Long-term Prognosis in Cerebral Venous Thrombosis
Follow-up of 77 Patients
Background and Purpose Very little is known about the long-term outcome of patients with cerebral venous thrombosis (CVT), particularly regarding the risk of residual epilepsy and further thrombotic events. We retrospectively studied 77 patients with CVT diagnosed by angiography and/or MRI.
Methods A cohort of 77 patients aged 18 to 77 (mean, 38.5) years with CVT, evaluated from 1975 through 1990, was followed up for a mean of 77.8 months (range, 14 to 204 months; median, 63 months). Information on death, neurological status, seizures, recurrent CVT, other thrombotic events, and subsequent pregnancies was obtained from direct observation, mail questionnaire, or telephone interviews.
Results Sixty-six of 77 patients (85.7%) had no neurological sequelae during follow-up. Eleven patients (14.3%) remained neurologically impaired. Two who initially presented with isolated intracranial hypertension had blindness due to optic atrophy. The other 9 had focal signs at the time of CVT and were left with various cognitive or focal deficits. Four of 28 (14.3%) patients who had seizures at the acute stage had recurrent seizures. One of the 51 patients with lateral sinus thrombosis developed a dural arteriovenous fistula. Nine of the 77 patients (11.7%) suffered a second CVT, all but one in the first year. Noncerebral thrombotic events occurred in 11 patients (14.3%). No recurrence of CVT occurred during later pregnancies, but 1 patient had a postpartum deep vein thrombosis.
Conclusions In the present series, CVT has an essentially good long-term prognosis. The frequency of long-standing epilepsy was low, suggesting that long-term anticonvulsant treatment is not necessary in the majority of cases. A second CVT or another thrombotic episode occurred in 20% of patients, stressing the need in a minority of cases for long-term anticoagulation.
Cerebral venous thrombosis (CVT) has for many years been diagnosed mainly at autopsy. This has led to the description of a very rare and lethal disease characterized clinically by headache, papilledema, seizures, focal deficits, coma, and death and pathologically by hemorrhagic infarction contraindicating the use of anticoagulants.1 2 3 4 The introduction and widespread use of cerebral angiography,5 CT of the brain, and more recently MRI6 made early diagnosis of CVT possible and completely modified our knowledge of this condition.7 8 Whereas in the early literature mortality still ranked between 30% and 50%,1 2 4 it was between 5.5% and 30% in more recent series.6 7 8 9 10
Very little is known about the long-term outcome of patients with CVT, since there is no study specifically dedicated to the subject. We report the long-term prognosis of 77 patients with CVT from the above-mentioned series of 110 patients.8
Subjects and Methods
The primary patient pool consisted of 110 consecutive adult subjects examined by us between 1975 and 1990 for a first CVT, confirmed by cerebral angiography and/or MRI with MR angiography.8 Six patients died during the acute event, and 2 died shortly thereafter of their underlying disease. Thus, there were 102 patients (from now on referred to as the “original patient group”) available for the study. A mail questionnaire and telephone interviews were used to obtain information from patients who could be located with the help of hospital records or publicly accessible sources. We specifically asked for presence of any of the following: headache, epileptic seizures, and sensorimotor and visual disturbances. We also collected data on the occurrence of CVT recurrence (diagnosed on new clinical, angiographic, or CT scan signs), deep vein thrombosis, pulmonary embolus, recurrence of CVT, any other health problem and hospitalization, and subsequent pregnancies. Other questions referred to professional activity and prior and current medication.11 By use of this method, 46 of the original 102 patients could be reliably located; of those, 43 returned the questionnaire, for a response rate of 93%.
For additional patients who had regular clinic follow-up visits, data were obtained during their outpatient visits. In addition, follow-up of 15 patients was based on chart review alone; these were former clinic patients who had either moved and could not be contacted or who had died in the meantime of known causes, often after extensive follow-up periods. Thus, a total of 77 patients of the original 102 could be entered into the data pool. Twenty-five of the original 102 subjects could not be located; 60% of the 25 were not French residents.
Table 1⇓ lists the main characteristics of the follow-up group and those lost to follow-up. The 77 patients of the follow-up group comprised 30 men and 47 women (ratio, 0.64); length of follow-up ranged from 14 to 204 months (median, 63 months; mean, 77.8 months [6.5 years]). There were no significant differences between the followup group and the primary group of 110 for any studied variable (sex, age, mode of onset, clinical presentation, type of seizure, site of occlusion, presence of a parenchymal brain lesion on CT/MRI, suspected cause of thrombosis, and heparin treatment).
As far as treatment was concerned, 62 patients (80.5%) received anticoagulants: heparin for a few days followed by oral anticoagulants for 3 to 4 months, unless there was an underlying disease carrying a thrombotic risk (Behçet’s disease, thrombophilia, etc), in which case anticoagulation was not stopped. Among these 62 patients, 16 had hemorrhagic infarcts before treatment, and no worsening was observed. Anticonvulsant agents were used only in the 28 patients who had seizures in the acute stage. There was no fixed duration of treatment. Anticonvulsants were tapered off gradually after 3 to 4 months unless attacks persisted.
