Preventability of stroke during pregnancy and postpartum

Abstract

Background

The preventability of strokes during pregnancy and postpartum is understudied.

Methods

We identified patients with ischemic and hemorrhagic strokes in pregnancy or within 12 weeks postpartum between 2009 and 2020 at an obstetric teaching hospital. We described the clinical course of the included cases. A multi-disciplinary panel adjudicated each occurrence of stroke according to a modified classification system for preventability.

Results

Fifteen patients experienced a stroke during pregnancy or postpartum. Precisely, 60% presented with a headache, and 47% of events were complicated by severe obstetrical hypertension. Two cases were thought to be “possibly preventable” due to delays in presentation to care and miscommunication regarding hypertension management goals.

Conclusion

In a cohort of pregnant and postpartum patients with stroke at a tertiary-care center, strokes may be prevented by future initiatives focusing on patient and provider education regarding early warning signs of pregnancy-associated stroke and hypertension thresholds and management specific to pregnancy.

Background

Maternal stroke, together with heart disease, accounts for over a third of maternal deaths in the United States (US), and has been on the rise.^1,2 It is estimated that stroke complicates 30 per 100,000 pregnancies in the US, with about half occurring during the antepartum or peri-delivery period and half during the postpartum period. ³ Due to several physiologic changes, including venous stasis, hypercoagulability, and hemodynamic alterations, pregnant patients have a threefold higher risk of stroke than the non-pregnant population.^4–8 Furthermore, whereas the distribution of stroke in the general population in the US is ∼ 87% ischemic and 13% hemorrhagic, in the pregnant population, up to 66% of strokes are hemorrhagic, further underscoring unique risk factors, pathophysiology, and management challenges of stroke in pregnancy.^3,9 The multifaceted sequelae of stroke can be devastating and efforts to prevent maternal stroke are urgently required.

Previous studies examining the preventability of maternal strokes have focused on state or national-level data on maternal deaths. One study in California examining maternal deaths due to stroke from 2002 to 2007 reported that 66% of preeclampsia-related stroke deaths (n = 33) were possibly preventable. ¹⁰ Severe range blood pressures (BPs) (systolic BP (SBP) ≥ 160 mmHg or diastolic BP (DBP) ≥ 110 mmHg) preceded 25/26 cases of fatal stroke, but only 6/31 patients had received first-line antihypertensive treatment as recommended by the American College of Obstetricians and Gynecologists (ACOG). ¹⁵ Of the 33 cases, 30 experienced providers delayed recognition of and response to clinical warning signs. ¹⁰ Another study in Japan assessed national data from 2010 to 2014 and found that 34% of pregnancy-related stroke deaths (n = 35) were possibly preventable, with 33% of preventable cases attributed to inadequate BP management. ¹¹ While these studies highlighted preventable etiologies of stroke-related maternal mortality, the preventability of maternal strokes where death did not occur needs to be described.

Improved treatment of hypertension in pregnancy may reduce the risk of stroke. ACOG guideline updates from 2013 to 2020 have included treatment of severely elevated SBP, defined first-line pharmacologic treatment, and highlighted time to acute treatment as within 30–60 min.^12–15 With these significant changes in obstetrical hypertension treatment recommendations in the last decade, as well as increasing focus on reviewing the preventability of severe maternal morbidity,^16,17 the preventability of maternal strokes must be evaluated.

Thus, the aim of this study was to assess the preventability of stroke and transient ischemic attack (TIA) during pregnancy and postpartum in a contemporary US cohort and to identify factors to be addressed to prevent pregnancy-associated stroke going forward.

Methods

Study population

Our study was conducted at a single urban tertiary obstetric teaching hospital, caring for approximately 8500 deliveries annually, a substantial portion of which are considered high-risk. As a facility specializing in obstetrics and gynecology, this hospital has a stand-alone emergency room where obstetricians are the first to evaluate and stabilize patients with acute conditions. The treatment of severe hypertension is protocolized and in line with recommendations published at the time (i.e. most recently the ACOG evaluation and management algorithm updated in 2020). ¹⁵ Currently, trained medical staff perform BP measurements via standardized technique. Any BPs where SBP ≥ 160 mmHg or DSP ≥ 110 mmHg are considered a severe episode of hypertension and are immediately reported to a physician or advanced practice provider. If persistently severe for ≥15 min, intravenous labetalol, intravenous hydralazine, or immediate-release oral nifedipine are administered. BPs are repeated every 10–20 min and treated with additional antihypertensive medications until they are no longer severely elevated. In addition, magnesium sulfate infusion is typically initiated for standard eclampsia prophylaxis indications. Furthermore, there is full-time access to diagnostic imaging, as well as Obstetric Internal Medicine specialists on-site to help guide the diagnosis and management of stroke in pregnancy and postpartum. Other adult medical specialty services, including neurology, neurosurgery, and interventional radiology, are in a neighboring medical-surgical hospital.

