Madjid Samii, M.D., Cordula Matthies, M.D.
Department of Neurosurgery, Nordstadt Hospital, Hannover, Germany
OBJECTIVE: To identify the actual benefits and persisting problems in treating vestibular schwannomas by the suboccipital approach, the results and complications in a consecutive series of 1000 tumors surgically treated by the senior author were analyzed and compared with experiences involving other treatment modalities.
METHODS: Pre- and postoperative clinical statuses were determined and radiological and surgical findings were collected and evaluated in a large database for 962 patients undergoing 1000 vestibular schwannoma operations at Nordstadt's neurosurgical department from 1978 to 1993.
RESULTS: By the suboccipital transmeatal approach, 979 tumors were completely removed; in 21 cases, deliberate partial removal was performed either in severely ill patients for decompression of the brain stem or in an attempt to preserve hearing in the last hearing ear. Anatomic preservation of the facial nerve was achieved in 93% of the patients and of the cochlear nerve in 68%. Major neurological complications included 1 case of tetraparesis, 10 cases of hemiparesis, and caudal cranial nerve palsies in 5.5% of the cases. Surgical complications included hematomas in 2.2% of the cases, cerebrospinal fluid fistulas in 9.2%, hydrocephalus in 2.3%, bacterial meningitis in 1.2%, and wound revisions in 1.1%. There were 11 deaths occurring at 2 to 69 days postoperatively (1.1%). The techniques that were developed for avoidance of complications are reported. The analysis identifies preexisting severe general and/or neurological morbidity, cystic tumor formation, and major caudal cranial nerve deficits as relevant risk factors.
CONCLUSION: The current treatment options of complete tumor resection with ongoing reduction of morbidity are well fulfilled by the suboccipital approach. By careful patient selection, the mortality rate should be further reduced to below 1%. (Neurosurgery 40:11-23, 1997)
Key words: Acoustic neuroma, Caudal cranial nerves, Complications, Facial nerve, Middle fossa approach, Mortality, Suboccipital approach
Although the diagnosis of a vestibular schwannoma (VS) (acoustic neuroma) was like a death sentence at the beginning of the century and a sentence to permanent disability at the middle of the century, the situation has dramatically changed within only 3 decades. The introduction of the operating microscope at the beginning of the 1960s and the revolutions in radiological diagnostics in the 1970s and 1980s contributed to making earlier diagnosis and adequate, refined treatment possible. At the end of this century, management of patients with neuromas seems to be almost routine.
There is a general agreement that completeness of resection and preservation of the facial nerve are the major goals (40), and they are being met at increasing rates. There is also agreement that any of the available approaches, such as the suboccipital, the middle fossa (MF), and the translabyrinthine (TL) approaches, and their modifications have their indications. With training and experience, surgeons can develop these approaches to high standards with optimum patient safety regarding mortality and morbidity.
During the past 16 years, the suboccipital approach has been routinely used for all variants of VSs at our institution. Based on 1000 VSs that were surgically treated by one surgeon (the senior author, M.S.), one of the most variable factors, statistically speaking, is excluded and a unique opportunity is offered to analyze the data on a far more homogenous basis. Moreover, the development, with increasing experience, of the latest refined technique is presented with an analysis of postoperative outcomes, complications, and their learning value.
From 1978 to 1993, 1000 VSs (acoustic neuromas) were removed in 962 patients using the suboccipital approach. Eighty-two patients had evident neurofibromatosis-2 (NF-2), and 120 bilateral tumors were treated surgically at our institution. Eight hundred eighty patients were free of any NF-2 features and were treated for unilateral tumors.
With the patient in the semi-sitting position (45, 49, 56) and the patient's head tilted and rotated 30 degrees to the involved side, a suboccipital craniectomy is performed exposing the borders of the transverse and sigmoid sinuses. The dura is opened by a laterally convexed dural incision, and after cerebrospinal fluid (CSF) drainage is achieved by opening the cerebellomedullary cistern, the cerebellum is mildly retracted so that the posterior wall of the internal auditory canal (IAC), possibly a part of the tumor, and the brain stem become visible. The dura from the posterior aspect of the IAC is removed, and the canal is drilled open by the use of decreasing sizes of diamond drills until the intrameatal tumor extension is exposed; the posterior and lateral semicircular canals as well as the jugular bulb are respected (51), and the distance to the fundus is intermittently measured by platelet-shaped knives 2 mm in diameter. If the tumor is of normal consistency, partial tumor mobilization inside the IAC is initiated. The extrameatal intracapsular tumor reduction with either the platelet-shaped knife or the Cavitron ultrasonic surgical aspirator is then accomplished, thereby continuously decreasing compression of surrounding neural and vascular structures. If the tumor is of highly increased consistency, primary mobilization in the IAC might be impossible without severance of the cochlear nerve; enucleation in the IAC or extrameatal then needs to be started. Under continuous saline irrigation by the "third hand" (the neurosurgical assistant), precise bimanual nerve preparation is performed and bipolar coagulation is reduced to a minimum and left up to the end of surgery for final hemostasis. Hereby, maximum safety to brain stem and nerve vascular supply is guaranteed; early coagulation of tumor capsule vessels branching from nerve feeding vessels is especially prohibited, and induction of vasospasm is prevented. As soon as the tumor mass is largely removed, the relation between tumor border and neural structures is relaxed. The final preparation and liberation of neural structures is performed by strictly gripping only the arachnoid sheath with the forceps and by limiting overstretching in one direction (25). The tightest junction between the tumor and the nerve vessel bundle is encountered just before their entrance into the IAC (54) is prepared at the end; at that point, sharp dissection might become necessary to prevent overstretching of the facial nerve. Its integrity is tested by continuous facial electromyography and by electrical stimulation at the end. Palpation with the platelet-shaped knife at the fundus and, in very few cases, where drilling is limited by the labyrinthine structures (24), testing with a micromirror or the use of an angulated microendoscope (53) ensure completeness of intrameatal tumor removal. Jugular venous compression is performed by the anesthesiologist to make any opened or torn veins visible for final hemostasis. The opened mastoid cells at the IAC are carefully closed, for prevention of an intradural CSF otorhinological fistula, using bone wax and a piece of muscle sealed with fibrin glue. After a last careful inspection of the cerebellopontine angle (CPA) and the removal of the retractor, continuous water-tight dural closure is achieved.
