Diagnosis and Treatment of Cerebral Venous Thrombosis

INTRODUCTION

Cerebral venous thrombosis (CVT) is estimated to account for less than 1% of all strokes.^1,2 Unlike arterial strokes, CVT tends to affect young patients with a female predominance, is often nonapoplectic in onset, and has a wide spectrum of clinical presentations.³ These and other features make CVT a challenging disease to diagnose without an understanding of its evolving epidemiology, clinical features, associated conditions, and the neuroimaging findings typically needed to confirm the diagnosis.

ANATOMY AND PATHOPHYSIOLOGY

The condition of CVT includes clots in both the large dural sinuses and cortical veins.⁴ The most prominent superficial sinus is the superior sagittal sinus, which drains into the transverse (lateral) sinuses and then out of the skull via the sigmoid sinuses into the internal jugular veins on each side. The superior sagittal sinus is the most frequent CVT location.⁵ The largest of the many tributary cortical veins that drain into the superior and transverse sinuses are the bilateral veins of Trolard (veins draped vertically over the parietal lobe, which drain into the superior sagittal sinus) and Labbé (veins situated horizontally over the temporal lobe, which drain into the transverse sinus). The deep venous system includes the straight sinus, vein of Galen, inferior sagittal sinus, internal cerebral veins that drain the thalami, and basal veins of Rosenthal (figure 7-1⁶).⁷ It is thought that venous clots often originate in dural sinuses and then propagate to smaller veins resulting in venous infarction, increased intracranial pressure, or both.⁴

IMAGE GALLERY1/5

FIGURE 7-1. Dural sinus and venous anatomy. Three-dimensional contrast-enhanced magnetic resonance venography (MRV) using time-resolved imaging contrast kinetics shows the sagittal view (A) and coronal view (B) of the same patient. Arrows indicate named sinuses and veins. In this patient, the right transverse sinus is dominant. Asymmetric transverse sinuses are a frequent finding on MRV in unselected patients with one study reporting that 61% of patients had a greater caliber right transverse sinus as compared with the left.6

EPIDEMIOLOGY

Recent population-based studies have shown a higher incidence of CVT than previously reported. The latest annual CVT incidence ranges from 1.32 to 2 per 100,000 adults based primarily on data from high-income countries.^1,2,8-10 Rising CVT incidence over time is likely partially due to better disease ascertainment with increasing availability of neuroimaging, although it is also possible that changes in the prevalence of known, emerging, or as yet unknown CVT-predisposing conditions may be a contributing factor.¹¹ In one of the recent studies using US data from two states, 0.66% of all stroke admissions from 2005 to 2016 were for CVT, with an estimated annual CVT incidence of 1.4 to 2 per 100,000 people.¹ In this study, the authors found that the proportion of stroke admissions due to CVT had increased by 70%, from 0.47% in 2005 to 0.80% in 2016, with the largest increases in CVT incidence noted among men and older women. Epidemiologic data from lower- and middle-income countries are lacking and represent an important area for future CVT research.¹

Rates of CVT among women of reproductive age are consistently 3 times higher than those for similarly aged men.^2,8,12,13 The incidence of CVT among women age 18 to 44 years in the aforementioned US study remained virtually unchanged over the study time period at 2.9 to 3.3 cases per 100,000 people.¹ Higher CVT rates among young women are consistent with known sex-specific factors (eg, oral contraceptive use, pregnancy, and the puerperium) associated with CVT.¹² To date, racial differences in CVT incidence have been underexplored. Data from a 2020 study suggest that the incidence of CVT may be disproportionately higher in Black people compared with people of other races in the United States.¹ Prevalence rates of systemic venous thromboembolism, deep venous thrombosis, and pulmonary embolism are generally higher in African Americans compared to other racial and ethnic groups, but the reasons for these differences as well as whether or not they are similar in patients with CVT require additional study.¹⁴

ASSOCIATED CONDITIONS

Conditions associated with CVT can be classified as either predisposing (eg, genetic prothrombotic diseases, antiphospholipid syndrome, cancer) or precipitating (eg, oral contraceptives, infections). In 80% of patients with CVT at least one associated condition is found, and in nearly half of patients with CVT more than a single condition is identified.⁵ Thus, the identification of one condition known to be associated with CVT should not deter clinicians from looking for additional conditions, particularly inherited thrombophilias.¹⁵ Indeed, the American Heart Association/American Stroke Association (AHA/ASA) CVT guidelines note that testing for prothrombotic conditions, including protein C, protein S, or antithrombin deficiencies antiphospholipid syndrome, and prothrombin G20210A and factor V Leiden mutations, can be beneficial for the management of patients with a first CVT (class 2a; level of evidence B).¹⁶ The more recently published European Stroke Organization (ESO) guidelines, however, note that good clinical practice includes performing thrombophilia testing in patients with a high probability of carrying severe thrombophilia (eg, those with a personal or family history of venous thrombosis, or those with CVT without a transient or permanent risk factor) to prevent recurrence based on the existing low-quality evidence surrounding thrombophilia testing.¹⁷ Conditions associated with CVT have been previously detailed in a comprehensive meta-analysis of case-control and cohort studies from 2018¹⁸; table 7-1^18-25 summarizes and expands upon these findings. In the future, genetic and lifestyle data may identify additional conditions associated with CVT and, perhaps, facilitate precision medicine in this space. For example, a recent genome-wide association study using genetic data from 11 European and 1 US research center identified a locus on chromosome 9 that was strongly associated with a nearly threefold increased CVT susceptibility.²⁶ The single-nucleotide polymorphisms with the largest associations were in the coding regions of the ABO blood type gene and, after determining the blood group distribution of cases and controls, these researchers found that a non-O blood group was more prevalent in CVT cases.²⁶

IMAGE GALLERY2/5

TABLE 7-1. Conditions Associated With Cerebral Venous Thrombosis^a

COVID-19 AND CEREBRAL VENOUS THROMBOSIS

The COVID-19 pandemic has led to the identification of additional conditions associated with CVT.

