patients with eclampsia were at the highest risk of a seizure disorder, those with preeclampsia were at somewhat elevated risk, whereas those with gestational hypertension—in whom cerebral edema and endothelial dysfunction is minimal—were at no higher risk.23 Vaisbuch et al found similar levels of antiangiogenic factors in patients with eclampsia compared with patients with severe preeclampsia, which have been shown to directly affect the cerebral vasculature in pregnancy.4,5,7 Women with eclampsia and severe preeclampsia also exhibit white matter changes on magnetic resonance imaging several years postpartum, although no study has preconception imaging for comparison.11,15,21,24 Certainly, evaluating the recovery of antiangiogenic factors in patients with preeclampsia and eclampsia, and its relation to brain magnetic resonance imaging findings, might improve our understanding of the mechanisms behind how these factors predispose patients to seizure activity—short term and long term.
We included all births from within an entire population of patients with universal health coverage, of whom 99% delivered in a hospital setting.25 Accordingly, study outcomes could be assessed in these patients after a reasonably long duration of follow-up. In addition, we used a time-varying exposure status to maximize the capture of the exposure, reflecting the reality that a hypertensive disorder of pregnancy may affect one pregnancy, but not another, within the same patient. This approach has an additional benefit of reducing the immortal time bias associated with observational studies by providing more accurate measures of risk.26 Our study variables were identified using International Classification of Diseases, 9th Revision and International Classification of Diseases, 10th Revision, Canada coding systems. A Canadian validation using administrative data like those here showed a sensitivity of 87.9% (95% CI 85.0–90.4) and a specificity of 99.6% for any hypertensive disorder of pregnancy.27 In the time period of the current study, the definitions of preeclampsia and eclampsia were essentially unchanged. Although the observed prevalence of preeclampsia was lower than expected, partly as a result of our hierarchical classification approach, this would not have affected the eclamptic main exposure group, whose 0.10% prevalence of eclampsia was as expected. Furthermore, we considered a diagnosis of eclampsia up to 30 days after the birth discharge date, although 90% of cases of postpartum eclampsia occur within 7 days of birth.28 Thus, the majority of exposure events in our study were likely sufficiently captured and classified.
As a result of the limited number of eclampsia cases and the rarity of the study outcome, we were unable to evaluate the interaction between eclampsia and other perinatal factors or chronic conditions on the risk of a seizure disorder. Although magnesium sulfate efficaciously reduces the risk of eclampsia in patients with preeclampsia,29 we did not possess information about its use or how it may have modified a patient’s future risk of seizure disorder. Finally, the study's findings may not be generalizable to populations with different demographics.
The Maternal Eclampsia and Long-Term Seizure study is informative for clinicians who can now confidently counsel and reassure patients with eclampsia that although their long-term relative risk of a seizure disorder is higher than patients without hypertensive disorder of pregnancy, the absolute risk is extremely low (approximately one seizure/2,200 person-years).
1. Sibai BM. Diagnosis, prevention, and management of eclampsia. Obstet Gynecol 2005;105:402–10.
2. Magee LA, Pels A, Helewa M, Rey E, von Dadelszen P; Canadian Hypertensive Disorders of Pregnancy (HDP) Working Group. Diagnosis, evaluation and management of the hypertensive disorders of pregnancy. Pregnancy Hypertens 2014;4:104–45.
3. Zwart JJ, Richters A, Ory F, de Vries JI, Bloemenkamp KW, van Roosmalen J. Eclampsia in the Netherlands. Obstet Gynecol 2008;112:820–7.
4. Liu S, Joseph KS, Liston RM, Bartholomew S, Walker MC, Leon JA, et al. Incidence, risk factors, and associated complications of eclampsia. Obstet Gynecol 2011;118:987–94.
5. Levine RJ, Lam C, Qian C, Yu KF, Maynard SE, Sachs BP, et al. Soluble endoglin and other circulating antiangiogenic factors in preeclampsia. N Engl J Med 2006;355:992–1005.
6. Levine RJ, Maynard SE, Qian C, Lim KH, England LJ, Yu K, et al. Circulating angiogenic factors and the risk of preeclampsia. N Engl J Med 2004;350:672–83.
