Scott, Catherine A. BMBCh, MSc; Bewley, Susan MA, MD; Rudd, Anthony FRCP (London); Spark, Patsy BSc; Kurinczuk, Jennifer J. MBChB, MD; Brocklehurst, Peter MBChB, MSc; Knight, Marian MBChB, DPhil
Stroke is an important cause of morbidity and mortality, the incidence of which is likely to increase as a result of an ageing population. For many years it was believed that the risk of stroke was significantly higher during pregnancy. However, controversy now exists. Previous estimates of incidence of stroke associated with pregnancy vary widely (from four to 40 per 100,000 deliveries)1–12 and the quality of studies undertaken has varied.13 In early series, imaging was not available and distinction of subtypes was not clear. Larger studies, based on administrative data sets, are subject to coding errors and have been unable to collect information on individual cases or timings of stroke. Smaller studies, often recruiting from specialist referral centers where incidence is higher and estimates of the denominator population may be inaccurate, are subject to selection bias. Furthermore, comparing results of studies is problematic because groups have been defined differently. Some include antenatal strokes only and others also include some or all of the cases in the postpartum period.
Studying stroke incidence and outcome is challenging in any population.14 Like with other studies of rare disorders of pregnancy, it is difficult to identify a sufficiently large sample. The UK Obstetric Surveillance System is a resource for studying rare disorders of pregnancy. This study was designed to estimate the incidence of antenatal stroke in the United Kingdom and to identify risk factors associated with stroke during pregnancy. Information on the clinical features, current management, survival, and prognosis of antenatal strokes was also obtained to develop guidance and improve the care of women having an antenatal stroke.
MATERIALS AND METHODS
A population-based (nationwide) cohort and nested case–control study was conducted using the UK Obstetric Surveillance System between October 2007 and March 2010. The UK Obstetric Surveillance System methods have been described in detail elsewhere.15 Nominated clinicians in each consultant-led maternity unit in the United Kingdom were sent a card each month and asked to report all cases of antenatal stroke, thus covering the entire cohort of U.K. births in this 2.5-year period. Clinicians who reported a case were asked to complete data-collection forms for the case and up to two control patients, detailing demographic and other potential risk factors, treatment, and outcomes. Up to five reminders were sent if completed forms were not returned. All data requested were anonymous. On receipt of data-collection forms, cases were checked to confirm that they met the case definition. Duplicate reports were identified by comparing the woman's year of birth, date of stroke, hospital, and expected date of delivery.
Cases were defined as any pregnant woman who had an antenatal stroke and delivered between October 2007 and March 2010. The World Health Organization definition of stroke was used16; cases were required to be confirmed using suitable imaging (computed tomography, magnetic resonance imaging plus or minus lumber puncture for subarachnoid hemorrhage) or at postmortem. Control patients were defined as the women delivering immediately before the reported case in the same hospital who had not had a stroke. Identical data were collected from both case and control patients except for details of the stroke itself.
Putative risk factors for stroke were identified from published reviews and case series. Data were collected on the type of stroke, demographic and other potential risk factors, management, and outcomes.
Data about maternal deaths with antenatal stroke up to December 2008 were obtained from the Centre for Maternal and Child Enquiries and compared with maternal deaths reported through the UK Obstetric Surveillance System. No additional cases were identified through this source. Centre for Maternal and Child Enquiries mortality data for 2009 and 2010 were not available for comparison.
The main aim of the study was to estimate the incidence of stroke during pregnancy; therefore, the power of the study was limited by the observed incidence. This analysis had 80% power to detect odds ratios (ORs) between 3.5 or greater and 6.8 or greater at the 5% level of statistical significance, assuming a range of prevalence for the prospective risk factors of between 30% and 5%, respectively.
Incidence rates with 95% confidence intervals (CIs) were calculated using the most recently available maternity data for October 2007 to March 2010.17–19 Simple associations were assessed with Fisher's exact tests for two independent proportions. Means and medians (interquartile range) were used to summarize continuous variables and were compared with unpaired t tests. Odds ratios were estimated throughout using exact logistic regression. Exact methods were used as a result of the small numbers of cases with certain risk factors, which could not be assumed to be normally distributed. Continuous variables such as age were used as continuous variables for modeling and for simplicity as dichotomous variables for presentation.
