CVT was diagnosed using pre- and post-enhanced T1-weighted MRI (T1WI), CT, magnetic resonance venography, and/or cerebral angiography. The imaging study was performed at the time of hospital admission for initial diagnosis of CVT. To evaluate temporal changes to brain lesions, each patient was subjected to follow-up MRI and/or CT studies at discharge. Both direct and indirect signs of thrombosis in the intracranial venous system were used for the diagnosis of CVT. Direct evidence of CVT included thrombi in the cerebral vein and sinuses on any imaging modality.2 Intensified venous collaterals, the presence of cork screw veins, broken bridging veins, venous dilatation, and delayed venous emptying were considered as indirect evidence for CVT.2 Combined parenchymal lesions with hemorrhagic conversion were also confirmed by density or signal changes on CT, T1WI, and/or gradient echo images.2 An increase in the initial hemorrhage size and/or newly developed bleeding detected follow-up imaging was categorized as an LMWH-related hemorrhagic complication.

To evaluate the causes of CVT in patients, we assessed a history of risk factors that have been reported to be associated with CVT, such as oral contraceptive, parturition, infectious disease, ear disease, carcinoma, dehydration, marasmus, sepsis, acute lymphoblastic leukemia, chemotherapy, systemic lupus erythematosus, Behäet's disease, osteoporosis, diabetes mellitus, homocystinuria, and iron deficiency anemia.14 Laboratory blood tests were also performed to determine the presence of coagulopathy, such as complete blood count (CBC), liver function test (LFT), prothrombin time (PT), activated partial thromboplastin time (aPTT), levels of protein C and S, antithrombin III, lupus anticoagulant, and anticardiolipin antibody. If the above historical and laboratory evaluation for CVT revealed no abnormality, we classified the case as not-determined etiology.14

Twice daily subcutaneous injection of 7,500 IU nadroparine (Fraxiparine®, Sanoffi-Aventis, France) immediately followed the initial CVT diagnosis and was maintained for 2 weeks. For secondary prevention after the termination of the drug, oral anticoagulant (warfarin) or antiplatelet agent (aspirin) was prescribed depending on the individual risk factors.

Initial manifestations were evaluated at CVT diagnosis and closely monitored for changes during nadroparine treatment until the initial neurological symptoms were normalized or stabilized. We recorded the onset of symptoms and/or signs prior to CVT diagnosis. We also recorded the interval from initial nadroparine treatment to initial evidence of clinical improvement. This interval time was used to estimate the drug's efficacy at managing CVT. Initial improvement was defined as the first signs of recovery from subjective symptoms observed at admission, such as headache or nausea, and/or objective neurological deficits, such as motor deficit, aphasia, and mental deterioration. Finally, the interval from initial nadroparin treatment to normalization or stabilization of initial neurological alterations was evaluated. Exacerbations of initial neurological deficits and/or newly developed neurological deficits after treatment were classified as, "ineffective" or as, "side effect".

A total of 12 patients (male : female=5 : 7, age 21-76 years) were included in the present study (Table 1). Thrombi were noted in the cerebral venous sinus of all patients, serving as a direct evidence of CVT (Table 2). The thrombi were located in multiple sinuses in five patients. The venous thrombi were most commonly in the transverse sinus, followed by the superior sagittal and sigmoid sinuses. Associated parenchymal venous infarction was observed in nine patients. The parenchymal lesions were distributed in variable hemispheric regions, including the temporal, parietal, and occipital lobes. In patients 2 and 12, venous infarctions in deep structures, such as basal ganglia, thalami, and/or the centrum semiovale, were associated with the venous thrombi located in inferior cerebral veins (Table 2).