To construct our cohort, we first identified individuals with International Classification of Diseases (ICD-9 and ICD-10) codes consistent with a possible ischemic or hemorrhage stroke or TIA between 1 January 2009 and 10 March 2020 (Supplemental Table 1). We included both ischemic and hemorrhagic strokes as well as TIA in our cohort, as we anticipated best practices to prevent maternal ischemic and hemorrhagic stroke would likely overlap. We then manually reviewed the electronic medical record (EMR) of each potential case and included participants who were pregnant or within 12 weeks postpartum, with clinical findings consistent with stroke or TIA as documented in the medical chart by the clinical team. The postpartum cutoff of 12 weeks was chosen to account for pregnancy-related hypercoagulability. ¹⁸ Individuals were excluded if they were not pregnant or not within 12 weeks postpartum at the time of the event, or if they did not have a stroke nor a TIA.

Case description

A senior resident physician in obstetrics and gynecology abstracted detailed information about each included case from the EMR system. Baseline characteristics of interest included maternal age, gravidity and parity, chart-documented race and ethnicity, relevant obstetric, medical, surgical, family, and social history, and medications. Clinical characteristics included gestational age or days postpartum at the time of stroke, symptomatology, vital signs, exam findings at the time of presentation, laboratory and radiologic findings at the time of presentation, pharmacologic or surgical treatment for stroke, clinical course as each case evolved, and post-stroke follow-up, as available.

Statistical analysis and assessment of preventability

Descriptive statistics were used to summarize clinical characteristics. An expert panel comprising a Maternal-Fetal Medicine specialist, a Neurologist, and an Obstetric Internal Medicine specialist then assessed each case for preventability, which was defined as “not likely preventable,” “possibly preventable,” “probably preventable,” or “likely preventable,” a classification system which was modified from the established Harvard Medical Practice Study tool, originally developed to provide more optionality than a straightforward yes-or-no preventability dichotomy. ¹⁷ We adapted our definition of “preventability” from the Centers for Disease Control and Prevention Morbidity and Mortality Weekly Report as “[an event] is considered preventable if the committee determines that there was some chance of the [event] being averted by one or more reasonable changes to patient, community, provider, health facility, and/or system factors.”^19,20 In cases that were considered possibly preventable, the most likely factor, such as “patient,” “provider,” “system,” or “other” was determined. “Patient-related” factors included delayed presentation to care and documented nonadherence with medical recommendations, “provider-related” factors included diagnostic and treatment errors, and “systems-related” factors included failure to abide by established protocols or suboptimal communication between care teams. ¹⁷ Moreover, we adopted a systematic and conservative approach in our adjudication of cases—for example, if a patient with a stroke presented with severe hypertension and focal neurologic symptoms, we only considered the event preventable if a specific reventable element was documented in the EMR, and we did not assume a natural sequence of events based on probability. In cases where there was disagreement among two multidisciplinary panels regarding preventability, the third-panel member was adjudicated by majority expert consensus.

Results

A total of 2663 patients screened positive by ICD codes, of which 15 were found to have a stroke or TIA during pregnancy or within the first 12 weeks postpartum. Patients excluded largely represented a non-pregnant gynecologic cohort of patients receiving care for benign or malignant gynecologic conditions. 7/15 (47%) of strokes occurred antepartum, 2/15 (13%) peri-delivery, and 6/15 (40%) during the postpartum period (range 0–12 days) (Table 1). Among cases, the average age at diagnosis was 29 years (range 16–38 years). 4/15 (27%) participants reported a migraine history. 6/15 (40%) strokes were ischemic and 8/15 (53%) were hemorrhagic, with one case of TIA. Furthermore, 8/15 (53%) cases were complicated by hypertension (mild hypertension defined as SBP ≥ 140 mmHg or DBP ≥ 90 mmHg and severe hypertension defined as SBP ≥ 160 mmHg or DBP ≥ 110 mmHg) and 9/15 (60%) patients presented with a headache, most commonly severe and unrelenting. Most cases presenting with hypertension (7/8, 88%) and headache (7/8, 88%), suffered a hemorrhagic stroke (Table 1). There were no maternal deaths.

Table 1.

Clinical characteristics and preventability of strokes in pregnancy and postpartum.

