Acute thrombosis of the cerebral sinuses and veins (cerebral venous thrombosis, CVT) is considered to be the cause of an acute stroke in approximately 1% of all stroke patients. However, the incidence of CVT is not known, as population-based studies are lacking. It has been estimated that annually about five to eight cases of CVT are identified among stroke patients of tertiary care hospitals . Historically, CVT was considered a severe, almost inevitably fatal disease, as diagnosis in the pre-angiograph era was usually made post-mortem. However, modern neuroimaging techniques allow the diagnosis of CVT at an early stage and document that CVT is more frequent than was traditionally assumed, and that its prognosis is much better than is generally accepted, provided that the diagnosis is suspected, the respective neuroimaging examinations are performed in a timely manner, and therapy is initiated early, i.e. often with the diagnosis being clinically suspected only. The variety of clinical signs and symptoms renders the diagnosis of CVT a challenge to the physician. Diagnosis is still frequently overlooked or delayed due to the wide spectrum of clinical symptoms and the often subacute or lingering disease onset.
It is important to keep the diagnosis of CVT in mind in stroke cases that present with a fluctuating course, headache, epileptic seizures, or disturbances of the level of consciousness. With timely therapeutic intervention, CVT has a favorable prognosis, with an overall mortality rate of about 8% in recent studies . However, thromboses of the inner cerebral veins as well as septic CVT remain severe diseases with high mortality rates.
The cerebral venous system consists of two distinct groups – the superficial and the deep cerebral veins – which eventually drain into the cerebral sinuses. The superficial veins of the brain that drain the cortex and the underlying white matter form a network of anastomoses that drain into the cortical sinuses, but number, diameter, and topography of these veins vary among individual patients. However, two major superficial veins can be identified in the majority of patients: the upper anastomotic vein of Trolard, which drains into the superior sagittal sinus, and the lower anastomotic vein of Labbé, which drains into the transverse sinus. Cerebral veins do not possess valves and therefore allow blood flow in both directions. This is the main reason why even larger thrombotic venous occlusions may remain clinically silent for a long time. In contrast, the deep veins that drain the basal ganglia and other deep subcortical structures do not possess the diversity of the superficial venous network. The basal veins of Rosenthal and the internal cerebral veins drain into the great cerebral vein of Galen and the straight sinus, and from there the transverse and sigmoid sinuses, finally reaching the vena cava via the jugular veins. Blood supply to the cerebellum and brainstem is drained from the posterior fossa by veins reaching the vein of Galen, the petrose, or the lateral sinus. In contrast to veins, the cerebral sinuses are formed by duplication of the dura mater and are fixed to the osseous cranial structures. Thus, there is no possibility of influencing venous blood flow by means of vasoconstriction or vasodilatation.
Cerebral veins have a peculiar anatomy, as they do not follow the arteries as in other parts of the body.
CVT may be due to infectious and non-infectious causes. Septic CVT is observed as a complication of bacterial infections of the visceral cranium, namely otitis, sinusitis, mastoiditis, and bacterial meningitis. The infectious agents reach the cerebral sinuses ascending via the draining veins of the face, the sinuses, or the ear, or following local inflammation that destroys osseous structures that separate the infectious focus from the brain. Clinical signs and symptoms of septic CVT comprise signs of systemic infection and of meningitis. Septic CVT remains a rare disease with high mortality, in spite of modern therapeutic surgical and medical approaches (see below for details).
Aseptic CVT may stem from a variety of causes, all of them resembling those of extracranial thrombosis (Box 14.1). However, the cause of CVT remains unknown in approximately 15–20% of all patients, in spite of a thorough diagnostic workup [2–4].
Septic CVT may be caused by bacterial infections of the visceral cranium (e.g. otitis, sinusitis, mastoiditis, and bacterial meningitis). Aseptic CVT may be caused by the same causes as extracranial thrombosis (see Box 14.1).
Genetic prothrombotic conditions
Anti-thrombin III deficiency
Protein C and protein S deficiency
Factor V Leiden mutation and resistance to activated protein C
Prothrombin G20210A mutation
Mutations in the methylenetetrahydrofolate reductase (MTHFR) gene
Acquired prothrombotic states
Anti-phospholipid and anti-cardiolipin antibodies
Otitis, mastoiditis, sinusitis
Systemic infectious disease
Systemic lupus erythematosus
Inflammatory bowel disease (Crohn disease, colitis ulcerosa)
Polycythemia, primary and secondary
Anemia, including paroxysmal nocturnal hemoglobinuria
Hormonal replacement therapy
Cytotoxic drugs (e.g. asparaginase, tamoxifen)
Mechanical causes, trauma
Injury to sinuses or jugular vein, jugular catheterization
Dehydration, especially in children
Venous thrombosis of the CNS differs from arterial thromboses in many ways: venous thrombosis does not manifest acutely, as arterial thrombosis does, but is a subacute, often fluctuating process, in which endogenous pro-thrombotic and fibrinolytic processes occur concurrently. Regional cerebral blood flow (rCBF) is not significantly impaired, the auto-regulation of cerebral perfusion is nearly fully maintained, and administration of acetazolamide induces – in contrast to arterial thrombosis – a significant increase of rCBF . In venous congestion, disturbances of neuronal functional metabolism are tolerated for a much longer time than in arterial occlusion, and full recovery from severe focal and generalized neurological signs and symptoms may be observed in CVT even after weeks.