A parsimonious regression model was developed by including explanatory factors in a core model if there was a pre-existing hypothesis or evidence to suggest that they were causally related to antenatal stroke, for example, maternal age. Potential confounding factors (for example, ethnicity, marital status) were then added to the core model and removed in a backward stepwise manner with the results examined as each variable was removed; factors that did not contribute significantly to the fit of the model on likelihood ratio testing (P<.05) were excluded. Each variable left in the final parsimonious model was tested for interaction, continuous variables were tested for departure from linearity by the addition of quadratic terms to the model and subsequent likelihood ratio testing, and model-checking procedures were performed. As a result of the small number of cases in the hemorrhagic and nonhemorrhagic subgroups, there was insufficient power to carry out a multivariable analysis; therefore, unadjusted ORs were calculated only. All analyses were carried out using STATA 11 software. The UK Obstetric Surveillance System general methodology (04/MRE02/45) and this study (07/H0718/54) were approved by the London Multi-center Research Ethics Committee.
All 229 eligible U.K. hospitals contributed data to the UK Obstetric Surveillance System during the study period (100% response). Seventy-eight cases of stroke were reported; data collection was complete for 83% (Fig. 1).
During the period October 2007 to March 2010, there were 30 confirmed cases of antenatal stroke in an estimated 1,958,203 women delivering.17–19 Of the women whose cases were excluded, one did not meet the case definition and 31 had a postpartum stroke. Data were collected on 89 control patients. The overall estimated incidence of antenatal stroke was 1.5 cases per 100,000 women delivering (95% CI 1.0–2.1). The incidences of nonhemorrhagic and hemorrhagic stroke were 0.9 (95% CI 0.5–1.3) and 0.6 (95% CI 0.3–1.0) per 100,000 women delivering, respectively. Eighteen (60%) strokes were nonhemorrhagic and 12 (40%) were hemorrhagic. Reported causes are summarized in Table 1. All cases except one were first-ever strokes (one woman who had an ischemic stroke during pregnancy had previously had a cerebral venous thrombosis).
The median gestational age at the time of all strokes was 30 2/7 weeks (n=29, interquartile range 14–38) and was similar whether hemorrhagic or not. Figure 2 shows the timing of each stroke during pregnancy according to cause.
Basic demographic characteristics and risk factors for case and control patients are shown in Table 2. Women with an antenatal stroke were significantly older than control women with 33% aged 35 years or older compared with 15%, respectively (P=.03). Women having hemorrhagic stroke tended to be older than women with nonhemorrhagic stroke (n=6, 50% aged 35 years or older compared with n=4, 22%, P=.14). There were no multiple gestations. In terms of classical risk factors for stroke, none of the case or control patients were known to have pre-existing atrial fibrillation or ischemic heart disease. Two women who had ischemic strokes were known to have cardiac septal defects. Two of the women who had ischemic stroke (7%) and one control patient (1%) were reported to have a family history of stroke.
After adjustment, history of migraine (adjusted OR 8.5, 95% CI 1.5–62.1), gestational diabetes (adjusted OR 26.8, 95% CI 3.2–∞), and preeclampsia or eclampsia (adjusted OR 7.7, 95% CI 1.3–55.7) remained statistically significant risk factors for all antenatal stroke (Table 2). After adjustment for age, the risk of stroke increased by 8% (adjusted OR 1.08, 95% CI 1.03–1.13) for every mm Hg increase in the highest recorded diastolic blood pressure (BP) during pregnancy and by 3% (adjusted OR 1.03, 95% CI 1.00–1.05) for every mm Hg increase in the highest recorded systolic BP. None of the women were reported to have had an acute hypotensive episode before their stroke (watershed infarct), and none were preceded by a transient ischemic attack.
Univariable analysis revealed hemorrhagic strokes to be more strongly associated with maternal age, occupation, and hypertensive disorders including eclampsia and preeclampsia, whereas nonhemorrhagic strokes were more strongly associated with a history of migraine. Both were associated with gestational diabetes.
There were no significant differences in clinical features distinguishing hemorrhagic from nonhemorrhagic strokes. Of the ischemic strokes, 87% presented with weakness, 50% with speech disturbance, 44% had headache, 38% had visual disturbance, and 30% had disturbed consciousness or collapse. Of the hemorrhagic stroke 55% had weakness, 55% headache, 27% visual disturbance, 37% had disturbed consciousness or collapse, and 33% presented with a seizure. All three women with subarachnoid hemorrhage presented with disturbed consciousness or collapse.
As shown in Table 3, treatment regimens varied. Of the 12 women with ischemic arterial stroke, none received thrombolysis, although most (9 of 12 [75%]) were started on an antiplatelet agent. Three (25%) received no antiplatelet or anticoagulant medication. These women received treatment for preeclampsia with antihypertensive medications and magnesium sulfate. Of the 30 cases of stroke, six were admitted to an acute stroke unit with the remainder being cared for in maternity, medical or neurology wards, and intensive care units.