No.	Timing	Relevant clinical information	Type of stroke	Max BP if hypertensive	Headache	Other neuro symptoms	Preventability
1	35w	Asthma, iron deficiency anemia, former smoker	Ischemic	–	–	Disorientation	Not likely
2	PPD1	Gestational carrier, family history of epilepsy	Ischemic	150/68	Mild	Visual floaters	Not likely
3	28w	Migraines	Hemorrhagic	200/120	Severe	Weakness, numbness	Possibly
4	41w	Asthma	TIA	–	–	Visual floaters	Not likely
5	PPD12	Depression, family history of seizures	Hemorrhagic	188/90	Severe	–	Not likely
6	PPD0	Mild thrombocytopenia stable throughout pregnancy	Hemorrhagic	160/103	–	Altered mental status	Possibly
7	PPD8	Migraines, ITP	Hemorrhagic	176/98	Severe	Photophobia	Not likely
8	39w	Depression, former smoker	Ischemic	–	–	Antegrade amnesia	Not likely
9	PPD11	Alpha-1-antitrypsin deficiency, fatty liver, anxiety, ADHD	Hemorrhagic	180/120	Severe	Weakness	Not likely
10	33w	Migraines	Ischemic	–	Severe	Weakness	Not likely
11	13w	Hepatitis C, history of PSUD, active tobacco use, asthma	Hemorrhagic	–	Severe	Right visual loss	Not likely
12	8w	History of neonatal demise due to genetic condition	Ischemic	–	–	Altered mental status, weakness	Not likely
13	PPD3	Hashimoto's thyroiditis, depression	Ischemic	–	–	Weakness, numbness	Not likely
14	29w	Anxiety, iron deficiency anemia	Hemorrhagic	166/78	Severe	photophobia	Not likely
15	38w	Migraines, anxiety, iron deficiency anemia, former smoker	Hemorrhagic	170/100	Severe	Altered mental status	Not likely

BP: blood pressure; W: weeks' gestation; PPD: postpartum day; ITP: immune thrombocytopenic purpura; ADHD: attention deficient hyperactivity syndrome; PSUD: polysubstance use disorder.

After a multidisciplinary panel review, 2/15 (13%) of cases were deemed to be possibly preventable. In one occurrence of stroke, the patient presented with an ongoing headache for 4 hours despite rest and analgesia, with focal neurologic deficits; the patient was diagnosed with hemorrhagic stroke in the setting of hypertensive emergency shortly upon hospital arrival. The patient's late presentation to care was considered as a “patient-related” preventable factor (Patient 3 in Table 1). In another case, a critically ill patient was transferred to an adjacent adult, non-obstetrical medical-surgical hospital to access specialty services. There, the patient developed an obstetric hypertensive emergency, which went unrecognized by non-obstetrically trained clinicians accustomed to higher BP ranges outside of pregnancy. The patient was noted to have altered mental status and subsequently diagnosed with hemorrhagic stroke due to severe hypertension upon transfer back to the obstetrical hospital. In this case, lack of specificity around BP management goals by the obstetrical team upon transfer to a non-obstetrical consulting team was considered the most preventable element of the case, resulting in delayed therapeutic response, and was considered a “systems-related” factor (Patient 6 in Table 1).

Conclusion

In our cohort, we described 15 cases of stroke during pregnancy and the first 12 weeks postpartum over 12 years. Only two cases were adjudicated as being possibly preventable, contrasting with previously published rates of preventable stroke-related maternal deaths at 34%–66%, most commonly due to inappropriate treatment of severe hypertension in these studies.^10,11 We hypothesize that we had an overall lower than expected number of cases of stroke related to inappropriate BP management largely due to our hospital's high volume and acuity, with protocolized management and treatment of hypertension in pregnancy in place since 2015. Healthcare facilities with dedicated obstetrical services and large delivery volumes have reported lower rates of maternal and neonatal morbidity than rural or community centers.^21,22 Furthermore, the application of protocols driven by evidence-informed treatment algorithms improves clinical outcomes and reduces health disparities. ²³ In our cohort, 75% of cases complicated by hypertension were adjudicated as “not preventable,” due to timely and appropriate treatment in keeping with national guidelines, or simultaneous presentation with new-onset hypertension and stroke prior to any opportunity for BP management. As these charts did not explicitly describe specific barriers or delays in presentation to care, they were not considered preventable.