COVID-19 Infection

Infection as a precipitant of CVT (pyogenic CVT) has been well described.^27,28 Data from a 2021 study suggest that a rare but demonstrable association between CVT and COVID-19 infection exists, although the underlying mechanism of this association is uncertain.²⁹ Several thromboembolic pathways have been implicated in the pathophysiology of COVID-19 infection and cerebrovascular disease as well as systemic venous thromboembolism that may also play a role in CVT formation.^30,31 The true prevalence of CVT in patients with COVID-19 infection is not known. A recent systematic review of existing case reports and retrospective cohort studies using data from 34,331 patients hospitalized with COVID-19 estimated the frequency of CVT at 0.08%. Signs, symptoms, and diagnosis of CVT followed the onset of respiratory or systemic COVID-19 symptoms by 1 to 8 weeks in nearly all 54 patients with CVT included in this study.³² Patients with CVT were often noted to have altered mental status, with thrombosis of the deep cerebral venous system or involvement of multiple sinuses.³² Only one of the identified patients with CVT and COVID-19 had an isolated headache which may reflect the fact that patients with CVT without severe neurologic deficits were underdiagnosed during the pandemic (ie, selection bias) or that, among patients with active COVID-19 infection, CVT is clinically more severe.³³ Inpatient mortality was reported in nearly half of patients with CVT and COVID-19 infection, much higher than that of patients with CVT without COVID-19.³⁴ It is unclear if this high mortality rate is related to neurologic complications, care quality, or the severity of COVID-19 infection itself.³² Based on these limited data, in patients with neurologic symptoms and COVID-19 infection, a high index of suspicion for CVT should be encouraged, and treatment of CVT should be initiated as soon as possible.

COVID-19 Vaccination

Another CVT precipitant was identified during postauthorization surveillance of people who had received COVID-19 vaccines using adenoviral vectors (human Ad26.COV2.S and chimpanzee ChAdOx1 nCov-19) to encode the spike glycoprotein of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). In February 2021, reports emerged of patients with thrombocytopenia and venous thromboembolism in unusual locations following administration of the ChAdOx1 nCov-19 vaccine. Within months, three independent descriptions were published of 39 people with thrombosis, frequently CVT, and thrombocytopenia that developed 5 to 24 days after initial vaccination with ChAdOx1 nCov-19.^35-37 Shortly thereafter, 12 women aged 18 to 60 years who presented with CVT and thrombocytopenia 6 to 15 days postvaccination with Ad26.COV2.S were reported.³⁸ To date, no independent predictors for CVT development postvaccine have been identified, although vaccine recipients with CVT and thrombocytopenia were often younger than age 60 years and were women who lacked prothrombotic risk factors.³⁹ In a comprehensive study using data from postauthorization safety reports and official data from 30 European countries and the United Kingdom, the calculated risk of CVT with thrombocytopenia within 28 days of the first ChAdOx1 nCov-19 dose was 4.4 (95% confidence interval, 3.9 to 4.9) per million doses and 0.7 (95% confidence interval, 0.2 to 2.4) after Ad26.COV2.S vaccination. In this analysis, the risk of CVT with thrombocytopenia after ChAdOx1 nCov-19 vaccination was highest among people aged 18 to 24 years (7.3 per million first doses) and lowest in those aged 70 years or older (0.2 per million doses).⁴⁰ On a population level, these absolute numbers are small and, although data collection is ongoing as vaccination continues, findings continue to overwhelmingly support the safety and efficacy of vaccination with respect to reducing COVID-19 risk and reducing death due to COVID-19.^39,41 Additionally, the prevalence of CVT among hospitalized patients with COVID-19, as noted above, is far higher (60-fold to 230-fold) than that of people who received adenovirus-based COVID-19 vaccines.^4,29,42 However, given the link between adenoviral vector–based vaccines and CVT, mRNA vaccines (BNT162b2 and mRNA-1273) may be preferred for certain patient subgroups (eg, younger) since the risk of CVT with thrombocytopenia following administration of mRNA vaccines is nearly zero.⁴⁰

