Cerebral Venous Thrombosis

Elio Agostoni; Angelo Aliprandi; Marco Longoni


Expert Rev Neurother. 2009;9(4):553-564. 

In This Article

Physiopathology of Cerebral Venous Thrombosis

Anatomy of the Venous System

Venous drainage of the brain can be essentially divided into two main districts: a superficial system that receives blood from the brain convexity (mainly from the cerebral cortex); and a deep system, which instead drains the deep white matter and the grain nuclei. The superficial veins drive blood from the frontal and parietal lobes into the superior sagittal sinus, while blood from the temporal and occipital lobes is drained by the left and right transverse sinuses. The deep cerebral veins, on the other hand, drain blood into the straight sinus through the Galen vein or into the transverse sinuses. The venous circulation of the neural structure in the posterior region, including the brainstem and cerebellum, have a more variable and inconstant anatomy; however, the veins of this region ultimately end in the transverse sinuses. Finally, the venous blood from the anteroinferior areas of the brain (inferior surface of the fontal lobes) and from the face is collected in the cavernous sinuses, and from there is driven to the transverse sinuses through the petrous sinuses. The superior sagittal, transverse and straight sinuses are connected in the confluence of sinuses in the posterior fossa; therefore, thrombotic processes affecting this area can impair the venous drainage of most of the brain and are, thus, responsible for dramatic and often fatal clinical syndromes. Further anastomotic connections are often present: the superficial system is connected with the deep system through the vein of Trabant, while the superior sagittal and transverse sinuses are in communication through the vein of Labbè. These pathways may provide some collateral flow in case of venous occlusions, explaining the mild clinical conditions observed in many patients.

Most of the venous blood of the brain is eventually collected into the transverse sinuses, which continue into the sigmoid sinuses and deep jugular veins; one of the later sinuses is often prevalent and conveys more blood flow. A minority of venous blood is also drained through anostomotic connections with the vertebral plexus.

Pathophysiology of Cerebral Venous Thrombosis

The two main pathogenic mechanisms of cerebral venous thrombosis (CVT) are intracranial hypertension and parenchymal ischemia; these two aspects often coexist but are responsible for distinct clinical symptoms. The consequences of CVT depend on the prevalence of the former or the latter mechanisms, and on the site and extent of the venous occlusion in relation to the anatomical features described previously. In case of a limited sinus occlusion, an effective collateral flow through cortical collaterals or with the deep circulation is often possible. In this case, parenchymal damage is limited to a localized edema without ischemic lesions. If the thrombotic process is more widespread and involves more than one sinus, collateral flow is insufficient to allow venous drainage; the first consequence is a dramatic increase in capillary pressure, leading to leakage of fluids in the extracellullar compartment, brain swelling and intracranial hypertension.[1,2] In addition, in cases of thrombosis of a cortical vein, collateral circulation is impossible. In this case, brain ischemia develops rapidly and a venous stroke is invariably present. Venous strokes have peculiar characteristics compared with arterial ones: the degree of edema is usually greater and hemorrhagic complications, ranging from small petechial lesions to a proper hemorrhagic infarction, are much more frequent.[3,4] The higher degree of brain edema is probably due to the presence of a vasogenic component, which is absent in arterial strokes, as demonstrated by MRI techniques.[5,6] The vasogenic component can be rapidly offset when venous drainage is restored, without the development of a permanent parenchymal lesion; this might be the reason why some venous strokes apparently of large size have a more favorable outcome. The prevalence of vasogenic edema has also suggested a possible use of steroids to limit brain swelling, but clinical experience has been disappointing. The high frequency of hemorrhagic complications, the frequent cortical involvement and the severe brain edema are probably responsible for the elevated rate of seizure observed in venous stroke,[7,8] in contrast to arterial stroke; these features are also important for the different diagnosis of cavernous sinus thrombosis.

An important pathogenic mechanism in CVT is the development of intracranial hypertension. Elevation of intracranial pressure may be due to a reduction of liquoral absorption resulting from sinus occlusion or to the mass effect of swollen brain areas. Intracranial hypertension may be mild, usually in patients without parenchymal lesions and with limited clinical features, or may lead to a dramatic worsening of the clinical conditions and to coma. The latter condition is most often observed in patients with large supratentorial lesions with a severe mass effect or in patients with occlusion of vessels of the deep circulation. In fact, in cases of occlusion of an internal cerebral vein or of the straight sinus, the diencephalic structure responsible for awareness are directly injured, leading to early loss of consciousness in these patients. Furthermore, edema of these centroencephalic areas leads rapidly to compression of the brainstem, endangering patient survival. Finally, parenchymal, subdural and even subarachnoid hemorrhages have been described as a consequence of venous thrombosis.[9,10]