All three women with cerebral venous thrombosis received an anticoagulant agent. In addition, one received aspirin. Two were given phenytoin. Of the two cases secondary to arterial dissection, one received a therapeutic (high) dose of aspirin and unfractionated heparin, whereas the other received low molecular weight heparin only. Of the nine women with intracerebral hemorrhage, five underwent decompressive neurosurgery, three received antihypertensive medications, and three received magnesium sulfate. One woman with subarachnoid hemorrhage underwent endovascular coiling of an aneurysm.
Sixty-two percent of women who had a stroke (n=16) delivered by cesarean compared with 19% of comparison women (P<.001). Cesarean delivery was performed in 82% of women with hemorrhagic strokes and 47% of those with nonhemorrhagic (P=.1), most commonly (89%) for maternal compromise.
There were six stroke-related maternal deaths, four as a result of intracerebral hemorrhage and two as a result of subarachnoid hemorrhage. All women with nonhemorrhagic strokes survived. This gives a case-fatality rate of 20% (95% CI 4–36%) of all strokes and 50% (95% CI 10–90%) of hemorrhagic strokes and a mortality rate as a result of antenatal stroke of 0.3 (95% CI 0.1–0.6) per 100,000 women delivering. The median time from stroke to death was 5 days (range 0–85). Of the six women who died, three delivered liveborn neonates after the stroke, but before death, and three neonates did not survive.
Eight women with hemorrhagic (67%) and three women with nonhemorrhagic strokes (25%) were admitted to an intensive care unit (median duration 3 days, interquartile range 2–7). Data on disability at discharge from obstetric services (modified Rankin Scale score) were only provided for 20 of 30 women. Of those with data, 2 of 12 (17%) women having a nonhemorrhagic stroke had a modified Rankin Scale score of 3 or greater indicating moderate to severe disability compared with seven of eight (88%) of the hemorrhagic cases (P=.005). Among the survivors, information at discharge was available for 21 of 24 women. Twelve of 15 (80%) nonhemorrhagic cases were discharged home compared with only three of six (50%) of surviving hemorrhagic cases (P=.3). The remaining six (30% of total survivors) were discharged to rehabilitation facilities or for further care elsewhere.
In terms of outcome of pregnancy, five resulted in fetal or neonate death (as a result of termination of pregnancy, stillbirth, or neonate mortality), all in women with hemorrhagic strokes. Eight newborns (36%) were admitted to the neonatal intensive care unit compared with 7% of neonates born to women in the control group (P<.001). Antenatal stroke was associated with premature delivery; the median gestation at delivery for cases was 39 weeks of gestation (interquartile range 37–40) compared with 40 weeks of gestation in control patients (interquartile range 39–41; P=.01). Eight of 25 (32%) cases had premature delivery (less than 37 weeks of gestation) compared with 7% of control patients (P=.002) and 16% of women who had stroke had a very premature delivery (less than 32 weeks of gestation) compared with none of the control patients (P=.002).
The incidence of antenatal stroke as estimated by this study is 1.5 cases per 100,000 women delivering. Although the condition is rare, the case-fatality rate was 20% with 45% of survivors having significant disability at discharge and 30% requiring transfer to rehabilitation facilities. We used the “gold standard” core criteria for stroke incidence studies using the World Health Organization definition of stroke.14 It is frequently reported that the incidence of stroke is higher during pregnancy. However, most previous studies report both antenatal and puerperal strokes together or do not provide population-based data on gestational strokes.2,7,8 Recent evidence suggests that pregnancy itself is associated with a relatively low risk of stroke, whereas the risks are higher in the postpartum period. Kitner et al4 reviewed both pregnant and nonpregnant female hospital discharges and found the relative risk of cerebral infarction and intracerebral hemorrhage during pregnancy to be 0.7 and 2.5, respectively. Our results appear to agree; the incidence of antenatal stroke found in this study is lower than recently published rates in nonpregnant women of childbearing age in the United Kingdom (3.6 ischemic strokes per 100,000 person-years).20 It is, however, possible our lower incidence is the result of underascertainment; for example, minor strokes or those that occurred very early in pregnancy would be unlikely to have involved obstetric services and thus were unlikely to be reported to the UK Obstetric Surveillance System. The higher mortality rate than reported elsewhere3 may suggest a subset of more severe strokes may have been captured in our study. Conversely, low case-fatality reported elsewhere5 may be the result of the overinclusion of noncases.