All of the patients were prepared for surgery by a thorough clinical investigation, including otorhinolaryngological investigations, computed tomography (CT) at bone windows, contrast-enhanced CT or gadolinium-enhanced magnetic resonance imaging (MRI), and functional x-rays of the cervical spine. Patients with neurofibromatosis underwent preoperative MRI of the spine for diagnosis of tumors with spinal cord compression; then modified positioning with only slight inclination and rotation under neurophysiological control was considered. Additional craniocervical tumors were operated on either before or at the event of cerebellopontine angle surgery, with extended exposure achieved by opening the foramen magnum and possibly by the removal of the C1 and C2 arches.
Postoperative care included an average of 1 day in an intensive care unit and, thereafter, mobilization under physiotherapeutic guidance, control of audiometry and CT 1 week after surgery, and discharge from the hospital after 8 to 14 days. The neurological results at discharge are reported. Clinical, otorhinolaryngological, and MRI or contrast-enhanced CT follow-up examinations are scheduled for 1, 2, and 5 years postoperatively. Patients with facial nerve paresis and facial nerve reconstructions are examined every 3 to 6 months. Patients with special hearing problems are also scheduled for 3- to 6-month follow-up examinations.
Tumor sizes were measured considering intra- and extrameatal tumor extension; large tumors were larger than 30 x 20 mm, and small tumors measured up to 30 x 20 mm. Tumor extension was described as follows: Class T1, purely intrameatal; Class T2, intra- and extrameatal; Class T3a, filling the cerebellopontine cistern; Class T3b, reaching the brain stem; Class T4a, compressing the brain stem; Class T4b, severely dislocating the brain stem and compressing the fourth ventricle.
Fifty-six patients had undergone surgery in other hospitals; 24 patients underwent subtotal tumor resection, and 32 underwent biopsy.
In 979 cases, tumor removal was complete. In 21 cases, deliberate subtotal tumor removal was performed, with the aim of lifesaving surgery in 10 cases and of hearing-saving surgery in 11. In 10 cases (nine patients), brain stem decompression was performed in six elderly disabled patients and in three patients with NF-2. Bilateral tumor surgery was performed in one patient, and monolateral tumor surgery was performed in two patients (Fig. 1, Example 1). In 11 patients with NF-2, cochlear nerve decompression for preservation of hearing was attempted. By opening the IAC and by tumor reduction, as long as brain stem auditory evoked potentials were satisfying, hearing was preserved in 8 of the 11 patients, and hearing quality was stable in the last hearing ear 6 years postoperatively. In two patients, intraoperative brain stem auditory evoked potential and immediate postoperative hearing were lost. In one patient, brain stem auditory evoked potential and hearing were preserved but were lost 2 weeks after the operation; this same patient underwent a subsequent operation because of tumor regrowth 5 months later. Details of surgical and monitoring techniques for hearing-conserving surgery are reported separately (30, 46, 48).
Recurrences occurred in 6 of 880 patients (0.7%) who did not have NF-2. One patient with a large brain stem-compressing, hemorrhagic tumor and preoperative facial paralysis showed a recurrence of the same size and type of tumor within 1 year and underwent a subsequent operation. This patient remained free of any further recurrence for 18 months. One patient who was operated on at our institution and for whom hearing preservation was achieved noticed deterioration and final loss of her initially preserved hearing 4 years postoperatively. She underwent a subsequent operation for the recurrence of a 25-mm tumor. In one patient, the recurrence was discovered during a routine MRI study and surgical resection was performed a second time. Three patients have not undergone subsequent surgery. The recurrence rates and management of NF-2 are discussed in detail elsewhere (48).
The incidences of patients' subjective postoperative symptoms are listed in Table 1 and are compared with preoperative symptom incidences (29). During the first 2 to 8 postsurgical weeks, patients complained of headaches (9%), swallowing difficulties (4.5%), subjective trigeminal hypaesthesia (7%), trigeminal paraesthesia (2%), tinnitus (35%), and dizziness or unsteadiness (56%). Imbalance when walking was the most frequently reported vestibular disturbance (35%), vertigo was the second most frequent (25%), and dizziness was the third (22%).
TABLE 1. Summary of Cranial Nerve Symptoms
|Cranial Nerve||Symptom||Incidence Immediately before Surgery (%)||Incidence 2-8 wk after Surgery (%)|
|Trigeminal nerve disturbances||9-7||10-7|
|Subjective trigeminal hypesthesia in V2||9||7|
|Subjective trigeminal paresthesia in V2||3||2|
|IX-XII||Caudal cranial nerve disturbances||2.7||4.5|
In 1000 operations in 962 patients, regardless of hearing function, the neurological status was postoperatively normal in the majority of cases. In these patients, the general physical condition, the mental status, the level of alertness, the status of possibly involved cranial nerves (i.e., the trigeminal nerve), the caudal cranial nerves, and the facial nerve were normal or recovering.
The facial nerve was anatomically preserved in 929 patients (93%). Of these patients, 51% (470 of 929 patients) experienced normal facial nerve function immediately postoperatively and at the time of discharge from the hospital (House-Brackmann [HB] Score 1) (Table 2). Forty-five percent (420 of 929 patients) experienced reduced facial nerve function, with good potential for recovery occurring within 1 to 12 months. Their HB scores were as follows: 13% (120 patients), HB Grade 2; 15% (140 patients), HB Grade 3; 6% (60 patients), HB Grade 4; and 11% (100 patients), HB Grade 5. Four percent (39 of 929 patients) experienced facial paralysis despite nerve continuity. Of those, 2.3% recovered spontaneously but 1.7% showed no recovery. Of those showing no recovery, 11 received hypoglossal-facial nerve anastomosis achieving good reanimation. Six patients refused any nerve reconstruction procedure.