The majority of the patients (66 of 77 patients [85.7%]) had no neurological sequelae during follow-up. All these patients had been symptom-free at the time of discharge. Eleven patients (14.3%), while sometimes improving to a considerable extent, continued to suffer from various degrees of neurological impairment, as summarized in Table 2⇓. Two patients who initially presented with isolated intracranial hypertension were left with blindness due to optic atrophy. The 9 others, who had focal signs at the acute stage, were left with various cognitive or focal deficits as sequelae. One patient developed spasmodic torticollis 13 years after her initial CVT, which involved superior sagittal sinus, cortical and cerebellar veins, and left multiple cranial nerve palsy and cerebellar incoordination as permanent sequelae.
Seizures were observed only in patients who had seizures and focal signs in the acute stage, and only in a minority (4 of 28 patients [14.3%]). These seizures appeared in the first year after CVT, with the exception of 1 patient who developed breakthrough seizures 2 years after CVT during an attempt to taper this medication. The three patients were no longer on anticonvulsant agents when seizures recurred. These recurrent seizures were easily controlled with antiepileptic drugs.
One patient of 51 with a lateral sinus thrombosis developed an asymptomatic dural arteriovenous fistula that was seen on follow-up angiography 12 months later. In 3 other patients, a dural arteriovenous fistula was preexisting and thought to be the cause of the lateral sinus thrombosis.
Regarding the thrombotic risk, 9 patients (11.7%; 4 men, 5 women) suffered recurrence of CVT, all but one within 12 months of the first episode. Eight patients recovered completely; in 1, sequelae from the recurrent CVT added to the existing neurological impairment. None of the 9 patients with recurrence was on long-term anticoagulation. In 5, no cause could be found even during follow-up. By contrast, in 4 patients no cause was found during the acute stage, but a repeated workup performed for the recurrent CVT disclosed an underlying condition: Behçet’s disease in 3 and carcinoma in 1.
Proven thrombotic events other than CVT occurred in 11 patients (14.3%), including 2 of the above-mentioned patients with recurrent unexplained CVT who also suffered from systemic thromboses. These events ranged from retinal vein thrombosis to pulmonary emboli and repetitive thromboses of large arteries, and their morbidity was generally much more severe than that of the CVT. Four of the 11 patients were diagnosed as having Behçet’s disease; one event occurred during postpartum, one in a patient with inflammatory bowel disease, and one in a patient using oral contraceptives. The four other events remained without explanation despite a personal or family history of systemic thromboses.
Pregnancy After CVT
There were 16 pregnancies in 9 patients during follow-up. Two of these patients had originally presented with pregnancy-related CVT. Twelve pregnancies (during which 2 women received preventive subcutaneous heparin, whereas 2 were given aspirin alone) remained without complications and ended in normal delivery. The 4 other pregnancies terminated prematurely because of miscarriage (in 2 women) or voluntary abortion (in 2 women). No recurrence of CVT occurred during any pregnancy or during postpartum, regardless of previous etiology.
Although this was a retrospective study, it gives some insight into the long-term prognosis of CVT, which appears essentially good. The follow-up group constitutes a large and representative part of the primary series of 110 patients. The number of patients suffering from sequelae at the moment of discharge does not differ significantly in the two groups, and there is no reason to assume that the dropout patients as a group may have taken a substantially worse turn after discharge than the patients with follow-up.
In our series, 66 of 77 patients (86%) with CVT recovered without evidence of permanent neurological damage. Eleven patients were left with sequelae: optic atrophy in 2, seizures in 4, and cognitive or focal deficits in 5. The 2 patients with optic atrophy initially presented with isolated intracranial hypertension, whereas all the others had focal signs at onset. This confirms the well-known fact that these two main modes of presentation carry a very different prognosis regarding the risk of sequelae.7 8 12 Late neurological complications (other than the sequelae of the initial event) are extremely rare. One of our patients developed spasmodic torticollis 13 years after an extensive CVT, which involved cerebellar and cortical veins as well as superior sagittal sinus. The only other reported case is that of a 9-year-old boy who developed progressive dyskinesia 7 years after an internal cerebral vein thrombosis with bilateral thalamic lesions.13
Most of our patients (80%) received anticoagulant agents in the acute stage, whether or not they had an hemorrhagic infarct. No worsening was observed. However, since this was not a controlled study, it is difficult to conclude (although we favor this interpretation) that the low rate of sequelae observed in the present series is related to the wide use of anticoagulants.
The frequency of epilepsy beyond the acute thrombotic episode in our series is low: 5% of the whole series and 14% of those who had seizures at the time of their CVT. These results are in contrast to those in Nagpal’s series,14 the only one that yields some information about long-term outcome of a certain number of CVT patients. The author refers to 7 of 19 patients as “completely well” (36.8%), whereas 6 patients suffered from frequent epileptic seizures and 6 more from spasticity. It is difficult, from our series and from the scanty literature, to give firm recommendations about the duration of anticonvulsant treatment. However, given the low risk of recurrent seizures and the very low risk of late recurrences, it seems appropriate to maintain anticonvulsant therapy for a year and to taper off gradually thereafter. If seizures recur, anticonvulsants should be given on a long-term basis.