Intracranial hemorrhage is often observed in CVT, and its incidence may reach 40–50% [3, 6], a percentage significantly higher than in cerebral arterial thrombosis or embolism. The most common form of intracranial hematoma in CVT is intracerebral bleeding, but subdural and – rarely – subarachnoid hemorrhage may be observed. In general, intracerebral hematoma in CVT is atypically localized in cortical and subcortical regions that do not correspond to territories of cerebral arteries. From a pathophysiological point of view, these bleedings are caused by the diapedesis of erythrocytes through the endothelial membrane, following the increase of the venous and capillary transmural pressure after venous thrombosis. The rationale for anti-coagulant therapy with heparin or low molecular weight heparin (LMWH) is that preventing the re-occlusion of veins and sinuses reopened by endogenous fibrinolysis will result in a lowering of venous and capillary pressure. Thus, even in the presence of intracerebral hemorrhage due to CVT, immediate anti-coagulation results in clinical amelioration without increase in hematoma volume in the majority of patients.
Hemorrhages are frequent in CVT.
Abrupt occlusion of a cerebral artery results in the acute manifestation of focal neurological symptoms due to ischemia of the brain tissue perfused by this artery. In contrast, cerebral venous thrombosis may remain clinically silent, as long as venous drainage is maintained by collateral veins or sinuses. Eventually, failure of collateral venous drainage will result in the gradual, fluctuating, or progressive clinical manifestation of focal or generalized brain dysfunction. An exception to this rule is CVT in pregnancy and puerperium, where signs and symptoms of venous thrombosis may present within minutes or hours .
Clinical features of CVT differ according to the venous structures involved. Cortical CVT will present with signs and symptoms different from that of deep CVT, and septic CVT will show findings other than aseptic thrombosis.
In most prospective clinical series [2, 3, 6, 8], intense and diffuse headache was either the first (>70%) or the most common (75–90%) symptom of cortical venous thrombosis. Headache, as well as nausea, papilloedema, visual loss, or sixth nerve palsy, is due to increased intracranial pressure. The onset of headache in CVT is subacute over hours and may precede the manifestation of other symptoms and signs by days or even weeks. Acute appearance of epileptic seizures is observed in 40–50% of all cases of CVT [2, 3, 6, 8], a percentage much higher than in arterial thrombosis of the brain. Seizures in CVT may present as simple partial seizures with post-ictal limb paresis or as complex partial seizures, and in both cases secondary generalization is often observed. Focal neurological signs may be observed in 30–50% of CVT patients [2, 3, 6, 8], but their localizing value is limited, due to the excellent collateralization of cerebral veins and the lack of venous valves that allows inversion of venous drainage in the case of localized thrombotic occlusion. Furthermore, the intensity of focal signs and symptoms may fluctuate over time. Motor symptoms may initially present as a monoparesis that gradually develops into a full-blown hemiparesis. With cortical CVT, higher cortical functions may be impaired, and aphasia or apraxia may be observed. Impairment of the level of consciousness (any degree from somnolence to deep coma) may be present in 30–50% of patients, and acute delirium or psychotic symptoms are observed in 20–25% [2, 3, 6, 8]. As a rule, extended thrombosis of cortical sinuses will result in symptoms and signs of generalized brain dysfunction (headache and other signs of increased intracranial pressure, impairment of the level of consciousness, generalized seizures), while isolated cortical venous thrombosis will result in focal neurological signs or focal seizures.
The rare thromboses of the inner cerebral veins (veins of Rosenthal, great vein of Galen, straight sinus, etc.) will result in a severe dysfunction of the diencephalon, reflected by coma and disturbances of eye movements and pupillary reflexes, a condition usually associated with poor outcome .