The facial nerve was anatomically lost in previous outward operations in 11 cases. In two cases, a nerve reanimation was achieved by hypoglossal-facial nerve combination, and in nine cases, previous plastic procedures had been performed.
The facial nerve was anatomically severed in 60 cases. In 42 cases, facial nerve discontinuity was treated by reconstruction using a nerve graft. In 23 cases, the facial nerve was reconstructed in the CPA using a sural graft of 0.5 to 3 cm. In 16 cases, reconstruction was performed from the CPA to the mastoid segment (intracranial-intratemporal transplantation, Samii's method) (11, 43, 44), and in 3 cases, Dott's method of intracranial-extracranial transplantation from the CPA to the external segment at the stylomastoid portion was performed.
TABLE 2. Summary of Cranial Nerve Signs
|Cranial Nerve||Neurological Sign||Incidence Immediately before Surgery (%)||Incidence 2-8 wk after Surgery|
(%) of All 1000
Operations (of Those 929 with Nerve Continuity)
|V||Trigeminal nerve hypesthesia||11-16||8-12|
|Trigeminal paresthesias or dysesthesias||1-3||0.5-1.7|
|House-Brackmann Grade 1||83||47 (51)|
|House-Brackmann Grade 2||8||12 (13)|
|House-Brackmann Grade 3||6||14 (15)|
|House-Brackmann Grade 4||0.7||6 (6)|
|House-Brackmann Grade 5||1||10 (11)|
|House-Brackmann Grade 6||1.3||11 (4)|
|IX-XII||Caudal cranial nerve disturbances||3.5||5.5|
In 18 cases, the facial nerve was lost at its origin at the brain stem and the facial nerve was reanimated by a donor nerve (by the contralateral facial nerve in 2 cases and by the hypoglossal nerve in 16 cases). Details of the indications for and the techniques of reconstruction, of the guidance given to patients after surgery, and of the outcomes are reported elsewhere (49).
The cochlear nerve was anatomically preserved in 682 cases (68%), of which 102 ears were deaf preoperatively. The cochlear nerve was lost in 299 cases (30%), of which 147 ears were deaf and 152 were hearing preoperatively. In 19 cases (2%), the nerve had been severed previously in operations that were performed elsewhere.
Of 732 hearing ears preoperatively, 580 cochlear nerves (79%) were preserved anatomically and 289 nerves were preserved in function with 39 patients with good, 115 patients with fair, and 135 patients with bad hearing. Discrimination was useful in 79% of postoperatively hearing patients. Further details are reported separately (46).
In summary, anatomic cochlear nerve preservation was 68% and functional preservation was 39% (289 of 732 cases) and 50% (289 of 580 cases). Morphological aspects of the tumor influenced the preservation rates. In cases of cystic tumor formation, the anatomic preservation rate of the facial nerve was reduced from 93 to 88% and of the cochlear nerve from 68 to 55%.
Cystic tumor formation was important to notice, because it predisposes to complications with the facial nerve and to postoperative hemorrhage, as is illustrated by the case of a 69-year-old woman (Fig. 2, Example 2). The woman presented with a cystic medium-sized VS. Surgically, the tumor was difficult to handle, was adherent, was hemorrhagic, and lacked the usual capsule. The facial nerve could not be preserved, and a facial nerve transplantation was performed immediately using a 2-cm sural graft at the CPA. Three days after surgery, the patient deteriorated clinically because of a delayed CPA bleeding and associated cerebellar ischemia and edema, which needed revision. Because of the bleeding into the nervous tissue, especially the facial nerve transplant, hypoglossal-facial reanimation was performed 3 weeks later, with good reinnervation 9 months later. The patient also recovered very well from her cerebellar ataxia, and only some imbalance was revealed by Romberg and Unterberger testing.
Tumor extension was another decisive factor. Preservation rates were as follows for the facial and cochlear nerves (respectively): Class T1, 100 and 94%; Class T2, 96 and 85%; Class T3a, 98 and 86%; Class T3b 94 and 59%; Class T4a, 86 and 49%; and Class T4b, 84 and 41%. The trigeminal nerve was disturbed in 15.5.% of the patients; the second branch was most frequently affected.
The vestibular nerve was completely resected in the majority of cases. As far as intraoperative observation (especially in large distorting lesions) can be reliable, the superior part of the vestibular nerve was solely preserved in 13 cases, the inferior part of the vestibular nerve was solely preserved in 37 cases, and superior and inferior parts were partially preserved in an additional 40 cases. In summary, there were 90 patients with some vestibular nerve preservation. Of those with partial vestibular nerve preservation, there were 64% with hearing preservation, compared with 31% in the group with complete vestibular nerve section. In the group with partial vestibular nerve preservation, there were 20% more patients with normal or good performance of Romberg and Unterberger tests; at the same time, there was a slightly higher incidence of subjective dizziness of 56%, compared to 49% in those with complete vestibular nerve resection.
The cerebellum was hemorrhagic or ischemic in seven patients, with severe permanent ataxia in two of the patients and good recovery with the ability to walk freely in the other five. Compromised cerebellar dysfunction was rare. Also, in large tumors, long-term MRI studies provided proof of cerebellum integrity (Fig. 2, Example 2).
