Dural sinus thrombosis after resection of vestibular schwannoma using suboccipital retrosigmoid approach—thrombosis classification and management proposal

Dural sinus thrombosis is one of the complications after posterior fossa surgery. However, that topic is not described well with regard to vestibular schwannoma surgery using the unique suboccipital retrosigmoid approach. We analyzed retrospectively medical records and radiological investigations of 116 patients. The including criteria were histopathologically confirmed vestibular schwannoma operated on using the retrosigmoid approach, preoperative and postoperative contrast-enhanced MRI, and at least 1-year follow-up. The patient group included 36% males and 64% females. The average age was 47.3 ± 13.9 years. Sixty percent of the tumors were classified as T4b according to the Hannover scale and their mean volume was 13.73 ± 10.28 cm3. There were no signs of thrombosis preoperatively. Postoperative changes in the dural sinuses were found in 26 (22%) cases. In 7 (27%) cases, there was an external compression by the hemostatic agent, and in 19 (73%) cases, a thrombus was visualized in the sinus lumen. The size of the sinus, age, and the tumor size were not risk factors for thrombosis, whereas an intraoperative sinus injury was a statistically significant risk factor (p = 0.0012). All of the patients diagnosed with thrombosis were in good clinical condition in long-term follow-up, except one fatal case. Complete recanalization was observed in 58% of cases after 1-year follow-up. Postoperative changes in the dural venous sinuses are a frequent finding after vestibular schwannoma surgery using the suboccipital retrosigmoid approach. Intraoperative dural injury is a risk factor for thrombosis. Thrombosis in that group of patients is usually asymptomatic and does not influence the prognosis.


Introduction
Dural sinus thrombosis (DST) is a complication after posterior fossa surgery, especially in the cases of vestibular schwannoma surgery, and its reported incidence ranges from 4.7 to 38.9% [1][2][3]5]. Various intraoperative risk factors, which could cause such a complication, have been described: injury of the dural sinus, long duration of the surgery, surgical mobilization of the sinus, bone skeletonization around the sinus, thermal injury caused by coagulation, or even heat generated by a microscope lamp [1]. Depending on the size of the affected sinus, extent of the thrombus, and individual vascular anatomy, there could be a wide range of symptoms, from clinically asymptomatic through mild, like headache, to even severe, like intracerebral or intracerebellar hemorrhage, hydrocephalus, and death [9]. There is no evidence-based data supporting the management of postoperative sinus thrombosis. The first choice approach to spontaneous cerebral sinus thrombosis is aggressive anticoagulation; however, such treatment is contraindicated in the early postoperative period, because of the risk of hematoma. This is why the treatment strategies vary from more conservative approach to administration of anticoagulants. The literature lacks information focusing on presentation of the thrombus present in postoperative imaging and radiological classification, and there is no algorithm proposed for therapeutic approach to postoperative lateral sinus thrombosis after vestibular schwannoma resection using the suboccipital retrosigmoid approach.

