Cerebral Venous Thrombosis in Pregnant and Postpartum Women Presenting to the Emergency Department with Headaches

Abstract

Introduction:

Cerebral venous thrombosis (CVT) is a thromboembolic disease of the intracranial venous systems. The disease can be difficult to diagnose as it often requires a high index of suspicion. Risk factors for the disease include pregnancy, oral contraceptive pills, congenital thrombophilia, infection, cancer, polycythemia, head trauma, and recent surgery. However, there have been no studies in the United States that have examined whether pregnancy and the postpartum stage are truly a risk factor for CVT. The aim of this study is to determine whether pregnant and postpartum women presenting to the emergency department with headaches have a higher incidence of CVT to better risk stratify which patients need to have advanced imaging pursued.

Methods:

A retrospective, observational case–control study was performing by querying the electronic medical record at a large county hospital for patients presenting with a headache to the emergency department. Patients were stratified into groups based on whether they were diagnosed with CVT, pregnancy status, and comorbid conditions to determine the risk associated between pregnancy, the puerperium stage, and CVT.

Results:

A total of 20,955 males and females presented to the emergency department between January 1, 2016 and April 13, 2023, with a chief complaint of headache. There were 19,474 female patients and 9581 male patients. In the case group, there were 793 pregnant women and 53 postpartum women. In the control group, there were 18,628 women who were not pregnant. Of the 22 patients diagnosed with CVT, 1 was in the puerperium stage and no patients were pregnant. Pregnant and postpartum patients were 1.05 (0.14–7.80) times more likely to develop CVT. Pregnant and postpartum patients were 1.73 (0.23–13.52) times more likely to develop CVT when controlled for comorbidities. Patients in the puerperium stage were 26.48 (3.33–210.87) times more likely to develop CVT when controlled for comorbidities.

Conclusion:

Pregnant patients presenting to the emergency department with headaches do not have a significantly higher risk of CVT; however, puerperium patients have a significantly higher risk of CVT compared to the general population.

INTRODUCTION

Cerebral venous thrombosis (CVT) is a thromboembolic disease of the intracranial venous systems which circulates blood from the dural sinuses into the internal jugular vein [Figures 1 and 2]. The disease can be quite difficult to diagnose as it often requires a high index of suspicion and the wherewithal to pursue a computed tomography (CT) venogram or magnetic resonance venogram. The annual incidence of CVT is estimated to be 12.1 per million people per year as it only accounts for 0.5% of strokes.[1] One study examined 8829 patients diagnosed with CVT. The average age of patients with CVT was 32.9 and the male-to-female ratio was 2:3.[2] The most common presenting symptom is headache, which is present in over 90% of patients and in 25% of patients is the only reported symptom. Focal motor weakness is present in approximately 40% of patients. Additional signs and symptoms include papilledema and seizures. Approximately, 16% of patients with CVT have seizures.[3] Seizures were most common in patients with hemorrhagic conversion or with superior sagittal sinus vein and cortical vein thrombosis.[4,5] A cranial nerve six palsy is the most common cranial nerve palsy in CVT.[6]

Figure 1.

Magnetic resonance venogram showing thrombosis of the left transverse sinus

Figure 2.

Magnetic resonance venogram reconstruction showing thrombosis of the left transverse

The clinical presentation of CVT is the result of two factors based on the principle that cerebral perfusion pressure (CPP) is equal to the difference between the intracranial pressure (ICP) and mean arterial pressure. CVT causes increased venous pressure resulting in decreased CPP. Decreased CPP then causes ischemia and cytotoxic edema of the brain leading to headaches. In addition to increased venous pressure, obstruction of the cerebral sinuses by thrombosis results in impaired cerebral spinal fluid absorption leading to increased ICP.[7]

Historically, risk factors for the disease include pregnancy, oral contraceptive pills, congenital thrombophilia (Factor V Leiden, antiphospholipid antibody syndrome, or G20210A prothrombin gene), infection (sinusitis, otitis media, or mastoiditis), cancer, polycythemia, head trauma, jugular vein catheterization, and recent surgery.[8] Many of these risk factors are assumed rather than studied based on the premise established by Virchow's triad; thrombosis is influenced by a state of hypercoagulability, venous stasis, and vessel wall damage. However, there have been no studies in the United States that have examined whether pregnancy is truly a risk factor for CVT. A case–control study in Europe studied patients from five academic centers in Sweden, Canada, Finland, Netherlands, and Australia examined the risk of CVT in pregnancy. Approximately, 25% of patients with CVT were pregnant or in the postpartum period representing an odds ratio of 3.8. However, when stratified into pregnant and postpartum were associated with an odds ratio of 1.2 (0.6–2.3), which is insignificant. Meanwhile, the postpartum period was associated with an odds ratio of 10.6 (5.6–20.0).[9]

Approximately 15%–20% of pregnant women develop migraines.[10] Pregnant women presenting to the emergency department with headaches represent a vulnerable population. The D-dimer assay has been investigated as a potential test to rule out patients for CVT who are at low risk. One meta-analysis examined 101 patients with CVT presenting to the emergency department with headache. A false negative D-dimer was found in 13 individuals (12.9%) for a sensitivity of 87.1%.[11] Therefore, advanced imaging is the only method to reliably diagnose CVT.