The caudal cranial nerves showed some decrease of function in 5.5% of the cases. In cases of intraoperative suspicion, the general preventive measures were to leave the patient intubated using a nasal tube for the 1st day, test the caudal cranial nerves before extubation, use a gastric tube and feeding tube until recovery of nerve function commenced, and have regular examinations by an otorhinolaryngologist. Normal feeding was resumed by first drinking water and then eating soft meals. Aspiration apparently occurred in some patients before the onset of normal feeding by regurgitation of genuine gastric fluid. Administration of sucralfate and aluminium sacharose sulfate was not always effective.
Hemiparesis occurred in 10 patients, 2 of whom died (see below). Tetraparesis occurred long-term in one patient and temporarily in another. In one patient, after resection of a small intra- and extrameatal neuroma, there was an immediate transsection syndrome. There had not been any intraoperative signs of air embolism. A significant cervical disc and cervical hematoma were excluded by immediate postoperative myelography and CT. The recovery was poor, and the patient required the permanent use of a wheelchair. Another patient suffered a pontine hemorrhage after surgery, presumably because of an additional venous anomaly with intrapontine stasis before tumor removal. He made a slow recovery after revision with some residual tetraparesis.
CSF fistulas had an incidence of 9.2%; there were 7.6% paradoxical external fistulas occurring on Days 1 through 16 after surgery, and there were 1.6% internal fistulas occurring on Days 1 to 56. The paradoxical fistulas with some leakage at the dural closure were successfully treated by compressive dressing; in half of the cases, an additional lumbar drain was applied for 5 to 8 days. In case of an internal fistula, surgical revision was performed with exploration for possible open mastoid cells at the porus, which were sealed with muscle and fibrin glue.
Hydrocephalus needed treatment in 23 patients (2.3%) (9 patients before and 14 patients after surgery). In 12 patients, a temporary external drainage was effective. In two additional patients, this was applied first and thereafter changed to a shunt. In nine patients, a shunt was implanted immediately.
Meningitis had an incidence of 3% (1.7% for aseptic and 1.3% for bacterial meningitis). Aseptic meningitis started at Days 2 to 21 postsurgically and lasted usually for 10 days with good recovery. In three patients, there was some temporary facial paresis with complete recovery within 2 to 4 weeks. Bacterial meningitis occurred on Days 3 through 28 and lasted for up to 26 days. Whereas mostly bacterial meningitis was diagnosed clinically and on lumbar puncture, in two cases, temporary abscess formation in the CPA with typical ring enhancement was found.
Wound revision was indicated in eight patients (six because of subcutaneous wound abscess and two because of wax allergy). Since the incidence of those latter two cases, the use of bone wax at the mastoid has been completely abandoned and no further wound revisions have been necessary.
There were 22 patients (2.2%) with symptomatic postoperative hemorrhage; surgical revision was necessary in 15 (1.5%) of the 22 patients. In seven patients, the hemorrhages were of acute onset within the first 24 hours, between 4 to 9 hours after surgery, and were located in the CPA in four cases, intrapontine in two, and epidurally in one. All of these patients underwent surgery. Both patients with pontine bleeding recovered, but a mild hemiparesis and ataxia remained in one and a light tetraparesis in the other (mentioned above); in this latter patient, some preexisting vascular abnormality was assumed because of a circular hypodense region revealed by CT before tumor surgery. Of those patients with CPA hemorrhage, two recovered without problems but two died (Patients 2 and 3) (Table 3). One patient died after 1 week and one after 9 weeks because of pulmonary embolism. The patient with an epidural hematoma recovered quickly, gaining normal neurological status except for some hypacusis.
There were 15 subacute hemorrhages, 13 in the CPA and 2 in epidural locations, occurring 1 to 10 days after surgery and requiring revision in eight patients (Fig. 2, Example 2). One patient died. After surgical revision, she had developed transverse sinus thrombosis and contralateral cerebri media infarction. She had preexisting ovarian carcinoma (Patient 5). The other patients recovered with no deficits or with moderate deficits as follows: facial paresis, six patients; caudal cranial nerve palsy, one patient; and ataxia, two patients.
There were 11 deaths (1.1%) (Table 3). One patient (Patient 1) had a previous myocardial infarction and suffered a subsequent fatal infarct a few weeks after surgery. Three patients (Patients 2, 3, and 4) died of acute pulmonary embolism; one of these patients had undergone acute surgical revision twice for recurrent CPA hemorrhage. One patient (Patient 5) developed sinus thrombosis and contralateral media infarction with a large edema, secondary to revision surgery for a postoperative hemorrhage; she also had a known metastasizing ovarian carcinoma.
Five patients (Patients 6, 7, 8, 9, and 10) had severe caudal cranial nerve palsies with aspiration pneumonia. Two patients (Patients 6 and 7) were severely disabled after previous surgery that was performed elsewhere and were in a very critical state before surgery. Two patients (Patients 8 and 9) suffered a rapid progressive type of neurofibromatosis with bilateral CPA neuromas, which showed fast growth after surgery and aggravation of previous caudal cranial palsies. These patients could never be taken off the respirator. One of them, Patient 8 (Fig. 1, Example 1), a 20-year-old male patient, had been diagnosed as having NF-2 some years prior but had refused surgery. When he finally sought surgical help, he was unable to talk because of severe dysarthria and unable to eat normally because of caudal cranial nerve palsies. There were right-sided hemiparesis, right blindness, bilateral oculomotor palsy, left abducens, right accessory, bilateral glosspharyngeal, left recurrent and left hypoglossal nerve palsies, and severe ataxia and imbalance. Contrast-enhanced CT (Fig. 1, A and B) demonstrated, among other intracranial processes, optic nerve tumors, a right trigeminal nerve tumor, bilateral CPA tumors composed of neuromas and a meningioma on the left with severe brain stem engorgement, caudal tumor extension at the craniocervical junction and along the jugular neurovascular sheath (Fig. 1, C and D), and complete tumorous replacement of the cranial base. After undergoing a left suboccipital craniectomy, subtotal removal of the left CPA tumor and facial nerve transplantation from the CPA to the mastoid part (postoperative CT), the patient recovered quickly but, after a few days, needed reintubation because of respiratory insufficiency. The patient's condition deteriorated continuously during several weeks. He was sent home with respirator assistance, but he died after 3 months.