The entity implicated in these rare but potentially devastating cases of CVT and thrombocytopenia following adenovirus-based COVID-19 vaccine administration is now called vaccine-induced immune thrombotic thrombocytopenia (VITT) or thrombosis with thrombocytopenia syndrome. VITT is one of three rare but pathophysiologically related hypercoagulable states associated with thrombosis, low platelet counts, and disseminated intravascular coagulation. The other two hypercoagulable states are heparin-induced thrombocytopenia (HIT), which occurs after heparin exposure and autoimmune heparin-induced thrombocytopenia (aHIT), which refers to a condition where antibodies activate platelets in the absence of heparin.⁴ These three hypercoagulable states (VITT, HIT, and aHIT) are all mediated by platelet-activating antibodies to platelet factor 4 (PF4). In HIT, exposure to unfractionated heparin, a polyanion, causes complexes of PF4 and heparin to form, resulting in the development of IgG autoantibodies against this complex and eventually leading to platelet activation, aggregation, and release of procoagulant molecules. Similarly, in VITT, autoantibodies are thought to be generated from a not-yet-identified polyanion in the adenoviral vaccines or expressed by the cells infected by the vaccine that binds to PF4.⁴³ Why the cerebral veins and sinuses, as opposed to other sites, are a frequent location of thromboses in VITT, occurring in approximately half of patients with VITT at presentation, is not well understood⁴⁴; rates of CVT in VITT differ substantially from rates of CVT in HIT (1.6%) and aHIT (2.5%).^45,46 Prior to the COVID-19 era, thrombocytopenia at CVT presentation was very uncommon.⁴⁷

In a multicenter UK cohort study, patients with VITT-associated CVT had similar presenting features to those with non–VITT-associated CVT, including 84% presenting with headache.⁴⁸ Those with VITT-associated CVT had more intracranial veins thrombosed, more frequent extracranial thromboses, and higher rates of death or dependency as compared to other patients with CVT despite aggressive treatment.⁴⁸ Fortunately, mortality for patients with VITT-associated CVT has significantly decreased from nearly 50% in March 2021 to around 22% for cases diagnosed thereafter, likely due to the beneficial effect of earlier recognition and improved treatments, especially the avoidance of heparin anticoagulation to treat CVT in VITT.⁴⁹ Therefore, in patients with CVT with symptom onset within 4 to 42 days of having received a COVID-19 vaccine using adenoviral vectors, following an algorithmic approach to evaluate and treat VITT is advised (figure 7-2^43,50).^43,50,51

IMAGE GALLERY3/5

FIGURE 7-2. Algorithm for diagnosing vaccine-induced immune thrombotic thrombocytopenia.aPTT = activated partial thromboplastin time; CBC = complete blood cell count; ELISA = enzyme-linked immunosorbent assay; PF4 = platelet factor 4; PT = prothrombin time ; VITT = vaccine-induced immune thrombotic thrombocytopenia.^a Patients with thrombosis and a normal platelet count might be in the early stage of VITT; continued assessment for development of thrombocytopenia/VITT is required.Data from Rizk JE, et al, JAMA Cardiol43 and the American Society of Hematology.50

IMAGE GALLERY4/5

FIGURE 7-3. Imaging of the patient in case 7-1. A, Initial emergency department imaging revealing a left temporal lobe intraparenchymal hemorrhage with subarachnoid extension on axial noncontrast CT (arrow). B, Evidence of dense cord sign on axial noncontrast CT (arrow). C, CT venography demonstrating thrombosis of the left transverse sinus (arrow), sigmoid sinus, internal jugular vein, and the vein of Labbé. D, Coronal T1-weighted MRI sequence confirming thrombosis of the left vein of Labbé; the arrow indicates thrombosis of the transverse sinus extending into the vein of Labbé.

CLINICAL PRESENTATION OF CEREBRAL VENOUS THROMBOSIS

The signs and symptoms of CVT are diverse and may mimic other neurologic disorders, complicating diagnosis. Symptoms of CVT reflect the location of the vein or sinus affected and, in some cases, multiple locations may be affected simultaneously. Presentations of CVT can be roughly divided into four syndromes: (1) isolated headache or increased intracranial pressure, (2) focal neurologic presentations, (3) subacute encephalopathy, and (4) cavernous sinus syndrome with multiple cranial neuropathies.⁵ In ISCVT (International Study on Cerebral Vein and Dural Sinus Thrombosis), a multicenter registry, the median time from CVT symptom onset to diagnosis was 7 days (interquartile range 3 to 16 days), suggesting considerable diagnostic delay.⁵² More recently, a multicenter retrospective US study found that a probable CVT misdiagnosis occurred in 3.6% of patients who presented emergently mostly with headache symptoms.⁵³ Detecting CVT in patients with isolated headache presentations, particularly those with a prior headache history, remains a significant clinical challenge.^54,55

Headaches are present in approximately 90% of patients with CVT.^5,28,56 Failure of blood to drain properly from the brain can lead to increased intracranial pressure manifesting with headache, vomiting, and papilledema with or without visual loss or sixth nerve paresis. When any combination of these features without other neurologic signs is present, the syndrome is referred to as isolated intracranial hypertension and is estimated to occur in nearly one-third of patients with CVT.^5,57 Up to one-quarter of patients with CVT can present with isolated headache without any additional stigmata of raised intracranial pressure.^56,58 Headache presentations in CVT are notoriously diverse. Among patients with new headache, obtaining a detailed neurologic examination, assessing for features known to be associated with CVT, and including secondary causes of headache in the differential may help to improve diagnostic accuracy in CVT. The International Classification of Headache Disorders, 3rd Edition, explicitly notes that headaches attributed to CVT have no specific characteristics. Evidence of a causal relationship between CVT and headache per the International Classification of Headache Disorders, 3rd Edition, simply requires that the headache developed in close temporal relation to the CVT and either headache has significantly worsened in parallel with clinical or radiological signs of extension of the CVT or headache has significantly improved or resolved after improvement of the CVT.⁵⁹ Some key clinical features of CVT-associated headache include being exacerbated by Valsalva maneuver and recumbency, subacute onset of pain, and more often diffuse than unilateral headache location.⁶⁰ However, acute presentations consistent with migraine or thunderclap headache may also occur.⁵⁸ The transverse sinus is frequently a site of thrombosis among patients with CVT and isolated headache.^58,61 Interestingly, a prospective study from 2020 where all patients with CVT were treated with anticoagulation acutely found no significant difference in the frequency of headache of any type or headache with features of intracranial hypertension in patients achieving full recanalization as compared with those who did not fully recanalize.⁶² The relationship between venous recanalization and CVT-associated headache as well as the mechanism by which CVT causes headache pain require further investigation.