Data source and study design
We retrospectively analyzed all patients operated on for cerebellopontine angle occupying lesions using suboccipital retrosigmoid approach in the single institution between January 2014 and June 2020. All patients were positioned supine with the head rotated contralaterally. The inclusion criteria were as follows: (1) histopathological confirmation of the vestibular nerve schwannoma, (2) patients ≥ 18 years old, (3) exclusively retrosigmoid suboccipital craniotomy or craniectomy, (4) preoperative and early (during the first 3 days) postoperative contrast-enhanced MRI imaging available, (5) at least one follow-up imaging. The minimal interval between the surgery and follow-up imaging was set up to 1 year. We excluded the reoperations from the analysis.
A medical database was searched using the D33.3 code from the International Statistical Classification of Diseases and Related Health Problems ICD-10. Patients with a lesion occupying the cerebellopontine angle and histopathological confirmation of vestibular schwannoma were included into the study. Age, sex, comorbidities with special attention to thrombophilia and thrombosis history, size and volume of the tumor, postoperative anticoagulation therapy, symptoms of intracranial hypertension (headache, vomiting, disorders of consciousness), intraoperative injury of the dural sinus and its surgical management, time of surgery, and length of hospitalization were taken into analysis. The length of hospitalization took into account the patients' stay both in the neurosurgical unit and in other units (neurology, rehabilitation).
Sizes and volumes were measured using the commercial neuronavigation software (Cranial v 4.0, BrainLab, Germany).
In our institution, we perform gadolinium-enhanced MRI imaging using T1-weighted sequence without and with contrast, T2 sequence, and FLAIR preoperatively and postoperatively in 1 to 3 days after the surgery. In order to assess the patency of venous sinuses, contrast-enhanced T1 was used. Magnetic resonance venogram was not routinely used in most cases. Radiological investigations were analyzed independently by the first author (KK) and the consulting neuroradiologists (KS, CW), and focused on the dural venous sinus postoperative changes, preoperative and postoperative tumor characteristics (size according to the Hannover grading system, preoperative and postoperative cerebellar or brainstem edema, cyst presence, volume, extent of resection), signs of intracerebellar hemorrhage, and dural sinuses dominance. When a diagnosis of dural venous thrombosis was established, the location of the affected sinus, its dominance, presence of marginal flow, and dilatation of the cerebellar veins were analyzed using the radiological scale prepared by the main author (KK).
For patients with radiologically diagnosed postoperative changes (compression or thrombosis), further details about the clinical course were sought. Sign consents were obtained from each study participant before further proceedings. The patients were examined neurologically in the outpatient clinic. Long postoperative course was investigated for early and late complications like headaches, signs of intracranial hypertension, liquorrhea, or wound dehiscence. Radiological follow-up was analyzed and particular attention was paid to the width of the affected sinus, thrombus resorption, and the degree of resorption.

Statistical analysis
Continuous variables were presented as mean values with standard deviation. Univariate analyses were performed to compare patient characteristics and the rate of complications between patients with and without postoperatively compromised sinus. Pearson's chi-square test and Fisher's exact test were used for analysis of categorical data, in association with the occurrence of postoperative sinus thrombosis. Statistical significance was defined as a p value of 0.05 or less.

Ethics
The Institutional Ethics Committee was informed about the ongoing study, and the Committee stated no need for its approval. Informed consent was obtained from all individual participants included in the study.

Patients and tumor characteristics
During the analyzed period of time, 116 patients met the inclusion criteria and were included in the analysis. Among these patients, 74 (64%) were women and 42 (36%) were men. One patient had known thrombophilic coagulopathy-antiphospholipid syndrome. The mean age was 47.3 ± 13.9 years. The average tumor size was 31.4 ± 11.4 mm in the anteroposterior dimension, 28.8 ± 9.8 mm in the mediolateral dimension, and 28.6 ± 9.9 mm in the superoinferior dimension. Sixty-three tumors (54%) were operated on the right side and 53 (46%) on the left side. The mean tumor volume was 13.73 ± 10.28 cm 3 . According to the Hannover scale, 4 tumors (3%) were classified as grade T1, 4 (3%) as grade T2, 8 (7%) as grade T3a, 8 (7%) as grade T3b, 23 (20%) as grade T4a, and 69 as grade T4b (60%). The mean duration of the surgery was 4 h and 44 min ± 1 h and 37 min. The mean length of stay (LOS) was 13.9 ± 9.8 days (range 6 to 76 days).