One patient, Patient 10, a 47-year-old woman (Fig. 3, Example 3) suffered a bilateral prepontine and intrapontine epidermoid tumor (Fig. 3, A-D) and an intra- and extrameatal acoustic nerve tumor on the left. Because of increasing problems with balance and with swallowing and double vision caused by left abducens palsy, the patient underwent VS removal and partial epidermoid resection. Immediately postoperatively, there was no gag reflex and there was left recurrent and bilateral hypoglossal palsy. Despite gastric tube feeding, the patient developed aspiration pneumonia, probably by regurgitation, and did not respond to intensive antibiotic treatment but went into septicemia and died after 6 weeks. One patient (Patient 11) did not recover after extirpation of a postoperative CPA hemorrhage.
We discuss the relevance regarding CSF fistula, infection, facial and cochlear nerve function, and general complaints. The prevalence of certain problems and complications is generally correlated with the surgical approach, i.e., the TL, MF, and suboccipital/retrosigmoid (SO) approaches. Although the attitude of strictly insisting on one special approach has been abandoned, special advantages and disadvantages, such as the incidence of CSF fistula, infection, and facial and cochlear nerve function, are still ascribed to each approach.
By the extended TL-transotic approach, as reported by Chen and Fisch (6) in 147 patients, very good results with regard to mortality (1 of 147 patients) and morbidity and anatomic facial nerve preservation (95%) are achieved. Brackmann and Green (2) advocate the TL approach in patients who are deaf or who have poor chances of hearing preservation. They state that this approach has the lowest morbidity with regard to spinal fluid leaks and postoperative headaches and provides the advantage of "minimum cerebellar retraction, identification of the facial nerve proximally and medially, and the ability to repair immediately the facial nerve, if it is severed during acoustic tumor removal." Because the value of hearing preservation is now appreciated, a categoric decision for the TL approach has been given up. Briggs et al. (3) found a higher incidence of CSF fistulas (11.6%) and infection in large tumors (4.6%) with hydrocephalus. Hoffman (20) found no significant difference in CSF fistula incidence between the TL group (21%) and the retrosigmoid transmeatal group (16%).
Pellet et al. (38) reported on 228 patients, 85% of whom were operated on using the TL approach and 15% using the MF approach. The mortality rate was 1.75%. Anatomic facial nerve preservation was achieved in 94% of the patients. Hypoglossal-facial anastomosis was necessary in 10% and radicality was achieved in 99%. CSF leaks occurred in 7.5%. Five percent were cured by lumbar puncture, and 2.5% were cured by surgical revision. Meningitis occurred in 0.4%, postoperative hematomas occurred in 1.75%, and balance troubles occurred in 67%. Rodgers and Luxford (42) reported on 723 cases using the TL approach, with an incidence of CSF leak in 6.8% and of meningitis in 2.9%. The presence of a CSF leak did not predispose to meningitis. Technical factors seemed to account for postoperative CSF leak and meningitis; meningitis occurred more often in larger tumors. Ramsay and Luxford (40), reporting on the TL approach in 65 patients older than 70 years, reported radicality in 94%, CSF leak in 10%, and meningitis in 1.5% of the patients. They advocate complete removal with facial nerve preservation as the goal of choice.
Haid and Wigand (18) reported on 263 patients who were operated on using the MF approach for tumors that had up to 35 mm extrameatal extension, with radicality in 96%, excellent facial function in 78%, persistent paralysis in 6%, and good rates of hearing preservation. Mazzoni et al. (31) reported on a selected series of 90 of 300 patients operated on using the SO approach for a trial of hearing preservation. The facial nerve was preserved in 99%, with completely normal function in 78%. The cochlear nerve was anatomically preserved in 96% and functionally in 44%. CSF leak occurred in 6.6% of the cases, meningitis in 2.2%, paresis or paralysis of the IXth and Xth cranial nerves in 3%, and ataxia in 2%. Glasscock et al. (16) reported on the SO and MF approaches in 161 selected patients. They found a lower incidence of temporary facial nerve paresis using the SO approach but also a higher incidence of postoperative headaches using this approach.
The literature provides proof of the statement that results are best achieved by the technique with which the surgical team is most familiar. Therefore, regarding the general morbidity and mortality that is associated with VSs, the results are best (and are very similar from one study to another) if they are reported from institutions that have vast experience with this pathological abnormality and with special approaches. Lack of experience sets the results back to the beginnings of neuroma management, as is clearly demonstrated by Charabi et al. (5), who collected data of 59 patients who underwent surgery using the SO approach in various units in Denmark. Because almost all patients in Denmark are operated on using the TL approach, the results in the SO approach are dramatically poor, with a high mortality rate higher than 8% and a high morbidity rate. This does not fulfill the current minimal standard and is unacceptable. A review of the literature provides proof of the actual standard in SO and MF approaches. A surgeon who intends to achieve hearing preservation in increasing numbers of his patients (according to the standards set by the patients) must fulfill these surgical standards of minimal mortality and morbidity. A limitation for comparison exists in that most authors report on "selected series" that automatically exclude important aspects worth considering and comparing regarding technique and management.