We identified opportunities for improvement in two cases, which highlighted the importance of patient education around features of acute stroke, and improved clinician communication and collaboration in obstetrical patients especially in the treatment of obstetrical hypertension. In one case deemed possibly preventable due to “patient-related” factors, our panel highlighted the need for improved education around stroke prevention and hypertension awareness. Specifically, the patient had a headache lasting for several hours prior to developing focal neurologic symptoms and seeking medical attention. Several clinical entities associated with strokes in pregnancy, such as severe preeclampsia, posterior reversible encephalopathy syndrome, and reversible cerebral vasoconstriction syndrome can present with headaches prior to evolving into stroke.^24,25 Moreover, headache has become increasingly recognized as a feature of acute stroke, especially in younger and female individuals and patients with migraine history. ²⁶ To address the need for standardized, systematic education on urgent symptoms and signs in pregnancy, ACOG implemented the Alliance for Innovation on Maternal Health program in 2014 and has since developed a multifaceted approach toward patient education. For example, in their “urgent maternal warning signs” pictograph, the first icon describes a “headache that won’t go away or gets worse over time”. ²⁷ Empowering patients with this information, and elaborating on red flag features of headaches (e.g. sudden onset, difference from typical headache or migraine symptomatology, positionality, association with elevated BPs, or association with neurologic deficits) may contribute to reducing the future incidence of stroke. While some patient education initiatives have found similar disparities in access to maternal health and access to patient education, ²⁸ community-based initiatives have shown promising results. ²⁹ Most cases in our cohort presenting with mild or severe hypertension (7/8, 88%) or headache (7/8, 88%) suffered a hemorrhagic stroke. Proactive hypertension management and education for patients and providers about the importance of headache assessment in pregnancy and postpartum may contribute to decreasing the incidence of hemorrhagic stroke in pregnancy. In addition, providing patients with BP cuffs and information on BP thresholds warranting further evaluation may help to prevent further maternal morbidity. ³⁰ Engaging families and communities of pregnant patients and widely disseminating maternal warning signs may reduce delays in presentation to care. ³¹

Our second case was deemed possibly preventable due to “systems-related” factors. Our obstetrical hospital lies adjacent to our non-obstetrical adult hospital, and while clinicians collaborate on the care of pregnant patients, miscommunication between obstetrical and non-obstetrically trained clinicians, in this case, led to unrecognized severe obstetrical hypertension. Hypertensive urgency/emergency in the non-pregnant, adult population is typically defined as SBP ≥ 180 mmHg and/or DBP ≥ 120 mmHg. ³² As such, while the patient's maximum BP of 160/103 raised red flags for preeclampsia at the obstetric institution, it was not recognized as markedly abnormal by staff members of a general medical-surgical hospital, illustrating the need for education of healthcare providers at large about hypertension thresholds specific to pregnancy. Moreover, additional precautions during transitions of care may be required, such as support from an obstetric provider during non-obstetric procedures, and written orders for BP thresholds specific to pregnancy.

Strengths of this study included the comprehensive ICD codes from which cases were systematically identified, chart review and abstraction of clinical courses by a senior resident physician, and a multidisciplinary panel to adjudicate preventability. In contrast to previous case series of maternal mortality in the setting of maternal stroke that used national or state-level data, we had access to more granular data on events leading up to stroke. Moreover, we highlighted novel pregnancy-specific factors to address to reduce the incidence of maternal strokes moving forward.

Among possible study limitations, cases of strokes and TIA may have been missed if they were misclassified using other diagnostic codes. Second, the exact timing of BP elevation to BP medication administration to a reduction in BP was not recorded. However, the panel reviewed all instances of severe hypertension and defined timely BP medication administration as being in line with ACOG recommendations (i.e. within 60 min). As such, whether or not treatment of severe hypertension was deemed appropriate was adjudicated according to time-to-treatment. Third, the retrospective nature of the study limited the information available to the multidisciplinary panel. Lack of documentation of barriers and delays in presentation to care after the onset of focal neurologic symptoms may have led to an underestimation of preventability. Finally, our results may not be generalizable to non-tertiary obstetric hospitals, and those without an established protocol for treating severe hypertension in pregnancy.

While most cases of maternal stroke in our study were not deemed preventable, our multi-disciplinary panel identified several specific areas among our cases of pregnancy-related stroke that may warrant targeted intervention, including improved education and communication around features of acute stroke—including recognizing headache as a key feature—BP management specific to pregnancy, timely evaluation, and strong clinician-clinician communication when non-obstetrical specialists are involved in the care of pregnant individuals. To reverse the documented trend in worse stroke-related outcomes among women, ³³ barriers to care access, presenting warning signs warranting immediate medical attention, and BP management thresholds specific to pregnancy and the postpartum period must be recognized and addressed. The profound impact of stroke on affected individuals and their families deserves multifaceted efforts toward risk reduction, especially as the incidence of maternal strokes is rising. ³⁴ After ensuring that systems are in place for prompt recognition and treatment of obstetrical hypertensive emergencies, the crucial next step is to promote patient and provider education on stroke warning signs and stroke presentation specific to pregnancy and the postpartum period.