It is widely quoted that at 50 mGy there may be harm to the fetus. A CT head typically results in 0.001–0.01 mGy radiation exposure to the fetus [Table 1].[12,13] The classic finding on the CT head of CVT is known as the dense triangle sign, which represents a hyperdense focus in the superior sagittal sinus. This finding is uncommon and CVT is only diagnosed one-third of the time by a noncontrast CT head. While the radiation from a CT head is well below the theoretical threshold for harm to the fetus, there is no documented amount of radiation to the fetus caused by a CT venogram. The aim of this study is to determine whether pregnant and postpartum women presenting to the emergency department with headaches have a higher incidence of CVT to better risk stratify which patients need to have advanced imaging pursued.

Table 1.

Fetal radiation exposure based on radiographic study

Type of study	Fetal radiation dose (mGy)
Head or neck CT	0.001–0.01
Chest X-ray	0.005–0.01
Chest CT or CT pulmonary angiogram	0.01–0.66
Abdominal CT	1.3–35
Pelvic CT	10–50

Modified from Tremblay et al. CT: Computed tomography

METHODS

Study design and study area

This was a 7-year retrospective, observational case–control study performed by querying the electronic medical records for patients presenting to the emergency department with headaches from January 1, 2016, to April 13, 2023 at a large county hospital with over 900 beds.

Study participant characteristics

The inclusion criteria were defined as women regardless of age presenting to the emergency department with headaches. Exclusion criteria were defined as patients with incomplete records preventing which limited subgroup analysis and further investigation. Patients were then separated into case and control groups [Figure 3]. The case group included pregnant women and those in the puerperium stage defined as 6 weeks postpartum. The control group included nonpregnant women. Patients were then identified as being diagnosed with CVT or other headache syndrome based on ICD-10 codes. Patients were stratified into groups based on the number of risk factors and comorbidities present. Comorbid conditions identified included hypertension, hyperlipidemia, diabetes mellitus, and history of tobacco abuse.

Figure 3.

Case and control characteristics

Data analysis

MedCalc statistical software (Ostend, Belgium) was utilized to determine the odds ratio and statistical significance. Differences were considered statistically significant when P < 0.05.

RESULTS

A total of 20,955 males and females presented to the emergency department between January 1, 2016, and April 13, 2023, with a chief complaint of headache. There were 19,474 female patients and 9581 male patients. In the case group, there were 793 pregnant women and 53 postpartum women. In the control group, there were 18,628 women who were not pregnant. There were 42 patients diagnosed with CVT. Exclusion criteria limited the analysis to 22 patients diagnosed with CVT as the remainder of the patients were excluded due to incomplete records available for subgroup analysis and record review. Of the 22 patients diagnosed with CVT, 1 was in the puerperium stage and no patients were pregnant [Table 2].

Table 2.

Cases of pregnant/postpartum patients with headaches diagnosed with cerebral venous thrombosis

	CVT	No CVT
Pregnant/postpartum	1	846
Not pregnant	21	18,606

CVT: Cerebral venous thrombosis

Pregnant and postpartum patients were 1.05 (0.14–7.80) times more likely to develop CVT using a 95% confidence interval (P = 0.9640). Patients were stratified based on the risk factors and comorbidities. Comorbidities identified were a history of hypertension, diabetes mellitus, hyperlipidemia, and tobacco abuse. The single patient in the puerperium stage with CVT had no identifiable comorbid conditions. A total of 10 nonpregnant patients with CVT had no comorbid conditions. A total of 751 pregnant and 12,977 nonpregnant patients presenting to the emergency department had no identifiable comorbid conditions and were not diagnosed with CVT [Table 3]. When comorbid conditions were taken into consideration, pregnant and postpartum patients were 1.73 (0.23–13.52) times more likely to develop CVT using a 95% confidence interval (P = 0.6023). A total of 53 patients presented to the emergency department with headache who were in the puerperium stage, 50 of these patients had 0 comorbidities. Patients in the puerperium stage were 17.04 (2.25–129.02) times more likely to develop CVT using a 95% confidence interval (P = 0.0060) and 26.48 (3.33–210.87) times more likely when controlled for comorbid conditions [Table 4].

Table 3.