Alternatively, patients with CVT may present with focal deficits, seizures, or both. Symptoms depend on the area of the brain affected. Common focal symptoms in CVT include hemiparesis, aphasia, and loss of vision.⁵ A key feature of focal neurologic deficits due to CVT is that they are frequently progressive in contrast with arterial strokes, which tend to be maximal at onset.⁶³ The duration of CVT symptoms prior to presentation was greater than 48 hours in 53% of patients in the VENOST study,⁵⁶ another large CVT registry, and, in ISCVT, symptom onset of between 48 hours and 30 days was seen in slightly more than half of patients.⁵ Another feature that may distinguish the focal deficits of CVT from those of more commonly encountered arterial strokes is their bilateral nature, particularly when the superior sagittal sinus is affected.^28,52 Seizures, both generalized and focal, are far more common in CVT than with arterial cerebrovascular events, occurring in nearly 40% of patients with CVT at initial presentation.⁵²

Finally, patients with thrombosis of the deep venous system may develop a subacute encephalopathy with confusion and lethargy or experience a rapid neurologic deterioration progressing to coma due to edema of bilateral thalami, basal ganglia, or other deep structures typically drained by these veins.^64,65 Approximately 10% of patients with CVT have thrombosis of the deep cerebral venous system.^52,64 Timely neurologic imaging is an essential component of the diagnostic evaluation for patients who present in coma to distinguish deep cerebral vein thrombosis from other neurologic causes. Cortical vein thrombosis is thought to be even more rare than deep venous system thrombosis, occurring in only 116 patients in the published literature.⁶⁶ However, since most of the published cases of isolated cortical vein thrombosis have associated venous infarct, underreporting and potentially underdiagnosis of nonsevere isolated cortical vein thrombosis cases may occur.

RADIOGRAPHIC FINDINGS

Given the broad spectrum of presentations seen in CVT, recognizing the presence of CVT on clinical grounds requires a high degree of suspicion. Radiologic studies are crucial to establish the definitive diagnosis of CVT. In the emergency setting, CT is often the imaging test of choice for patients presenting with acute focal neurologic symptoms. However, ruling out CVT is difficult to do using CT alone.⁶⁷ Both direct visualization of a thrombus in a cerebral vein or sinus, often called the “dense clot sign” or “cord sign,” as well as visualization of a filling defect within a dural sinus after contrast administration, the “empty delta sign,” have high specificity but low sensitivity.⁶⁸ Intracerebral hemorrhage, which is well seen on head CT, is present in approximately one-third of patients with CVT.⁵ Increased suspicion for hemorrhagic venous infarcts from CVT should occur when multiple intraparenchymal hemorrhages are present; infarcts are ill-defined or in a nonarterial territory; or involve the bilateral thalami or bilateral basal ganglia, or are juxtacortical (case 7-1).^69,70

Advanced imaging is generally required to diagnose and rule out CVT with certainty, and noninvasive imaging is typically favored over cerebral angiography, which is the gold standard.⁷² CT venography, a contrast-enhanced helical CT examination performed with a time-optimized contrast bolus, allows direct visualization of a thrombosed cerebral vein, and is a fairly reliable alternative to angiography with high sensitivity but low specificity.⁷³ Magnetic resonance venography (MRV) can also be used to detect CVT without ionizing radiation exposure. Both CT venography and contrast-enhanced MRV are superior to time-of-flight MRV techniques since complex flow can produce artifacts in the latter. Contrast-enhanced MRV allows for a direct assessment of luminal filling similar to that of CT venography, with comparable sensitivity and specificity to CT venography.^74,75 Contrast-enhanced brain MRI provides detailed information about the brain parenchyma and is probably more accurate for diagnosing CVT than noncontrast-enhanced MRV sequences.⁶⁸ On brain MRI, the most common finding is visualization of the thrombus in the T1-weighted images. However, the timing of the CVT is an important consideration for thrombosis visualization on MRI: in the acute phase (first 5 days after CVT) thrombus is isointense on T1-weighted images and hypointense on T2*-weighted images; from 5 to 15 days the thrombus becomes hyperintense on T1-weighted and T2*-weighted images; finally, after 15 days it becomes homogeneous and hypointense in all image sequences. T2*-weighted and susceptibility-weighted imaging sequences can be useful to assist in the diagnosis of isolated cortical venous thrombosis on brain MRI.^16,76