Dural sinus thrombosis
In 26 (22%) cases, changes in the contrast filling of the lateral sinuses were noted. In 7 (27%) cases, the sinus was narrowed because of external compression, caused by the hemostatic agent used for the dural closure purposes (Fig. 1). It could be easily recognized when preoperative and postoperative images were compared. In the postoperative view, the internal margin of the sinus was not smooth; however, the sinus lumen was filled with contrasted blood. The remaining cases presented true thrombosis, with the area inside the lumen lacking the contrast filling. In 8 (31%) cases, the thrombus filled less than a half (Fig. 2) and in 6 cases (23%) more than a half of the sinus cross-section (Fig. 3). In 2 cases (8%), there was almost complete closure of flow through the sinus except for narrow marginal contrast filling (Fig. 4). In 2 cases (8%), a complete obstruction of contrast filling was observed, however, without engorgement of the draining veins ( Fig. 5), which was seen in one case (3%) (Fig. 6). A certain pattern could be noted. This is why we developed a classification of postoperative changes seen in the lateral sinuses ( Table 1).
None of the patients diagnosed with thrombosis had known thrombophilia. All cases were ipsilateral to the performed surgery. In 10 cases (38%), the affected sinus was dominating, and in 6 cases symmetrical (23%). In 12 cases (46%), the thrombosis was located on the right side, and in 14 on the left (54%). Dominance of the affected sinus was not a significant factor for thrombosis (χ 2 , p = 0.16). The patient group included 18 females (69%) and 8 males (31%). Differences between sexes did not reach the level of significance (χ 2 , p = 0.17). Two patients (8%) had T3a grade tumors, 1 patient (4%) had a T3b grade tumor, 6 patients had T4a (23%), and 17 (65%) had T4b tumors. The tumor size in the Hannover scale was not a significant risk factor for thrombosis (χ 2 , p = 0.82), nor was the tumor volume (13.51 ± 11.21 cm 3 for the affected group and 13.6 ± 10.1 cm 3 , U Mann-Whitney test, p = 0.57). The mean age of patients in the thrombotic group was 46 ± 14 years and did not differ significantly from the non-affected group (U Mann-Whitney test, p = 0.61). Mean time of surgery was also similar in both groups (U Mann-Whitney test, p = 0.58). Mean LOS was 14 ± 13 days and was similar to the control group (U Mann-Whitney test, p = 0.53). The dural sinus was injured intraoperatively in 6 cases (23%) in the thrombotic group and in 3 cases (3%) in the non-affected group, and was a statistically significant risk factor for lateral sinus thrombosis (χ 2 , p = 0.0012). The differences between the thrombotic and non-thrombotic group are presented in Table 2.
None of the patients presented with the symptoms suggesting thrombosis, except one patient, who developed signs of increased intracranial pressure on the first postoperative day. However, there was a major intraoperative venous bleeding from the incidentally avulsed superior petrosal vein and thereafter from the superior petrosal sinus. In the early postoperative course, the patient developed cerebellar edema, intracerebellar hemorrhage, and obstructive hydrocephalus. An emergency reoperation with the posterior fossa decompression was performed and an external ventricular catheter was employed. After 14 days of treatment in the ICU, the patient died. The postoperative CT revealed massive cerebellar edema, a small intracerebellar hematoma, and signs of the lateral sinus thrombosis (grade II), which was confirmed by contrast-enhanced MRI. The most possible explanation was a venous stroke, caused by intraoperative, incidental closure of the superior petrosal vein during the surgery, or lateral DST. The remaining patients were classified as 5 in the GOS (19 cases, 73%) and 4 (1 case, 4%). Five patients were lost during the follow-up (19%). The mean follow-up time was 21.75 ± 11.77 months.
Postoperative anticoagulation therapy was instituted in 10 cases (38%). Enoxaparin in prophylactic dose was used in all cases, according to the body weight. Oral   In 7 cases (27%), the thrombus was located only in the sigmoid sinus; in 6 cases (23%), exclusively in the transverse sinus; in 4 cases (15%), in the transverse-sigmoid junction and sigmoid sinus; in 3 cases (12%), in the transverse-sigmoid junction and transverse sinus; and in 6 cases (23%), in the sigmoid, transverse, and sinus junction. In 2 cases (8%), the thrombus was located also in the dural sinuses within the tentorium.
Nineteen patients (73%) reached at least 12 months of the follow-up and were under long-term postoperative control. Seven patients (23%) were lost in the follow-up-they changed the neurosurgical center, changed their phone number, or did not answer the call. In all cases, some degree of thrombus resorption was observed. In 8 cases (42%), there was a partial recanalization, and in 11 cases (58%), the postoperative changes resolved completely. None of them presented symptoms which could be connected with thrombosis.

Discussion
Suboccipital retrosigmoid craniotomy is an accepted approach for the vestibular schwannoma surgery. It provides wide exposure of the cerebellopontine angle and gives an opportunity to resect large tumors and preserve hearing in some cases. Lateral sinus thrombosis is a known complication after the posterior fossa surgery and has been noted in 4.7-11.6% of cases [2][3][4][5][6][7][8]. However, there is limited data about that complication exclusively in the vestibular schwannoma surgery. There is a series of studies [1,4,10]; however, they have some limitations. In most of them, non-contrast CT was used as a postoperative control [1]; the analyzed cohort included various surgical approaches [1,3,10] or translabyrinthine approach [5]. There are no large cohort studies with well-documented postoperative radiological investigation on the retrosigmoid approach in the vestibular schwannoma surgery. What is more, to the best of our knowledge, there is no classification of postoperative DST in the literature. The authors of previous studies managed the topic noting only the fact of presence or absence of the thrombus.
The risk factors of deep venous thrombosis after skull base surgery are well described [4]. The main concern is to prevent pulmonary embolism, which could be fatal. On the other hand, hyperemia and venous hypertension, which could cause intracranial hypertension and cerebral hemorrhage, are more problematic in thrombosis of the dural sinuses. What is more, the risk factors and management of the lateral dural venous sinuses are less evident than in the case of deep venous thrombosis.