7. Vaisbuch E, Whitty JE, Hassan SS, Romero R, Kusanovic JP, Cotton DB, et al. Circulating angiogenic and antiangiogenic factors in women with eclampsia. Am J Obstet Gynecol 2011;204:152.e1–9.
8. Wasseff S. Mechanisms of convulsions in eclampsia. Med Hypotheses 2009;72:49–51.
9. Macdonald RL. Soluble endoglin linking eclamptic women and brain vascular malformations? Ann Neurol 2009;66:2–4.
10. Cunningham FG, Twickler DM. Cerebral edema complicating eclampsia. Am J Obstet Gynecol 2000;182:94–100.
11. Aukes AM, de Groot JC, Aarnoudse JG, Zeeman GG. Brain lesions several years after eclampsia. Am J Obstet Gynecol 2009;200:504.e1–5.
12. Andersgaard AB, Herbst A, Johansen M, Borgstrom A, Bille AG, Øian P. Follow-up interviews after eclampsia. Gynecol Obstet Invest 2009;67:49–52.
13. Wiegman MJ, de Groot JC, Jansonius NM, Aarnoudse JG, Groen H, Faas MM, et al. Long-term visual functioning after eclampsia. Obstet Gynecol 2012;119:959–66.
14. Ray JG, Vermeulen MJ, Schull MJ, Redelmeier DA. Cardiovascular health after maternal placental syndromes (CHAMPS): population-based retrospective cohort study. Lancet 2005;366:1797–803.
15. Shah AK, Rajamani K, Whitty JE. Eclampsia: a neurological perspective. J Neurol Sci 2008;271:158–67.
16. Lawn N, Laich E, Ho S, Martin R, Faught E, Knowlton R, et al. Eclampsia, hippocampal sclerosis, and temporal lobe epilepsy: accident or association? Neurology 2004;62:1352–6.
17. Klingman WO, Suter C. Seizure states and pregnancy. Neurology 1957;7:105–18.
18. Sexton JA. Epilepsy as a sequel of obstetrical complications. J Ky Med Assoc 1976;74:595–8.
19. Allison PD. Survival analysis using SAS: a practical guide. 2nd ed. Cary (NC):SAS Institute; 2010.
20. Sibai BM, Sarinoglu C, Mercer BM. Eclampsia VII. Pregnancy outcome after eclampsia and long-term prognosis. Am J Obstet Gynecol 1992;166:1757–61.
21. Zeeman GG. Neurologic complications of pre-eclampsia. Semin Perinatol 2009;33:166–72.
22. Solinas C, Briellmann R, Harvey AS, Mitchell L, Berkovic S. Hypertensive encephalopathy: antecedent to hippocampal sclerosis and temporal lobe epilepsy? Neurology 2003;60:1534–6.
23. Melamed N, Ray JG, Hladunewich M, Cox B, Kingdom JC. Gestational hypertension and preeclampsia: are they the same disease? J Obstet Gynaecol Can 2014;36:642–7.
24. Aukes AM, de Groot JC, Wiegman MJ, Aarnoudse JG, Sanwikarja G, Zeeman GG. Long-term cerebral imaging after pre-eclampsia. BJOG 2012;119:1117–22.
25. Ray JG, Schull MJ, Kingdom JC, Vermeulen MJ. Heart failure and dysrhythmias after maternal placental syndromes: HAD MPS Study. Heart 2012;98:1136–41.
26. Liu J, Weinhandl ED, Gilbertson DT, Collins AJ, St Peter WL. Issues regarding ‘immortal time’ in the analysis of the treatment effects in observational studies. Kidney Int 2012;81:341–50.
27. Joseph KS, Fahey J; Canadian Perinatal Surveillance System. Validation of perinatal data in the Discharge Abstract Database of the Canadian Institute for Health Information. Chronic Dis Can 2009;29:96–100.
28. Al-Safi Z, Imudia AN, Filetti LC, Hobson DT, Bahado-Singh RO, Awonuga AO. Delayed postpartum preeclampsia and eclampsia: demographics, clinical course, and complications. Obstet Gynecol 2011;118:1102–7.
29. Altman D, Carroll G, Duley L, Farrell B, Moodley J, Neilson J, et al. Do women with pre-eclampsia, and their babies benefit from magnesium sulphate? The Magpie Trial: a randomised placebo-controlled trial. Lancet 2002;359:1877–90.