Using MRI black-blood thrombus imaging, a noncontrast-enhanced T1-weighted imaging method that allows for direct visualization of the thrombus itself, to detect CVT has very high sensitivity and excellent specificity compared to combined CT and MRI modalities in small studies.^77,78 Since a similar native contrast thrombus MRI technique is highly accurate for the diagnosis of new and recurrent lower-extremity deep venous thrombosis,^79,80 black-blood thrombus imaging may be of value for CVT detection if current findings are supported by additional data.⁶⁸ An interesting radiographic finding seen on brain MRI, the brush sign, has recently been described in CVT. This sign is an abnormal hypointensity of the subependymal and deep medullary veins in paramagnetic-sensitive MRI brain sequences that has been reported in patients with CVT, particularly those with thrombosis of the deep venous system (figure 7-4⁸¹). The deep medullary veins are small-caliber vessels located adjacent to the atrium and posterior body of the lateral ventricle, draining the white matter of the cerebral hemispheres to the subependymal veins of the lateral ventricles. The brush sign was associated with higher thrombus load as well as ipsilateral parenchymal lesions in one small study; this sign may be a marker of CVT severity and should prompt close monitoring during the acute phase.⁸¹

IMAGE GALLERY5/5

FIGURE 7-4. The brush sign in cerebral venous thrombosis. Axial T2*-weighted MRI shows bilateral brush sign (A, arrows) in a patient with acute thrombosis of the superior sagittal sinus, bilateral transverse sinus, jugular vein, vein of Galen, straight sinus, and internal cerebral veins; this extensive CVT is well seen on magnetic resonance venography (MRV) (B).Reprinted with permission from Aguiar de Sousa D et al, Stroke.81 © 2019 Wolters Kluwer Health.

TREATMENT

Medical, endovascular, and surgical treatments can be used to treat CVT.

Acute Anticoagulation

Anticoagulation remains the first-line treatment of choice for CVT in the acute setting, even when concurrent intracerebral hemorrhage is present.^16,17,82 Both the AHA/ASA and the ESO guidelines recommend initiation of parenteral anticoagulation with unfractionated or low-molecular-weight heparin (LMWH) prior to transitioning to oral anticoagulants for CVT treatment.^16,17 The ESO guidelines have a weak recommendation for LMWH over unfractionated heparin based on a meta-analysis suggesting a nonsignificant trend toward improved functional outcomes and mortality with LMWH without a difference in rates of bleeding.^17,83 This trend is in keeping with an analysis from ISCVT that suggested LMWH might be safer and perhaps more efficacious than unfractionated heparin.⁸⁴

The clot in the venous system is the primary target of acute anticoagulation. In a systematic review including data from 694 patients with CVT, vessel recanalization occurred in roughly 85% at follow-up. This study found a significant increase in the chance of favorable outcome (defined as a modified Rankin Scale [mRS] score of 0 to 1) in patients with CVT with recanalization.⁸⁵ New data suggest that vessel recanalization occurs early once anticoagulation is initiated. In a prospective study of 68 patients with CVT all treated acutely with anticoagulation who were imaged 0, 8, and 90 days from diagnosis, 43 (68%) had partial recanalization and 4 (6%) had full recanalization at day 8.⁶² At 90 days, 41% had partial recanalization and 54% had full vessel recanalization. Patients with early recanalization were at a lower risk of enlargement of nonhemorrhagic lesions and showed early regression of venous infarcts. These findings support the widely accepted hypothesis that recanalization improves regional perfusion.⁶² An additional important aspect of early anticoagulation initiation in CVT is prevention of other dangerous venous thromboembolisms in patients with CVT, particularly pulmonary embolisms, at index hospitalization.⁸⁶

Duration of Anticoagulation

The duration of anticoagulation following CVT has not been studied in any randomized controlled trials; current recommendations are based on extrapolation from venous thromboembolism data.^16,17 This extrapolation from venous thromboembolism to the CVT population, who are younger and likelier to have had a provoked event, has been challenged. EXCOA-CVT (EXtending oral antiCOAgulation treatment after acute Cerebral Vein Thrombosis) is a cluster randomized trial designed to evaluate the efficacy and safety of anticoagulation with vitamin K antagonists for 3 to 6 versus 12 months after CVT to clarify the optimal duration of this therapy.⁸⁷ Current AHA/ASA guidelines note that for patients with provoked CVT (associated with a transient risk factor) a 3- to 6-month duration of anticoagulation is reasonable but that for patients with unprovoked CVT anticoagulation may be continued for 6 to 12 months.¹⁶ Patients who have recurrent venous thrombosis or an associated prothrombotic condition with a high thrombotic risk may need permanent anticoagulation; specific recommendations for the prevention of recurrent venous thromboembolic events should be followed in those conditions.¹⁷