Symptomatology
Various symptomatology can be presented in the venous sinus thrombosis: headaches, seizures, intracranial hypertension, intraparenchymal hemorrhage [1,6,8]. Some cases of fatal course of sinus thrombosis have even been presented [9]. However, most studies emphasize mild or even asymptomatic course of lateral sinus thrombosis [1-3, 5, 8]. In our study, most of the patients, except one, were completely asymptomatic. It is important to emphasize that some non-specific signs and symptoms (vertigo, ataxia, headache) could be similar both for the uncomplicated postoperative course and DST. However, in none of cases, except one, the described symptoms did not correlate with radiological signs of venous hypertension (no cerebellar edema, infarction, or intracerebellar hemorrhage). In one case, a massive cerebellar edema developed, which could be explained by the large sigmoid sinus thrombosis. What is more, all of the patients remain  [1-3, 5, 8] that postoperative DST is rarely symptomatic in such cases. This is likely due to collateral venous drainage, making thrombosis well-tolerated.

Risk factors
Duration of the surgery, the tumor volume and size, and the patient's age or sex were not risk factors of postoperative thrombosis. Moreover, we did not observe the correlation between thrombosis and smaller sinus diameter observed by other authors [1,3]. It is an interesting fact, conflicting with the assumption that lower caliber of the vessel and slower flow could increase the risk for thrombosis in that group of patients. The most important risk factor seems to be an intraoperative injury of the sinus, which was observed in 23% of cases in the thrombosis group and only in 3% of cases in the non-thrombosed group. It means that sinus exposure is the main risk factor of thrombosis and the main prophylactic activities should focus on minimizing the sinus manipulation. Some authors theorized that even heat conduction during bone drilling could facilitate the thrombotic cascade [1]. In our group, the most common moment of sinus injury was the moment of passing the footplate parallel to the sigmoid sinus. We have noticed that in most of such cases, we injured not the sigmoid sinus itself, but the mastoid emissary vein which connects the venous drainage of the brain with the superficial scalp venous network. During craniotomy, the vein was usually ruptured and avulsed, causing also bleeding from the sigmoid sinus. We manage that complication by changing the strategy of craniotomy. Nowadays, we perform craniotomy to the point where the vein exits the skull through the foramen. The remaining part of the bone is drilled using the high-speed bone cut and diamond drills, to the level of the medial margin of the sinus, which is sufficient for the vestibular schwannoma surgery. After such modification of the craniotomy technique, we observed a markedly lower incidence of the postoperative sinus thrombosis. However, further analysis on that topic is needed.

Incidence
The incidence of DST after posterior fossa surgery was previously estimated at 11.6% [8]. Recent studies have noted more frequent diagnoses of thrombosis-32.4% [3] and 38.9% [9]; however, there are studies with lower frequency-6% [1] and 6.7% [10]. The reported incidence depends on the character of the study, retrospective or prospective; the imaging modality, contrast-enhanced or not; and on the surgical approach. In our retrospective study, where every analyzed case was investigated using contrastenhanced MRI, we noted postoperative changes in 22%.
It must be noted that in 7 cases the sinus was narrowed by the hemostatic agent used for the dural closure, which does not mean thrombosis at all. None of the patients from that group developed any symptoms. Long-term postoperative control revealed complete resolution of compression in 5 cases and partial resolution in 1 case. One patient was lost to follow-up. Such presentation seems to be benign and asymptomatic. To the best of our knowledge, such radiological symptom has never been described nor analyzed in the neurosurgical literature. True thrombosis was noted in 19 cases, which account for 16% of the studied patients. As mentioned above, the studies analyzed lesions of various surgical approaches, different pathologies, and non-contrast postoperative examinations. To the best of our knowledge, our study is the largest study focusing on vestibular schwannoma operated on using the retrosigmoid approach with postoperative contrast-enhanced MRI.