As a result of the rarity of antenatal stroke and despite 30 months of national data collection, we had limited power to analyze risk factors as a result of small numbers. Nevertheless, in agreement with previous studies,3 an association was found between stroke and preeclampsia and eclampsia, gestational diabetes, and migraine. Recent evidence suggests that variability of BP may be more important than mean BP on the risk of stroke in the nonobstetric population,21 and the increased risk of both hemorrhagic and ischemic strokes with preeclampsia and eclampsia illustrate that acute increases in BP can cause cerebral hemorrhage and ischemic stroke. We found the highest risk of antenatal stroke to be in the third trimester (Fig. 2) and this is likely to reflect the association with preeclampsia and eclampsia and the timings of these conditions. Clinicians should be aware that this time period carries the highest risk and of the associations of antenatal stroke with diabetes and preeclampsia as the risk profile of the childbearing population continues to change.
We also found a history of migraine was associated with antenatal stroke, although subgroup analysis found the association appeared to persist for nonhemorrhagic but not hemorrhagic strokes. A recent database study based on discharge diagnostic codes found a strong association between migraines occurring during pregnancy and stroke during pregnancy or the puerperium (OR 15.05, 8.26–27.4) and in particular for ischemic stroke (OR 30.7, 17.4–34.1).22 The mechanism for this association is unknown, but the authors suggest that a subset of women with migraine is predisposed to poor compensation to vascular stresses such as those during pregnancy. Although this association is potentially important, it should be interpreted with caution; the database study22 and ours both report a relatively low prevalence of migraine in control patients compared with rates reported in the literature. In agreement with a previous study,3 antenatal stroke was associated with gestational diabetes (adjusted OR 26.8, 95% CI 3.2–∞). However, because there was only a small number of associated cases, we cannot exclude the possibility that this was a chance finding. Outside of pregnancy, insulin resistance is associated with an increased risk for stroke,23 but it also is plausible that treatments started during pregnancy increase the risk of stroke rather than gestational diabetes per se.
Although there are case reports of thrombolysis during pregnancy,24 no woman in this study received thrombolysis, presumably because pregnancy is one of the contraindications to licensed treatment with thrombolysis. Although there is currently no consensus on use of antiplatelet and antithrombotic agents in pregnancy,25 it is widely believed that aspirin is safe after 12 weeks of gestation. However, just 8 of 12 (67%) women with ischemic arterial strokes were treated with aspirin in this study, whereas 42% received an anticoagulant, which is only recommended for acute venous stroke.
In nonpregnant populations, hemorrhagic strokes account for approximately 22% of strokes.26 In our study, in agreement with others studying obstetric populations,3,7,8 we found a higher proportion (40%) of strokes was hemorrhagic. As shown in other studies,12 the prognosis of hemorrhagic strokes was worse than for nonhemorrhagic strokes. The outcomes appear to be poorer for pregnant women (case-fatality rate of 20% of all strokes and 50% of hemorrhagic strokes) than those reported for nonpregnant populations.27 It is plausible that the physiologic changes associated with pregnancy make the prognosis worse for pregnant women, or the high case-fatality rate may be related to reluctance to provide specialist stroke treatments. It is important that pregnant women with stroke receive specialist stroke care for optimal monitoring and rehabilitation, yet only 6 of 30 women in this study were treated on an acute stroke unit. U.K. national guidelines state that all patients with an acute stroke should be admitted directly to a specialist stroke unit after initial assessment28; clearly this may be less practical in the obstetric population but because high-quality care in acute stroke units is known to improve outcomes,29 it is vital that pregnant women are given the best treatment possible. An alternative explanation for the high case-fatality rate is that our study picked up only a severe subset of strokes; larger database studies have reported lower mortality rates (2.7–3.0%) but included patients with transient ischemic attacks.5 A study with overlapping notification of sources to ensure complete ascertainment would address this issue and enable comprehensive study of both antenatal and postnatal strokes; advancement in computerized medical record linkage systems may aid the study of pregnancy-related stroke incidence in the future.
1. Cross JN, Castro PO, Jennett WB. Cerebral strokes associated with pregnancy and the puerperium. Br Med J 1968;3:214–8.
2. Jaigobin C, Silver FL. Stroke and pregnancy. Stroke 2000;31:2948–51.
3. James AH, Bushnell CD, Jamison MG, Myers ER. Incidence and risk factors for stroke in pregnancy and the puerperium. Obstet Gynecol 2005;106:509–16.
4. Kittner SJ, Stern BJ, Feeser BR, Hebel R, Nagey DA, Buchholz DW, et al.. Pregnancy and the risk of stroke. N Engl J Med 1996;335:768–74.