Common risk factors responsible for CVT were found in all but two patients (cases 6 and 7) (Table 1). Protein C and/or S deficiency were determined to be the most frequent cause, affecting 7 out of 12 patients. Deficiencies of both proteins were detected in three patients. Two patients had not only protein S and/or C deficiency, but also elevated levels of blood factor V (case 11) and blood factor VIII (case 12). Postpartum and oral contraceptives were checked in cases 4 and 8, respectively. Additionally, chronic otitis media in case 5 and osteomyelitis in the left lower tibia in case 9 were associated with protein C and/or S deficiency. Iron deficiency anemia (Hb=7.3 gm/dl) was detected in case 5. Case 2 had been diagnosed with stomach cancer 6 months previously but had no abnormality on the evaluated history and laboratory work-up. In all patients, warfarin was maintained at levels 2 to 3 of the international normal range of prothrombine time for secondary prevention of CVT.

Headache was the most frequent presenting symptom, and associated with neurological deficits, such as convulsive movement, sensory or global aphasia, hemiparesis, and hemianopsia. However, convulsive movement was the dominant manifestation without the headache symptoms in three patients. A mental status of deep drowsiness was the presenting manifestation in case 2. Interestingly, before the availability of imaging diagnosis for CVT, the likely presumptive diagnosis for case 2 was a seizure disorder. Other patients were initially determined as having variable diagnoses, with headaches in two patients, metastasis of stomach cancer, brain tumor, encephalitis, cerebral infarction, and meningitis (Table 1).

CVT was diagnosed rapidly (within 1 to 3 days) when the patients showed characteristic neurological deterioration at the hospital. However, the diagnosis was usually delayed (11 to 30 days from the initial onset of symptoms), when the patient first suffered from non-specific neurological manifestations, such as headache and nausea. Improvement of the initial symptoms was usually detected within 2 to 8 days after taking nadroparine, irrespective of a delay in the diagnosis of CVT. In two patients (case 3 and 8), improvement was detected after 20 days of nadroparine because clinical evaluations were impossible due to the inability to communicate. One patient showed sensory aphasia at admission, while another patient presented with a deep drowsiness during the management with nadroparine. Normalization to the previous neurological condition occurred rapidly in two patients, within 2 and 4 days in cases 1 and 12, respectively. Most patients normalized their initial neurological state within 11 to 60 days. Patient 9, who had marked thrombi and hemorrhagic conversion in both cerebral cortices, showed stabilization of right side hemiplegia by 222 days of nadroparine treatment, improving to Grade III.

Hemorrhagic conversion was detected in five of the nine patients wth venous infarctions: cases 3, 7, 8, 9, and 10. In case 9, severe thrombi in the superior sagittal sinus, inferior sagittal sinus, and sigmoid sinus (Fig. 1-B) and parenchymal lesion without hemorrhagic conversion were detected in the initial imaging study (Fig. 1-A), Unfractionated heparin was administered immediately following the CVT diagnosis, and 2 days later the mental state had decreased from alert to deep drowsiness. Motor weakness on the right extremities had also aggravated to hemiplegia. On the follow-up brain CT study, hemorrhagic conversion had developed in the previous infarcted lesions (Fig. 1-C). The mental state markedly improved 2 days after replacing heparin with nadroparin and no aggravation of clinical symptoms was detected during the maintenance with nadroparin.

Associated hemorrhage was observed in the initial brain image of four patients. Again, nadroparine did not cause clinical or radiological aggravation. In case 3, a marked temporal lobe hemorrhagic infarction was observed on the initial CT, (Fig. 2-A) and the left lateral sinus was evident in cerebral venography performed at admission. (Fig. 2-B) The initiation of nadroparine treatment resulted in no aggravation of mental status, eventhough the surrounding edema increased in the follow-up study (Fig. 2-C). The patient recovered to an alert and cooperative status without neurological findings after termination of the nadroparine schedule. Encephalomalatic change at the hemorrhagic converted parenchymal lesion remained in case 9 (Fig. 1-D) and 4. (Fig. 2-D) Convulsive movement developed after hospital discharge in both of these patients, and an anti-epileptic drug was prescribed to prevent recurrence. Hemorrhagic-converted cerebral lesions were observed in the initial imaging study in three other patients. These patients showed clinical improvement after taking nadroparine, and no aggravation of the initial manifestations as well as radiological problems were observed. However, hemorrhagic converted lesions remained as encephalomalatic changes in the follow-up imaging studies.