TABLE 3. Mortality Related to Preoperative Conditions and Postoperative Complicationsa
|Time from Surgery until Death|
|1||Myocardial infarct||Myocardial infarct||2 d|
|2||Pulmonary embolism||4 d|
|3||Pulmonary embolism||4 wk|
|4||Acute hemorrhage until surgery||Pulmonary embolism||10 wk|
|5||Ovarian cancer, metastasis||Subacute hemorrhage||Sinus thrombosis, contralat. media infarct||3 wk|
|6||Disabled||Caudal cranial nerve palsies||Aspiration pneumonia||11 wk|
|7||Disabled||Caudal cranial nerve palsies||Aspiration pneumonia||8 mo|
|8||NF-2, disabled||Caudal cranial nerve palsies||Aspiration pneumonia||Tumor regrowth, brain-stem dysfunction||7 wk|
|9||NF-2, disabled||Caudal cranial nerve palsies||Aspiration pneumonia||2 wk|
|10||Bilateral epidermoids||Caudal cranial nerve palsies||Aspiration pneumonia, septicemia||5 wk|
|11||Acute CPA hemorrhage until surgery||1 wk|
NF-2, neurofibromatosis-2; CPA, cerebellopontine angle.
The "feasability" (50) of total tumor removal and hearing preservation is under suspicion by some colleagues because of histological findings of tumorous nerve infiltrations (13, 15, 36, 37). Some authors (28) completely reject the principle of hearing preservation surgery in "favour of total tumor removal" because, according to their immunohistochemical studies (monoclonal mouse antibodies to human neurofilaments), "the cochlear nerve-tumor interface showed no clear cleavage in those cases in which macroscopically visible adherences between the cochlear nerve and the tumor were present." However, the minor incidence of recurrences (6 within 880 patients who did not have NF-2) detected in our series illustrates that by total microsurgical removal as it is performed in our patients, relevant tumor rests evidently were not left behind for the sake of hearing preservation. Moreover, in the majority of cases, the so-called acoustic neuroma is a VS originating from the vestibular nerve and only compressing the cochlear nerve (14).
Deliberate subtotal tumor removal should be performed only in rare cases of lifesaving decompressive surgery (10 cases = 1% in our study) or in case of a special agreement with the patient in whom a tumorous cochlear nerve is anticipated (11 cases = 1.1% in our study). Several successful cases have been reported of partial tumor removal (26, 33). The details and difficulties are presented and discussed elsewhere (46).
The attitude of removing the tumor "immunohistologically" complete would imply the occasional resection of a macroscopically intact cochlear nerve. That nerve, according to our experience and the experience of others (9), never shows any tumor recurrences and would function well for decades, as is evident in increasing numbers.
Regarding CSF leak, we stress that only 1.6% of the patients in this study needed surgical revision because of a "real" internal fistula. In each of these cases, the cause of the fistula was some unsealed air cells at the posterior wall of the IAC. At surgical revision of this area, the previously inserted piece of muscle was found to be too small or was not fixed well enough. This can happen in cases of strong CSF pulsation or if one restricts the distribution of the fibrin glue to avoid its contact with the cochlear nerve. In 7.6% of the patients, there was a so-called paradoxical rhinorrhea. In these patients, the CSF was leaking through a dural opening to the epidural space and from there to the mastoid cells. Here, an external treatment is effective by compressive dressing and sometimes by additional lumbar drainage. To reduce the incidence of such CSF leak, we also fix a flat layer of muscle over the suture line and over the mastoid cells.
The incidence of bacterial meningitis was also very low (1.3%). Aseptic meningitis has not been encountered since the use of bone wax for the closure of opened mastoid cells has been discontinued.
The desired mortality rate is below 1%; the current mortality rate is approximately 1 to 2% (12) and, rarely, 1 to 3%. Careful selection of patients and decision against surgery for patients who are at high risk, such as patients with preexisting nerves deficits, hemiparesis, or cardiopulmonary compromise, would, as in all surgical disciplines, improve the numbers. Of the 11 patients who died in this series, 4 would most likely have died soon without surgery (Table 3). They were accepted for surgery because of the patients' and relatives' insistence on surgical treatment, because of the lack of any alternative treatment in view of the huge tumor sizes, and because of the hope for improved life quality (there were two malignant cases of NF-2 in one patient with a carcinoma and one with previous severe myocardial infarction). There were two additional patients who had been severely disabled since partial tumor removal that had been performed elsewhere.
According to our experience with this series, there are three groups of patients who are at major risk. First, there are premorbid patients with previous surgery or with severe brain stem compression, either because of delayed indication or especially in some patients with NF-2. Patients in this group have become less frequent because of early diagnosis, but they will usually be sent away from most units and to departments with high prevalence of CPA surgery. The only solution to this would be the option for surgery or radiosurgery if such patients are sent to specialized centers early to decide for or against surgery. The patient with neurofibromatosis presented in this article (Fig. 1, Example 1) is a typical example of a patient being admitted to our institution late, with virtually no chance of rescue. In conclusion, it is not the "chronological age" (39) but the preoperative general biological state that advantageously affects outcome, and advantageous results can be obtained in patients who are advanced in age.
Second, there are tumors with caudal cranial nerve involvement and complete palsy after surgery. Patients with such tumors are at high risk of aspiration pneumonia. There is no absolute prophylaxis possible; even tracheostomy and tube feeding cannot prevent it completely. This risk was considerably higher when the caudal cranial nerves were intact before surgery and complete paralysis started after the operation, whereas chronic preoperative palsy development was far better tolerated. All of these unfavorable circumstances are presented in Example 3, in which an acoustic tumor was associated with pre- and intrapontine epidermoid. The patient's complications were not caused by the acoustic tumor surgery but the epidermoid resection.