Cases of women presenting to the emergency department with headaches based upon comorbidities

Number of Comorbidities	0	1	2	>3
Number of comorbidities	0	1	2	>3
Pregnant/postpartum women who did not have CVT	751	90	2	2
Nonpregnant women who did not have CVT	12,977	4121	1232	277
Women who did have CVT	11	4	6	1

CVT: Cerebral venous thrombosis

Table 4.

Odds ratios comparing pregnant and postpartum women with headaches diagnosed with cerebral venous thrombosis

	OR of CVT	P
Pregnant and postpartum women	1.05 (0.14–7.80)	0.9640
Pregnant and postpartum women when controlled for comorbidities	1.73 (0.22–13.52)	0.6023
Postpartum women	17.04 (2.25–129.02)	0.0060
Postpartum women when controlled for comorbidities	26.48 (3.33–210.87)	0.0020

OR: Odds ratio, CVT: Cerebral venous thrombosis

DISCUSSION

This retrospective review of 20,955 patients suggests that pregnant women presenting to the emergency department with headaches were not significantly more likely to be diagnosed with CVT; however, postpartum women were significantly more likely to be diagnosed with CVT.

The strengths of this study include that the data were examined over a 7-year period with a robust control group. The most glaring issue with this retrospective study is that it is underpowered as there was only 1 woman in the postpartum stage diagnosed with CVT and no pregnant women were identified. The case group only had 42 patients; however, it is important to recognize that CVT remains a relatively rare disease. Thus, there simply is not a large amount of cases available in a disease that has an incidence of 12.1 cases per one million people.

Further limitations of the study include the age of the population studied. The county hospital where the study took place supports a community that has an average age of 48.9 years. Meanwhile, the average age in the United States is approximately 38.1 years.[14] It would be beneficial if a similar study was conducted that only evaluated women of childbearing age from 15 to 40 years. Unfortunately, the lack of cases of CVT was prohibitive for this evaluation due to the lack of a national registry for medical data that is found throughout many countries in Europe. The hope is that in the future a meta-analysis may be able to offer additional perspective and analysis. Additional limitations of the study are related to the querying of the EMR. There were limited and incomplete records from 2001 to 2016, which affected the amount of data that could reliably used during the study and 20 cases had to be excluded due to incomplete records for review. It is also unclear how many patients may have had CVT; however, were not diagnosed because the correct imaging study was not pursued. Finally, cases may have been missed related to patients who did not have a primary complaint of headache. This highlights the fact the providers need to not only be hypervigilant for this diagnose, but they also need tools to help aid them in diagnosis.

One area of interest is the clinical CVT score developed by Heldner et al. A risk score was assigned based on 6 factors: seizure (4 points), thrombophilia (4 points), oral contraceptive use (2 points), symptom duration for more than 6 days (2 points), worse headache ever (1 point), and focal neurologic deficits (1 point). A high score was defined as more than 6 points, moderate score as 3–6 points, and a low score as <3 points. The study examined 539 patients, 5.9% of patients in the low-score group had CVT, 28.3% in the moderate-score group, and 92.5% in the high-score group. When the D-dimer assay (3 points) was added to the clinical score, the utility of the study became even more noticeable. There was a 100% positive predictive value for CVT for patients with a clinical score between 9 and 17. Interestingly, not a single patient with a low score diagnosed with CVT had a negative D-dimer. While Heldner et al. does capture the primary risk factors for thrombosis, the data from our study are inconclusive whether pregnancy should be added to the risk score but suggests that postpartum status should be included in the risk score.[15]

While previous suspicions indicate that pregnant women are in a hypercoagulable state increasing their risk of thromboembolism, it is unclear to what degree this influences the risk of development of CVT. Pregnant women presenting to the emergency department with headaches were not more likely to have CVT; however, women in the puerperium state were significantly more likely when comorbidities were taken into consideration. Our recommendation remains that CVT should be on the differential diagnosis of any patient presenting to the emergency department with headache. The presence of pregnancy alone in a patient presenting to the emergency department with a headache does not necessitate that advanced imaging be performed. Similar to all patients, pregnant patients and those in the postpartum period must have a detailed history taken and a physical exam performed to evaluate for the numerous life-threatening causes of headache. Only by considering the patient's presentation, as a whole can we hope to recognize what is overall regarded as a rare diagnosis and ensure the best possible care for pregnant and postpartum patients presenting to the emergency department with neurological complaints.

CONCLUSION

Pregnant patients presenting to the emergency department with headaches did not have a significantly higher risk of CVT; however, postpartum patients had a significantly higher risk of CVT compared to the general population.

Research quality and ethics statement

This study was approved by the Institutional Review Board/ Ethics Committee (IRB# 2033768). The authors followed applicable EQUATOR Network (https://www.equator-network.org/) guidelines during this research project.

Financial support and sponsorship

Nil.

Conflicts of interest

Dr. Sagar Galwankar is a member of the editorial board of the Journal of Emergencies, Trauma and Shock.