5. Kuklina EV, Tong X, Bansil P, George MG, Callaghan WM. Trends in pregnancy hospitalizations that included a stroke in the United States from 1994 to 2007: reasons for concern? Stroke 2011;42:2564–70.
6. Lanska DJ, Kryscio RJ. Risk factors for peripartum and postpartum stroke and intracranial venous thrombosis. Stroke 2000;31:1274–82.
7. Liang CC, Chang SD, Lai SL, Hsieh CC, Chueh HY, Lee TH. Stroke complicating pregnancy and the puerperium. Eur J Neurol 2006;13:1256–60.
8. Ros HS, Lichtenstein P, Bellocco R, Petersson G, Cnattingius S. Pulmonary embolism and stroke in relation to pregnancy: how can high-risk women be identified? Am J Obstet Gynecol 2002;186:198–203.
9. Simolke GA, Cox SM, Cunningham FG. Cerebrovascular accidents complicating pregnancy and the puerperium. Obstet Gynecol 1991;78:37–42.
10. Wang KC, Chen CP, Yang YC, Wang KG, Hung FY, Su TH. Stroke complicating pregnancy and the puerperium. Zhonghua Yi Xue Za Zhi (Taipei) 1999;62:13–9.
11. Wiebers DO, Whisnant JP. The incidence of stroke among pregnant women in Rochester, Minn, 1955 through 1979. JAMA 1985;254:3055–7.
12. Sharshar T, Lamy C, Mas JL. Incidence and causes of strokes associated with pregnancy and puerperium. A study in public hospitals of Ile de France. Stroke in Pregnancy Study Group. Stroke 1995;26:930–6.
13. Grosset DG, Ebrahim S, Bone I, Warlow C. Stroke in pregnancy and the puerperium: what magnitude of risk? J Neurol Neurosurg Psychiatry 1995;58:129–31.
14. Feigin VL, Carter K. Editorial comment—Stroke incidence studies one step closer to the elusive gold standard? Stroke 2004;35:2045–7.
15. Knight M, Kurinczuk JJ, Tuffnell D, Brocklehurst P. The UK Obstetric Surveillance System for rare disorders of pregnancy. BJOG 2005;112:263–5.
16. Hatano S. Experience from a multicentre stroke register: a preliminary report. Bull World Health Organ 1976;54:541–53.
17. Office for National Statistics birth summary tables, England and Wales, 2010. Available at: www.ons.gov.uk
. Retrieved November 2011.
20. Nightingale AL, Farmer RD. Ischemic stroke in young women: a nested case-control study using the UK General Practice Research Database. Stroke 2004;35:1574–8.
21. Rothwell PM. Limitations of the usual blood-pressure hypothesis and importance of variability, instability, and episodic hypertension. Lancet 2010;375:938–48.
22. Bushnell CD, Jamison M, James AH. Migraines during pregnancy linked to stroke and vascular diseases: US population based case–control study. BMJ 2009;338:b664.
23. Kernan WN, Inzucchi SE, Viscoli CM, Brass LM, Bravata DM, Horwitz RI. Insulin resistance and risk for stroke. Neurology 2002;59:809–15.
24. Johnson DM, Kramer DC, Cohen E, Rochon M, Rosner M, Weinberger J. Thrombolytic therapy for acute stroke in late pregnancy with intra-arterial recombinant tissue plasminogen activator. Stroke 2005;36:e53–5.
25. Helms AK, Drogan O, Kittner SJ. First trimester stroke prophylaxis in pregnant women with a history of stroke. Stroke 2009;40:1158–61.
26. O'Donnell MJ, Xavier D, Liu L, Zhang H, Chin SL, Rao-Melacini P, et al.. Risk factors for ischaemic and intracerebral haemorrhagic stroke in 22 countries (the INTERSTROKE study): a case–control study. Lancet 2010;376:112–23.
27. Rothwell PM, Coull AJ, Giles MF, Howard SC, Silver LE, Bull LM, et al.. Change in stroke incidence, mortality, case-fatality, severity, and risk factors in Oxfordshire, UK from 1981 to 2004 (Oxford Vascular Study). Lancet 2004;363:1925–33.
28. NICE (2008) clinical guideline 68. Stroke: diagnosis and initial management of acute stroke and transient ischaemic attack (TIA). Available at: www.nice.org.uk/CG68
29. Middleton S, McElduff P, Ward J, Grimshaw JM, Dale S, D'Este C, et al.. Implementation of evidence-based treatment protocols to manage fever, hyperglycaemia, and swallowing dysfunction in acute stroke (QASC): a cluster randomised controlled trial. Lancet 2011;378:1699–706.