Third, there is a certain danger of postoperative hemorrhage that is higher in cystic tumor formation but it could occur in any VS. This factor can be controlled best and the incidence can be largely reduced. Careful inspection at the completion of surgery for venous bleeding that may not be apparent is essential during jugular venous compression, and subtle anesthetic guidance after surgery will prevent blood pressure crises in hypertensive patients. Our postoperative management has therefore developed toward subtle but fast anesthesia termination and extubation as early as possible, usually immediately in the operating room or within 1 to 2 hours in the intensive care unit. The patient's wakefulness can thereby be controlled as the most reliable parameter for any deterioration and related developing hemorrhage. Some tumors that lack the normal tumor capsule predispose especially to postoperative hemorrhage, even after several days, as presented in Example 1. In such cases, special care for these patients in the early postoperative period is indispensable. The reasons for this specific danger in cystic tumors are subject to speculations (4). Tumor enzyme activity might be destructive to the tumor-nerve barrier and might be a cause for the lack of a normal tumor capsule and the severe softening of the brain tissue that is much more prone to hemorrhages during and after surgery. Moreover, the rates of nerve preservation in cystic tumors are poorer in our patients and those of other authors (4).
Regarding safety of the brain stem, the technique of dissection under continuous irrigation is especially helpful because cauterization may be reduced to a minimum. Identification and control of vascular supply to the brain stem are most reliably provided by the SO approach (41), whereas overview as well as access are limited in the MF approach and the TL/transotic approach (the latter bearing a substantial risk in large tumors with tight brain stem connections) (55).
The argument that the MF or posterior fossa approaches, the only routes enabling hearing preservation (with few exceptions) (32), are more dangerous to the patient or especially to the facial nerve than the TL route was discarded some time ago (8, 17, 27, 35). Regarding the facial nerve, the TL and the SO approaches offer the best opportunities for preservation as well as for reconstruction of the facial nerve (2, 7-10, 12, 21-23). Tos et al. (54) found suction to be a damaging process during nerve preparation. They emphasize that the most dangerous part for facial nerve severance is just medial to the porus. Glasscock et al. (16) report on the SO and MF approaches in 161 selected patients. They found a lower incidence of temporary facial nerve paresis in the SO approach but a higher incidence of postoperative headaches. In summary, preservation rates by the suboccipital approach are reported to be higher than 90% independent of tumor size, but, in general, these reports include tumors larger than 30 mm in diameter (1).
Cohen (7) was one of the first and few (34) who reported on the indications for using the MF and SO approaches for hearing preservation when the majority of his colleagues still insisted on the TL approach. Even his early results were very favorable, with only 8% facial nerve paralysis and no mortality. Many authors now report functional cochlear nerve preservation rates from 25% (12) to higher than 50% (52). Additional experience with both approaches enabled Cohen (8) to formulate the advantages of the SO approach as providing visualization, access, and control of all relevant structures and the chance of functional nerve preservation.
The incidence of postoperative headaches is apparently higher with the SO approach than with the TL or MF approach (2, 54). Harner et al. (19) investigated 331 patients for postoperative headaches after the osteoplastic SO approach and found a decrease from 23% 3 months after surgery to 9% after 2 years. In our series, immediately postoperatively, 9% complained of headaches compared to 12% before surgery. Currently, a study is being performed at our department on the benefits and problems in osteoplastic procedures, either by osteoplastic craniotomy or by craniectomy with bony replacement plasty. By these measures, long-term development of subcutaneous submuscular scarring and tearing at the dura will be prevented. Whether this mechanism is responsible for persisting headaches after craniectomies is a matter of speculation.
In the majority of cases, an experienced surgical team can safely and successfully resect VSs with a minimal risk of mortality or recurrence. At the time of discharge from the hospital, 73% of the patients in this series had satisfactory facial nerve function with complete eye closure; 59% showed good function (HB Grades 1 and 2), and 14% with HB Grade 3 were expected to recover within weeks to months.
In cases of nerve discontinuity or loss of the facial nerve stump at the brain stem, it is recommended that reconstruction be performed immediately by transplantation or by nerve reanimation with a donor nerve, respectively. Cochlear nerve deafness can be prevented in increasing numbers by increasing the rates of anatomic (68%) and functional (39%) nerve preservation. The suboccipital approach is the only one that enables hearing preservation regardless of tumor size. Good preoperative hearing and small tumor sizes are favorable factors.
The mortality rate was 1.1% in this study. It may be lowered if patients with severe preoperative morbidity do not undergo surgery or if they are not transferred from other centers at a last desperate moment but at a time when surgery still offers a realistic lifesaving chance. Cystic schwannomas require special surgical and postsurgical attention, because they are more dangerous to remove with regard to brain stem and facial nerve integrity and are more likely to lead to hemorrhage in the acute and subacute postoperative periods.
We cordially thank S. Brinkmann for producing the photographs.
Received, November 9, 1995.
Accepted, August 6, 1996.
Reprint requests: Madjid Samii, M.D., Department of Neurosurgery, Nordstadt Hospital, Haltenhoffstr. 41, D-30167, Hannover, Germany.
This article is important for the health care planner as well as the neurosurgeon. In some ways, it is the core of this remarkable series presented by Samii and his colleagues on vestibular schwannoma (VS) surgery. There is a 1.1% mortality rate and a 22.3% morbidity rate for surgery in the hands of this master surgeon. Although some surgical complications are not very significant (e.g., wound revision), others can be life threatening (meningitis) or troublesome for the patient (IXth, Xth, and XIth cranial nerve weakness). It is hoped that surgical complications would not include facial weakness or hearing loss.
Samii and Matthies present an honest and detailed report, which I think
is one of the best works regarding VS surgery. Even with a master
surgeon, the management of these tumors has significant morbidity.
Surgery should be performed only in centers that have experienced
surgeons, residents, and the attending personnel able to detect and
handle complications and sophisticated operations. This surgery is not
appropriate for community hospitals and occasional neurosurgeons.