Causes of CVT

The main causes of CVT are essentially similar to those of venous occlusion in other regions and may be caused by pathologies that either affect the vessel wall, reducing blood flow, or induce a state of systemic hypercoagulation. The causes of CVT have been recorded in two recent studies. The first was a multicentric prospective record that enrolled 650 patients mainly in Europe and South America, but also with contributions from China, Australia and Canada;[11] the International Study for Cerebral Venous Thrombosis (ISCVT). The study had a prospective design and a median follow-up of 16 months. The second study analyzed a total of 232 patients from the USA, 157 by retrospective chart review and 75 by prospective enrollment; only 182 patients, however, fulfilled the predefined diagnostic criteria and were included in the study.[12]

Before discussing the different causes identified, some preliminary methodological considerations are needed. First of all, the study of Wasay et al. included a majority of patients identified with retrospective analysis: this feature may be responsible for the high rate of patients labeled as 'unidentified cause' (43 vs 12% of the ISCVT), and to the underestimation of same causative factors, such as oral contraceptive (OC) use, which is present in only 5% of patients compared with 54% of the ISCVT. Furthermore, while more than one causative factor was allowed in the ISCVT, only one cause was attributed to each patient in the study by Wasay and some risk factors were not tested in all patients. However, the presence of multiple risk factors is very likely and is probably a better description of the actual clinical reality. As a result of these differences, direct comparison of the two studies is difficult as far as etiology is concerned and, for this reason, we did not pool the data from these two sources. Also, the earlier cohort studied by Ameri and Bousser reported a much lower rate of contraceptive use, but as the patients were recruited retrospectively it may be influnced by the same bias.[13] Finally, in our own group of 44 patients, we found a similar distribution as the ISCVT.[14] The causes of CVT ascertained in the two larger studies and in the two retrospective groups are reported in Table 1.

Despite these methodological problems, however, some conclusions can indeed be drawn. First of all, the main cause of CVT is linked to pregnancy or puerperium (6.3 and 13.8%, respectively, in the ISCVT) or to assumption of OCs. The ISCVT also reported a 4.3% rate of patients assuming hormone-replacement therapy. All of these conditions share the state of hypercoagulation induced by estrogens as the causative factor of CVT. The role of estrogens in determining CVT was indeed first suggested by the rapid increase of the female-to-male ratio of the disease after the widespread use of OCs (465 out of the 625 patients in the ISCVT were females); and has more recently been confirmed in two case-control studies.[15,16] As for peri- and post-partum, the risk has been estimated to be approximately 12 cases for every 100,000 births.[17]

Other causes include inherited or induced states of hypercoagulation; the former are more prevalent (Table 1) and are present in 10-20% of patients. These include a deficit of proteins C or S, antithrombin III and resistance to factor V. It is important to note that in ISCVT 40% of patients had more than one risk factor; the most frequent association is a combination of OCs or puerperium-peripartum with an inherited coagulopathy. In particular, prothrombin gene mutation 20210A has been found to be the most frequent genetic alteration in CVT patients and to significantly increase the risk when combined with use of OCs.[15,18] Therefore, the search for causal factors should be extensive in all patients and should always include a screening for coagulopathies; the identification of one factor should never stop the diagnostic work-up. Interestingly, the risk associated with OC use does not seem to be increased by cigarette smoking, in contrast to arterial stroke. As a matter of fact, a case-control study on 43 young women with CVT whose only risk factor was OC use and 255 healthy OC users did not show any difference in the rate of smokers (Table 2).[19] This may reflect the different pathogenic mechanism in arterial and venous stroke, but more data from epidemiological and clinical studies in this field are needed. Other causes of hypercoagulation that must always be considered are tumors, systemic infections and acquired hypercoagulation states, which include antiphospholipid antibodies, nephrotic syndrome and hyperhomocysteinemia. A similar mechanism, which involves an increase of blood viscosity, is present in case of polycythemia or leukemia, or of severe anemia or dehydratation. The role of severe anemia as a risk factor of CVT has been specifically tested in a prospective study on 120 subjects (odds ratio [OR] = 1.10; 95% CI: 1.01-2.22; p < 0.05).[20]

Inflammatory conditions may cause CVT by direct damage of vessel walls or by inflammatory hypercoagulative states, including systemic lupus erythematous, Behçet's disease and rheumatoid arthritis. Also, sarcoidosis and intestinal bowel diseases and vasculitis in the setting of AIDS are thought to cause venous thrombosis by such mechanisms. Local causes instead lead to direct damage of a sinus or cortical vein or to the impairment of local venous circulation. The former mechanism is involved in the case of CNS neoplasms and head trauma or surgery, while the latter is more frequent in the case of brain or dural arteriovenous malformations. Finally, some infectious diseases may spread to the brain sinuses by contiguity and lead to their thrombosis. Such a mechanism is responsible for cases of transverse sinus occlusion in patients with middle ear otitis, or of cavernous sinus syndrome in those with ocular or facial infections. Infectious conditions were among the first causes of CVT in the past, but their incidence was reduced in industrialized countries by the introduction of antibiotics. However, the ISCVT still reports an 8.2% rate of middle ear, face or neck infection as causative factors, probably reflecting the persistence of such pathologies in other parts of the world, such as South America or China (only 1% was reported in the USA study).


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.