Anticoagulation and management
The main concern in DST is the prevention of the thrombus propagation, obstruction of the blood drainage, and increased intracranial pressure due to hyperemia or intracerebral hemorrhage. Such a situation could be fatal, as reported [5]. The role of anticoagulation is to prevent such thrombus progression. However, the usage of anticoagulants in the early postoperative course is always a dilemma. What is more, there are no guidelines managing that topic. Despite the relative clinical safety of the low molecular weight heparin (LMWH) in the prophylactic dosage, the use of higher, therapeutic doses seems to be more risky and thus is avoided postoperatively. However, even low dosage could cause hemorrhagic complications, as reported by other authors [1]. In our group, we did not observe any complications which could be associated with the used anticoagulation. We manage thrombosis basing on its extent. The sinus compression or a small thrombus (grades I and II) was managed only by IV administration of fluids and clinical observation. In more prominent cases (grades III-VI), we administered LMWH in prophylactic dosage and fluids. The strongest indications for anticoagulation include especially venous stasis and engorgement of the draining veins (grade VI). In our group, anticoagulation was administered in 10 cases (38%), never earlier than 24-48 h after the surgery, and was discontinued before the discharge. In all cases, enoxaparin in prophylactic dose was used, according to the body weight. Oral anticoagulants were not prescribed as a continuation of the therapy. Based on our results, it is debatable if anticoagulation is necessary in such cases. Almost all of them were asymptomatic. In some of the cases, the thrombus was noticed retrospectively and such patients had been managed without anticoagulation. Only one case, described earlier, could be associated with venous thrombosis. However, in that particular case, the superior petrosal vein was closed during the surgery, so the explanation of that case is complex. On the other hand, we did not observe any hemorrhagic complications in that group of patients. Using our thrombosis classification, we hypothesize that the cases where a thrombus is causing the engorgement of other draining veins (grade VI) are the most important indication for the administration of anticoagulants in that group of patients. On the basis of our results obtained in such cases, we recommend maintaining the LMWH in the prophylactic dose for 3 months and evaluate the sinus patency and draining vein size using MRI + C and MRV. In the case of normalization of the vein size and recanalization of the sinus lumen, we would discontinue the therapy. With respect to partial recanalization with normalization of the draining vein size, we also recommend discontinuation of the anticoagulant. We would maintain the therapy for another 3 months and increase the dose to therapeutic, if the veins were still enlarged prominently, which suggested venous congestion in the surrounding brain tissue.
The cases with progressive neurological deterioration caused by venous hypertension and progressive intracranial hypertension are another issue. It is important to emphasize that we did not observe such a manifestation of DST in our group, except one case, which was debatable and discussed earlier. However, from the pathophysiological point of view, such clinical presentation is possible and should be discussed. It is important to admit that no large, controlled clinical studies addressing that topic have been conducted. In our clinical practice, we applied the guidelines for spontaneous DST [11]. In the cases of progressive neurological deterioration caused by postoperative DST, with progressive intracranial hypertension not controllable by medical treatment (mannitol therapy, anticoagulation), invasive therapy should be considered. If a prominent and progressive mass effect is seen in the serial CT examination, a surgical decompression should be employed. Especially if the mass effect is present in the cerebellar hemisphere, an urgent posterior fossa craniectomy and external ventricular drainage should be performed. If a supratentorial mass effect is seen, a decompressive hemicraniectomy could be considered a salvage therapy for intracranial hypertension.
In the cases of a massive, symptomatic, and life-threating thrombosis, endovascular treatment would be indicated. Either direct administration of fibrinolytic agents or a mechanical removal of the thrombus could be considered, as described by other authors [12,13].

Conclusions
Postoperative changes in the dural venous sinuses are a frequent finding after vestibular schwannoma surgery using the suboccipital retrosigmoid approach. Intraoperative dural injury is a risk factor for thrombosis. Thrombosis in that group of patients is usually asymptomatic and does not influence the prognosis. Availability of data and material Available.
Code availability Not applicable.

Declarations
Ethics approval The Institutional Ethics Committee was informed about the ongoing study, and the Committee stated no need for its approval.
Consent to participate Informed consent was obtained from all individual participants included in the study.