Use of Direct Oral Anticoagulants

Since the noninferiority of direct oral anticoagulants to prevent systemic venous thromboembolism as compared to vitamin K antagonists has been shown,⁸⁸ a great deal of interest has been expressed in using these newer agents to reduce the risk of venous thromboembolism after CVT. Direct oral anticoagulants have a favorable safety profile, predictable pharmacokinetics, and are easier to administer than vitamin K antagonists.⁸⁹ Some support for the use of direct oral anticoagulants after CVT comes from recent clinical trials. RESPECT-CVT (Clinical Trial Comparing Efficacy and Safety of Dabigatran Etexilate with Warfarin in Patients With Cerebral Venous and Dural Sinus Thrombosis) was a prospective, randomized, open-label trial to evaluate the efficacy and safety of dabigatran compared to dose-adjusted vitamin K antagonists to prevent recurrent venous thromboembolism.⁹⁰ In this trial, 120 adult patients with CVT were randomized in a 1:1 fashion to dose-adjusted warfarin to maintain an international normalized ratio (INR) of 2 to 3 or dabigatran 150 mg twice a day for 24 weeks. Patients were eligible for inclusion if they were stable after 5 to 15 days of treatment with therapeutic heparin, able to swallow, did not have CVT associated with either CNS infection or major trauma, and did not have any surgical treatments planned. At the end of the study, no recurrent venous thromboembolisms were observed in either treatment group, but three major bleeding events occurred. These major bleeding events involved one patient with intestinal bleeding in the dabigatran group and two patients with intracranial (subdural) hemorrhages in the vitamin K antagonists group. The two treatment groups had nearly identical rates of CVT recanalization on imaging. While RESPECT-CVT could not detect a statistically significant difference between the two treatments for recurrent venous thromboembolism, it did suggest that dabigatran is a reasonable option to prevent recurrent venous thromboembolism after CVT.

The safety and efficacy of rivaroxaban in CVT has also been studied. EINSTEIN-Jr (Oral Rivaroxaban in Children With Venous Thrombosis), a multicenter, parallel-group, open-label, randomized study, compared rivaroxaban to dose-adjusted vitamin K antagonists in children (aged 0 to 17 years) with venous thromboembolism; a prespecified substudy of patients enrolled with CVT was published in 2020.⁹¹ The main trial assigned patients in a 2:1 fashion to body weight–adjusted rivaroxaban in a 20 mg equivalent dose or standard anticoagulation. Of the 114 children with confirmed CVT, symptomatic recurrent venous thromboembolism occurred in none of the 73 children in the rivaroxaban group versus one (2.4%) of the 41 children in the standard anticoagulation group. Clinically relevant, nonmajor extracranial bleeding was observed in five rivaroxaban recipients (6.8%) and one child (2.4%) in the standard anticoagulation arm had a major bleeding event (subdural). No new venous infarcts occurred in either group and repeat imaging demonstrated a similar effect on clot resolution with rivaroxaban as compared to vitamin K antagonists.⁹¹ Although it is challenging to apply the findings from a pediatric trial (particularly one that was not powered for hypothesis testing) to an adult population, the CVT substudy of EINSTEIN-Jr supports the overall trend that direct oral anticoagulants are effective after CVT. Currently, SECRET (Study of Rivaroxaban for CeREbral Venous Thrombosis), an open-label, randomized, controlled, phase II clinical trial designed to evaluate the safety of rivaroxaban in adults, is ongoing.⁹² Another trial called RWCVT (Rivaroxaban versus Warfarin in CVT Treatment) has completed enrollment but has not yet published results. An ongoing international observational study, DOAC-CVT (Direct Oral Anticoagulants for the Treatment of Cerebral Venous Thrombosis, NCT04660747), may also inform future practice.

Many practitioners have already started using direct oral anticoagulants to treat CVT based on existing data. A systematic review of direct oral anticoagulant use in CVT found a substantial increase in observational cohorts and case series that included patients with CVT treated with apixaban, dabigatran, edoxaban, or rivaroxaban since 2019. The observational cohorts included in the systematic review reported a similar risk of death in direct oral anticoagulant versus standard therapy arms and noted that favorable outcomes (mRS 0 to 2) were more likely in direct oral anticoagulant–treated patients with CVT.⁹³ ACTION-CVT (Direct Oral Anticoagulants Versus Warfarin in the Treatment of Cerebral Venous Thrombosis), a large multicenter retrospective study, provides additional reassurance to those who use direct oral anticoagulants after the acute phase of CVT. The ACTION-CVT study included 845 patients with CVT across 27 centers in four countries. It found that direct oral anticoagulant treatment was associated with a similar rate of recurrent venous thromboembolism (5.26 versus 5.87 per 100 patient years), a lower risk of major hemorrhage (2.44 versus 4.70 per 100 patient years), and similar rates of death and recanalization as with vitamin K antagonists.⁹⁴ In ACTION-CVT, two-thirds of patients on direct oral anticoagulants were treated with apixaban. Although ACTION-CVT and other retrospective treatment studies are prone to confounding by indication, no major safety issues have been found with the use of direct oral anticoagulants as opposed to vitamin K antagonists in clinical practice.⁹⁵ It is important to note that patients with antiphospholipid antibody syndrome should not be treated with direct oral anticoagulants based on two randomized trials showing an increase in arterial thrombotic events when these patients were treated with rivaroxaban instead of warfarin.^96,97 Additionally, pregnant patients with CVT should likely be continued on LMWH rather than any other agent due to potential teratogenicity.⁹⁸