Thrombosis of the cavernous sinus may present with the characteristic combination of ocular chemosis, eye protrusion, painful ophthalmoplegia, trigeminal dysfunction, and – occasionally – papilloedema and visual disturbances. Cavernous sinus thrombosis may be unilateral, but the good collateralization between the cavernous sinuses usually leads to bilateral symptoms, while extension of the thrombosis into the large cerebral sinuses is the exception. Most cases of cavernous sinus thrombosis are due to ascending infection from the orbita, the paranasal sinuses or other structures of the viscerocranium and are accompanied by signs of acute local or systemic infection.
Symptoms of CVT are manifold: they may remain clinically silent as long as venous drainage is still maintained. Headache is the most common and frequently the first symptom of CVT. Epileptic seizures, focal neurological signs, impairment of the level of consciousness, and psychotic symptoms can occur.
Septic thrombosis of other sinuses is found as a complication of bacterial infection (e.g. otitis, mastoiditis, bacterial meningitis), and is nearly always accompanied by symptoms and signs of systemic infection. Septic CVT accounts for about 5% of all cases of cerebral thrombosis only, but is characterized by excess mortality in spite of maximum treatment.
Septic CVT is accompanied by symptoms of systemic infection.
Due to the multitude of clinical manifestations as well as etiologies, the diagnosis of CVT remains a challenge to the clinical physician. The less distinct the clinical presentation is, the more difficult is the diagnosis of CVT. CVT may be suspected in the presence of headache and other signs of intracranial hypertension, alone or in combination with epileptic seizures and fluctuating neurological signs, especially if conditions are present that may favor thrombogenesis (e.g. bacterial infection, pregnancy and puerperium, malignancies, and known pro-thrombotic states; see Box 14.1). However, mono- or oligosymptomatic cases of CVT may be difficult to diagnose. In patients with signs and symptoms of systemic infection, CVT may be mistaken for meningo-encephalitis. The presence of CVT has to be suspected in young stroke patients, in painful stroke, in stroke with unusual presentation, and in patients with first-ever headache in combination with seizures or subtle focal signs.
The differential diagnosis of aseptic CVT comprises benign intracranial hypertension, but also all forms of intracranial hypertension due to neoplastic diseases. Aseptic thrombosis of the cavernous sinus leading to painful uni- or bilateral ophthalmoplegia has to be differentiated from the Tolosa-Hunt syndrome.
Cerebral computed tomography (CCT) is widely available and is feasible in critically ill patients. Thus, CCT is often the first neuroimaging technique applied to patients with CVT and should be performed before and after the intravenous application of iodinated contrast media. However, CCT findings in CVT are often non-specific and may consist of one or more of the following: localized or diffuse brain edema, focal hypodensities that do not comply with the boundaries of cerebral arterial territories, atypical hemorrhagic infarctions or hematomas (Figure 14.1). CT venography (CTV) increases the diagnostic yield of contrast-enhanced CCT in the diagnosis of CVT [1, 9]. Using modern CT scanners, CVT is considered to be equally sensitive to MRV in the corroboration of the clinical diagnosis of CVT . However, CCT may be entirely normal in up to 25% of patients with angiographically proven CVT. Thus, the main indication of CCT in CVT remains to rule out other conditions that may mimic or be confounded with CVT.
Figure 14.1 Unenhanced cranial computed tomography scan showing an atypical right temporal hemorrhagic venous infarction in a patient with isolated cortical venous thrombosis. Note the cord sign.
However, there are two CCT findings that – if present – are highly suggestive of CVT. The thrombotic occlusion of an isolated cortical vein may present as a thread-like hyperdense structure on no-contrast CCT (“cord sign,” Figures 14.1 and 14.2A). After intravenous application of iodinated contrast media, the dura mater of the sinuses will show a distinct enhancement (Figure 14.2B), and the non-enhancing intravenous thrombus may be discriminated as a triangle (“empty triangle” or “Delta-sign,” in analogy to the design of the Greek capital letter Delta [Δ]). While the cord sign is found in up to 20% of CVT cases only, the Delta-sign has been described in 15–45% of CVT patients .
Figure 14.2 Cranial computed tomography in a patient with thrombosis of the straight sinus: the straight sinus presents as a hyperintense thread (cord sign) in non-enhanced CCT (A), while after intravenous injection of iodinated contrast media the surrounding sinus structures show a distinct enhancement surrounding the thrombus (B).
A “cord sign” (a thread-like hyperdense structure on no-contrast CCT) and a “Delta-sign” (a triangle-shaped non-enhanced structure showing after application of contrast media) are highly suggestive of CVT. Other findings are non-specific, such as brain edema. The main indication of CCT in CVT is to rule out other conditions.