The authors present their impressive results with 1000 tumors treated surgically using the suboccipital approach. They note that the suboccipital approach is the only route permitting hearing preservation when removing both large and small tumors. We think that the suboccipital approach also lends itself to higher rates of facial nerve preservation than other approaches that place the facial nerve more directly in the operative field. With the suboccipital approach, there is frequently a layer of eighth nerve between the tumor and the facial nerve. This makes dissection along the facial nerve safer than when the facial nerve is more directly exposed in the operative field throughout the procedure, as in the middle fossa or translabyrinthine approaches, for which the residual bundle of eighth nerve is often not positioned to act as a protective layer for the facial nerve. The facial nerve was preserved in 93% of the 1000 operations reported here. At discharge from the hospital, 73% of the patients had House-Brackmann Grade 1, 2, or 3 facial nerve function. The cochlear nerve was preserved anatomically in 68% and functionally 39% of the cases. We have also relied on the suboccipital approach and have achieved results similar to those reported by the authors. In our most recent 100 patients with acoustic neuromas, it was possible to preserve the facial nerve in continuity in 97% the of patients, with the majority of patients having House-Brackmann Grade 1 or 2 facial nerve function at discharge from the hospital. Only three grafting procedures were needed.
There has been a trend in the United States away from the use of the
sitting position for acoustic neuroma removal, as reported in this
series. In recent years, we have used the three-quarter prone position
routinely. We gain many of the advantages of the sitting position by
tilting the table with the patient in the three-quarter prone position
so that the head is significantly above the heart. The head is high
enough that the surgeon often must place his stool on a low platform at
the head of the table. Initially, we surgically treated only small
tumors in the three-quarter prone position but, in recent years, have
used it for tumors of all sizes and have been able to obtain
satisfactory relaxation of the brain with elevation of the head, modern
neuroanesthetic techniques, and drainage of cerebrospinal fluid (CSF)
from the cisterns. Our method of closure of the meatus is somewhat
different from that reported here. We carefully wax the air cells and
lay a fat graft over the drilled meatal lip. We have operated on more
than 200 consecutive patients without encountering a CSF fistula and
have not used fibrin glue or spinal drainage in any patient. The goal
has been total removal, and in 100 consecutive patients who underwent
magnetic resonance 1 year postoperatively, only 2 were found to have
small residual fragments of tumor. Both fragments were located in the
cerebellopontine angle rather than within the lateral end of the
meatus, which we were able to clear of tumor in every case by the
suboccipital route. Our incidence of caudal cranial nerve deficits has
decreased below those reported here with the shift from the sitting to
the three-quarter prone position for surgery. This may be because the
lower cranial nerves are more directly in the operative field with the
sitting position than with the three-quarter prone position. The
authors have achieved excellent results with the suboccipital
This excellent article by Samii and Matthies presents their series of 1000 consecutive acoustic neuromas that were carefully followed and for which the complications were documented. It will stand as a benchmark for all surgeons performing surgery on acoustic tumors.
Although a fair number of complications do occur, they are manageable, with very low mortality. Samii and Matthies consider a 1.1% mortality rate for 1000 acoustic tumors to be acceptable. Of some interest is that our series of 500 cases has practically identical complication rates. Our mortality is 1.0%. I think this article underscores that meticulous surgery by an experienced operator can reduce the risk of major intracranial surgery to that of acute appendicitis.
Samii has long been recognized as a pioneer and leader in the surgical management of vestibular nerve schwannomas. This review by Samii and Matthies of 1000 such tumors treated during a 15-year period should serve as a benchmark for all of us involved in the treatment of VSs. The results described in this article should not be expected by those who are early in their learning of such procedures but should represent a goal. There is no question that the surgical outcomes of the last several hundred patients on whom I have operated have significantly improved from series of similar size that were reported earlier in our experience at the Mayo Clinic. This improvement in outcome took place even though the surgical time continued to decrease and the tumor size has stayed the same; some of the smaller lesions are being treated with gamma knife or by surgeons who initially see the patient.
I agree with many of the points made in this review, and I have little to add as far as constructive criticism. I, too, think that cystic tumors are much more difficult to dissect free from the facial nerve. At this point, however, I have not seen cases of delayed hemorrhage as described by Samii and Matthies in this review. I also agree that cerebellar dysfunction is extremely uncommon, even after a large tumor resection. Nevertheless, there is the very rare patient who does develop a hemorrhagic infarction or ischemia of the cerebellum that can result in devastating complications. Our incidence of such an unfortunate outcome is also 1 in 500 cases. Fortunately, we have not seen problems with hemi- or tetraparesis after surgery. The incidence of CSF fistula has dramatically decreased as our experience has increased. At the Mayo Clinic, we also had the incidence rate of approximately 10% in the early 1980s; however, during the years, this has dropped to approximately 3%. Essentially, all of our CSF leaks continue to manifest themselves with rhinorrhea, which we think is secondary to opened mastoid air cells, usually in the porus. Unlike Samii and Matthies, we usually treat this with early surgery of the mastoid air cells, which are obliterated by a fat graft. This mastoid procedure eliminates the need for reentering the posterior fossa and is almost always successful with the first attempt. Spinal drainage is not performed either before or after mastoid obliteration. We also have not found problems with "wax allergy" and continue to use large amounts of bone wax to obliterate not only the mastoid air cells at the porus but also at the craniectomy site. I have not seen problems associated with leaving large amounts of wax at these sites, but the concerns presented by Samii and Matthies warrant consideration.
I also agree that the best results are achieved by a technique with which the surgical team is most familiar. The experience of the team using a particular approach is probably more important than the surgical approach used.
This is a very good review of how far we have come in our ability to
surgically manage VSs. Although good preoperative hearing
and small tumor size increases the likelihood of hearing preservation
with surgery, we also attempt to preserve hearing in all patients. The
suboccipital approach allows such an attempt, without necessarily
increasing the surgical risks of facial nerve injury or other
neurological morbidity. Discussions with Samii and many other surgeons
who perform many procedures for VSs have convinced me that regardless
of the particular surgical approach, an individual needs to perform a
large number of procedures to minimize morbidity and maximize retained
Please submit your thoughts or comments to William Chandler, Internet Moderator at firstname.lastname@example.org.
Return to Neurosurgery Home Page
Return to Table of Contents