Endovascular Therapy

Endovascular therapy, defined as mechanical thrombectomy with or without thrombolysis, has long been used to rapidly recanalize occluded sinuses in patients with severe CVT who worsen despite anticoagulation or who cannot receive anticoagulation. The results of the TO-ACT (Thrombolysis or Anticoagulation for Cerebral Venous Thrombosis) trial were recently published.⁹⁹ This multicenter, open-label, blinded-endpoint, randomized trial was designed to assess the safety and efficacy of endovascular therapy. Patients with CVT were randomized 1:1 to receive either endovascular therapy with anticoagulation or anticoagulation alone. Adult patients with radiologically confirmed CVT who had one or more prespecified features associated with poor outcome (mental status disorder, coma state, intracerebral hemorrhage, or thrombosis of the deep venous system) were included in the study. TO-ACT was halted after the first interim analysis for futility. A total of 67 patients were randomized into TO-ACT, accounting for approximately 16% of patients with CVT (67 of 420) seen at the sites during the study period. The number of patients with a favorable functional outcome (mRS 0 to 2) at 12 months was very similar between treatment groups (85% versus 82%). The mortality rates at 6 and 12 months were numerically higher in the endovascular therapy group than the medical management group but did not reach statistical significance. Because of the small sample size, TO-ACT is somewhat difficult to interpret. It remains possible that improved methods of patient selection or different endovascular techniques (in TO-ACT, more modern stent retrievers and aspiration catheters were not used) may increase the frequency of favorable outcomes after CVT. Indeed, symptomatic hemorrhage was higher in the medical arm (9% versus 3%) in TO-ACT, suggesting that EVT may be effective in reducing the risk of further bleeding. Retrospective data, however, have suggested that the role of endovascular therapy in CVT is limited. Using data from the Nationwide Inpatient Sample from 2004 to 2014, researchers found that 3% of identified patients with CVT were treated with endovascular therapy. These researchers found that endovascular therapy was independently associated with an increased risk of death (odds ratio 1.96) after adjustment to account for measured confounders.¹⁰⁰ Further data as to the role of endovascular therapy in select patients with CVT are likely needed before treatment recommendations can be made.

Treatment of Vaccine-Induced Immune Thrombotic Thrombocytopenia–Associated Cerebral Venous Thrombosis

Early recognition, diagnosis, and treatment of VITT has led to favorable patient outcomes.¹⁰¹ As in HIT, therapeutic anticoagulation with nonheparin anticoagulants is the primary treatment for VITT with or without CVT. Based on data from patients with aHIT, administration of IV immunoglobulin (IVIg) as soon as VITT is diagnosed or under consideration is recommended by both the American Society of Hematology (ASH) and the AHA/ASA at a recommend dose of 1 g/kg for two days.^50,51 In patients with severe VITT (extensive thrombosis with platelets <50 × 10³/μL) or resistant VITT (continued thrombosis despite medical therapy), treatment with IVIg as a sole immunotherapy might not be enough because of the high antibody burden.⁴³ Although the 2020 AHA/ASA guidelines for VITT-associated CVT do not address plasma exchange as a treatment modality, the ASH notes that plasma exchange can be considered if thrombosis continues despite nonheparin anticoagulation and IVIg.¹⁰² Aspirin should not be used in patients with VITT. Platelet transfusions should be avoided with careful risk and benefit assessments made in patients with bleeding, who require surgical intervention, or both.⁵¹ Some cases of VITT-associated CVT with extensive clot burden treated with endovascular therapy have been reported with mostly favorable outcomes.¹⁰³ Given the rarity of thrombocytopenia and CVT, all care should be individualized for each patient.

Decompressive Surgery

Herniation attributable to unilateral mass effect produced by large edematous venous infarctions or parenchymal hemorrhages is the major cause of acute death in CVT.¹⁰⁴ Although the vast majority of patients with CVT have a favorable outcome, about 4% develop cerebral edema severe enough to cause brain herniation. In these instances, decompressive craniectomy, hematoma evacuation, or both have been used to prevent death.⁵ Although a potential disadvantage of craniectomy is that it precludes anticoagulation for the immediate postoperative period, support for decompressive surgery exists. In a retrospective study that included a systematic review of the literature, a total of 69 patients with CVT were identified.¹⁰⁵ Of those, only 12 (17%) had a poor outcome (mRS 5 to 6) at a median of 12 months of follow-up. Nearly one-third of patients who were comatose prior to surgery recovered completely at follow-up.¹⁰⁵ An updated systematic review published in 2019 identified 169 patients with CVT who were treated with decompressive surgery, mostly from low- to middle-income countries, and similarly found a low mortality rate of 16% at follow-up.¹⁰⁶ Despite the low quality of evidence, the ESO guidelines now strongly recommend using decompressive surgery for patients with acute CVT and parenchymal lesions with impending herniation to prevent death as a randomized controlled trial is unlikely for ethical and feasibility reasons.¹⁷

Chemical Prophylaxis

Whether individuals with a history of CVT would benefit from targeted prophylaxis in scenarios associated with increased venous thromboembolism risk is uncertain and represents an important area for future research. Pregnant women have been the focus of some research regarding the use of chemical prophylaxis to prevent venous thromboembolism or CVT recurrence.¹⁰⁷ In a small retrospective study of 63 women who became pregnant after their diagnosis of CVT and were treated with LMWH for the entire gestational period, two (3%) had venous thromboembolisms and none had bleeding complications.¹⁰⁸ In an update of a systematic review published in 2017 that included a total of 393 patients, an analysis stratified according to antithrombotic prophylaxis showed a trend toward lower rates of recurrent CVT and extracerebral venous thromboembolism in patients receiving antithrombotic prophylaxis with heparin.^109,110 Although limited, based on these and other data both AHA/ASA and the ESO recommend LMWH prophylaxis in pregnant patients with a previous history of CVT.^16,17 The optimal dose of LMWH in pregnant women with moderate to high risk of recurrence of venous thromboembolism is the subject of substantial debate with an ongoing open-label randomized controlled trial comparing two different doses of LMWH in pregnant patients with a history of venous thromboembolism.¹¹¹