Magnetic Resonance Imaging
Cerebral magnetic resonance imaging (MRI, Figure 14.3) and magnetic resonance venography (MRV) are extremely sensitive in detecting CVT as well as the underlying parenchymal alterations. The ability of MRI and MRV to obtain images in various planes facilitates the visualization of the different cerebral sinuses. It is important to obtain – at least initially – tri-planar MRI in sagittal, axial, and coronal T1 and T2, T2*, and FLAIR sequences in combination with MRV, in order to minimize confusion of CVT with sinus aplasia or hypoplasia, and not to mistake the T2-weighted hypointense signal of deoxyhemoglobin and intracellular methemoglobin with flow voids [10, 11]. MRI and MRV allow direct imaging of the thrombus, whose signal intensity depends on clot age. Initially (days 1–5), thrombotic material gives an isointense signal on T1 images instead of the normal intraluminal flow void and a strongly hypointense signal on T2 images, indicating the presence of deoxyhemoglobin in erythrocytes of the thrombus. During the second week after clot formation, red blood cells are destroyed, and deoxyhemoglobin is metabolized into methemoglobin, and the thrombus yields a hyperintense signal on both T1- and T2-weighted images. After 2 weeks, the thrombus becomes hypointense on T1- and hyperintense on T2-weighted images, and recanalization may occur with the reappearance of flow void signaling. Partial or total recanalization is observed within 4–5 months after thrombosis [10–12]. MRI and MRV are non-invasive neuroimaging techniques and may easily be repeated for follow-up and re-evaluation of the course of the disease. However, MRI and MRV are – in most cases – unable to detect isolated cortical venous thrombosis.
MRI and MRV are highly sensitive in detecting CVT. They allow direct imaging of the thrombus; the signal intensity depends on clot age.
Figure 14.3 Magnetic resonance imaging (T1-weighted images after intravenous injection of paramagnetic contrast media) in a patient with thrombosis of the superior sagittal, straight, and right transverse sinus.
Digital Subtraction Angiography
Until recently, digital subtraction angiography (DSA) has been the gold standard for the diagnosis of CVT, documenting the partial filling of cerebral venous structures after intra-arterial injection of iodinated contrast media (Figure 14.4). However, DSA is an invasive diagnostic procedure, associated with a periprocedural risk of death or stroke of about 1%. Furthermore, the interpretation of DSA (as of MRV or CT venography) may be complicated by the presence of anatomical variations, e.g. the hypoplasia of a transverse sinus . Often, indirect signs of thrombosis, e.g. the dilatation of venous collaterals, or the regional prolongation of venous transition time are the only findings that indicate the presence of CVT. Thus, the role of DSA in the diagnosis of CVT remains restricted to those patients where the clinical suspicion cannot be corroborated by other neuroimaging techniques.
Figure 14.4 Digital subtraction angiography in a patient with isolated thrombosis of the right inferior anastomotic vein of Labbe (right), in contrast to physiological imaging of the cerebral vein findings of the contralateral hemisphere (left).
DSA, the former gold standard in CVT diagnosis, is nowadays restricted to patients where other neuroimaging techniques are not feasible.
Other Diagnostic Findings
The definite diagnosis of CVT is based on the detection of venous thrombosis by the neuroimaging techniques described above. As differential diagnosis of CVT comprises a large number of diseases, diagnostic workup in patients with the final diagnosis of CVT requires extensive laboratory exams as well as other auxiliary testing: lumbar puncture, electroencephalography (EEG), and transcranial duplex ultrasound are often performed, but most findings are non-specific.
Most routine laboratory findings in the acute phase of aseptic CVT are non-specific: mild leukocytosis, elevated erythrocyte sedimentation rate, and CRP are the most common abnormalities. Acute thrombosis may be suspected if the D-dimers, a fibrinogen degradation product, are found to be elevated. However, elevated D-dimers just indicate active thrombosis (anywhere in the body), and normal values for D-dimers do not exclude acute CVT. Recently, measuring D-dimers before neuroimaging in patients with suspected CVT has been suggested, except in those with isolated headache and in case of prolonged duration of symptoms (i.e. >1 week) before the test [1, 14, 15].
Other laboratory markers for acute thrombosis include PAI-1, thrombin-anti-thrombin (TAT), and plasmin-anti-plasmin (PAP) complexes. However, their diagnostic value in the acute phase of CVT is under debate, and testing is not widely available.
Venous thrombosis (including CVT) may be associated with an increased risk of thrombophilia, i.e. of recurrent thrombosis. However, routine thrombophilia screening in all CVT patients is not recommended by the current guidelines . Thrombophilia screening may be feasible in patients with high pre-test probability of carrying severe thrombophilia (i.e. a personal and/or family history of venous thrombosis, young age at CVT, CVT without a transient or permanent risk factor) to prevent recurrent venous thrombotic events (VTEs). A suggested panel of laboratory examinations to exclude major inherited or acquired thrombophilia is given in Box 14.2.