One patient of 51 with lateral sinus thrombosis developed a dural arteriovenous fistula, a well-established but rare long-term complication of lateral sinus thrombosis.15 16 The relationship between these two conditions is a complex one because, as in 3 of our patients, a dural arteriovenous fistula can be the underlying cause of the sinus thrombosis.17
The risk of recurrent CVT in our series is also low, and it all but disappears after the first year; the prognosis of the second CVT seems as good as that of the first event, regardless of whether there is a known etiology. None of the patients with recurrence was taking long-term anticoagulation because no underlying thrombotic disorder had been found at the acute stage. In four of these cases, however, repeated investigations performed for the recurrent CVT disclosed an underlying disease, such as Behçet’s disease, which would have required long-term anticoagulation.
Other thrombotic events (arterial and venous) occurred in 11 patients during follow-up. Seven of these events had an underlying cause or predisposing factors such as Behçet’s disease or postpartum. This stresses the need for a close long-term follow-up of patients with CVT to disclose and treat a cause that might have initially been overlooked. If such an underlying thrombotic condition is found, long-term anticoagulation is usually required after the first CVT, definitely if there is a recurrence. When no cause is found, long-term anticoagulation is not indicated after the initial CVT. However, if there is a recurrent CVT or another venous thromboembolic event, then there is a case for long-term anticoagulation, even in the absence of a known thrombophilia.
An important observation is that 12 pregnancies in 7 women with a history of CVT (only 2 of whom prophylactically received subcutaneous heparin) were carried out without recurrence of CVT. Although figures are low, these data suggest that systematic heparin treatment during pregnancy is not warranted unless there is an underlying thrombophilia or a known cause of venous thrombosis. However, given the fact that 1 of our patients developed deep vein thrombosis during postpartum, it might be justified to give low-dose heparin (or low-molecular-weight heparin) as preventive treatment during the first 2 postpartum weeks.
In conclusion, the long-term prognosis of CVT in the present series of 77 patients with a mean follow-up of 77.8 months was essentially good. The risks of long-standing epilepsy and of CVT recurrence were low. In particular, no recurrence was observed in women who later became pregnant. In most cases, there is no need for long-term anticoagulant or anticonvulsant treatment. However, these are required on a case-by-case basis in the few patients who have recurrent thrombotic (cerebral or systemic) events or recurrent seizures. In the absence of known thrombophilia, heparin treatment during further pregnancies does not seem to be required, but a short-term preventive treatment during the first 2 postpartum weeks is probably judicious. It should, however, be emphasized that our data are based on a retrospective study and that prospective studies are needed to better assess the long-term prognosis of patients with CVT.
Reprint requests to Pr M.G. Bousser, Service de Neurologie, Hôpital Saint-Antoine, 184 rue du Faubourg Saint-Antoine, 75571 Paris, Cédéx 12, France.
Review of this manuscript was directed by Editor-in-Chief Mark L. Dyken.
- Received April 26, 1995.
- Revision received October 9, 1995.
- Accepted October 9, 1995.
- Copyright © 1996 by American Heart Association
Barnett HJM, Hyland HH. Non-infective intracranial venous thrombosis. Brain. 1953;76:36-49.
Garcin R, Pestel M. Thrombophlébites cérébrales. Paris, France: K. Masson; 1949.
Kalbag RM, Woolf AL. Cerebral venous thrombosis. London, UK: Oxford University Press; 1967.
Krayenbühl HA. Cerebral venous and sinus thrombosis. Clin Neurosurg. 1967;14:1-24.
Krayenbühl HA. Cerebral venus thrombosis: the diagnostic value of cerebral angiography. Schweiz Arch Neurochir Psychiatry. 1954;74:261-287.
Bousser MG, Chiras J, Bories J, Castaigne P. Cerebral venous thrombosis: a review of 38 cases. Stroke. 1985;16:199-213.
Einhäupl KM, Villringer A, Haberl RL, Pfister W, Deckert M, Steinhoff H, Schmiedek P. Clinical spectrum of sinus venous thrombosis. In: Einhäupl K, Kempski O, Baethmann A, eds. Cerebral Sinus Thrombosis: Experimental and Clinical Aspects. New York, NY: Plenum Press; 1990:149-155.
Preter M. Langzeitprognose bei Hirnvenen und Sinusthrombosen. Eine Katamnese von 77 patienten. Munich, Germany: Ludwig-Maximilians-University; 1995. Doctoral thesis.
Barinagarrementeria F, Cantu C, Arredondo H. Aseptic cerebral venous thrombosis: proposed prognostic scale. J Stroke Cerebrovasc Dis. 1992;2:34-39.
Solomon GE, Engel M, Hecht HL, Rapoport AR. Progressive dyskinesia due to internal cerebral vein thrombosis. Neurology. 1982;32:769-772.