CLINICAL OUTCOMES

In general, CVT has a favorable outcome with an in-hospital mortality rate ranging from 1% to 4% and from 8% to 10% during long-term follow-up.^{2,5,10,13,104,112,113} Mortality rates after CVT have been declining. One systematic review found an inverse correlation between mortality and the calendar year in which patients with CVT were recruited into a particular study.¹¹⁴ The frequency of presenting with focal neurologic deficits and coma also decreased significantly over time. Possible explanations are improvements in treatment, a shift in risk factors, and the identification of less severe cases by improved diagnostic methods.¹¹⁴ Most studies reporting outcomes after CVT have reported mRS scores at follow-up which may not accurately capture the morbidity that follows CVT. A single-center study of 161 patients with CVT in Finland found that even though 82% of patients had an mRS of 0 to 1 at 6 months, as many as 68% of patients reported residual symptoms which frequently included neuropsychological difficulties and headache.¹¹⁵ Older, smaller-cohort studies have identified cognitive impairments, headaches, and seizures after CVT, frequently resulting in unemployment.^116-118 Future research detailing functional outcome after CVT and evaluating interventions to improve patients’ ability to return to the workforce is warranted.

CONCLUSION

The epidemiology of CVT is changing, including more frequent detection among older patients and increased reported incidence rates. The presentation of CVT can be subtle and usually differs from that of other cerebrovascular diseases, making detection of CVT on advanced neuroimaging an essential component of diagnosis. The use of direct oral anticoagulants to treat CVT is an important advance that has recently been shown to be safe and effective, and evidence supports a shift toward their use in clinical practice. To date, the rates of CVT as a complication of adenovirus-based COVID-19 vaccination are very low (<5 per million vaccine doses) and essentially zero with mRNA-based vaccines, with the benefits of COVID-19 vaccination far outweighing the risk of VITT.

KEY POINTS

• Unlike arterial strokes, cerebral venous thrombosis (CVT) has a wide spectrum of clinical presentations, tends to affect younger patients with a female predominance, and is often nonapoplectic in onset.
• The latest annual CVT incidence ranges from 1.32 to 2 per 100,000 adults based primarily on data from high-income countries.
• Conditions associated with CVT can be classified as either predisposing (eg, genetic prothrombotic diseases, antiphospholipid syndrome, cancer) or precipitating (eg, oral contraceptives, infections).
• Data from a 2021 study suggest that a rare but demonstrable association between CVT and COVID-19 infection exists, although the underlying mechanisms of this association are uncertain.
• In patients with neurologic symptoms and COVID-19 infection, a high index of suspicion for CVT should be encouraged, and treatment of CVT should be initiated as soon as possible.
• The entity implicated in the rare but potentially devastating cases of CVT and thrombocytopenia following adenovirus-based COVID-19 vaccine administration is now called vaccine-induced immune thrombotic thrombocytopenia (VITT), or thrombosis with thrombocytopenia syndrome.
• In patients with CVT with symptom onset within 4 to 42 days after having received a COVID-19 vaccine using adenoviral vectors, following an algorithmic approach to evaluate and treat VITT is advised.
• Presentations of CVT can be roughly divided into four syndromes: (1) isolated headache or increased intracranial pressure, (2) focal neurologic presentations, (3) subacute encephalopathy, and (4) cavernous sinus syndrome with multiple cranial neuropathies.
• A key feature of focal neurologic deficits due to CVT is that they are frequently progressive in nature in contrast to arterial strokes which tend to be maximal at onset.
• Contrast-enhanced brain MRI provides detailed information about the brain parenchyma and is probably more accurate for diagnosing CVT than non-contrast-enhanced magnetic resonance venography sequences.
• Both the American Heart Association/American Stroke Association (AHA/ASA) and the more recently published European Stroke Organization (ESO) guidelines recommend initiation of parenteral anticoagulation with unfractionated or low-molecular-weight heparin prior to transitioning to oral anticoagulants for CVT treatment.
• The ACTION-CVT study and other retrospective treatment studies are prone to confounding by indication; nevertheless, there do not seem to be major safety issues with the use of direct oral anticoagulants as opposed to vitamin K antagonists in clinical practice.
• As in heparin-induced thrombocytopenia, therapeutic anticoagulation with non-heparin anticoagulants is the primary treatment for VITT with or without CVT.
• Despite the low quality of evidence, the ESO guidelines now strongly recommend using decompressive surgery for patients with acute CVT and parenchymal lesions with impending herniation to prevent death as a randomized controlled trial is unlikely for ethical and feasibility reasons.
• In general, CVT has a favorable outcome with an in-hospital mortality rate ranging from 1% to 4% and from 8% to 10% during long-term follow-up.