The Journal of Laryngology and Otology
March 1991, Vol. 105, pp. 165-173

Acoustic neuroma (schwannoma) surgery 1978-1990
J. E. M. DUTTON F.R.C.S., R. T. RAMSDEN F.R.C.S., R. H. LYE F.R.C.S., K. MORRIS F.R.C.S.

A. O. KEITH F.R.C.S., R. PAGE F.R.C.S., J VAFADIS F.R.C.S. (Manchester)

Abstract

A series of 151 patients with 154 acoustic schwannomas have been operated upon in Manchester Royal
Infirmary by a joint Otological and Neurosurgical team, employing either the translabyrinthine or the suboccipital approach. The perioperative mortality rate was 3 per cent. Anatomical preservation of the facial
nerve was achieved in 89 per cent of tumour removals and a good to normal functional result in 79 per cent
of cases. Attempts at hearing preservation have been unsuccessful, largely because of the small number of
patients in the series in whom useful hearing was present preoperatively. Complications included major
brain stem ischaemia (1.2 per cent), CSF fistula (5 per cent) and facial dysaesthesia (7 per cent). The incidence of mortality and morbidity is directly related to tumour size and to the experience of the surgeons. A
number of patients experienced an unusual type of post-operative dreamlike state which appeared to be a
form of hypnagogic hallucination, and the possible neurophysiological mechanism responsible for this
phenomenon is discussed. The continuing failure to attain the ideal of early diagnosis is lamented, and the
importance of a flexible bidisciplinary surgical approach emphasized.

Introduction

Material (Table 1)

The twentieth century has seen an evolving pattern in
the management of acoustic neuromas. In the first two
or three decades the picture was one of almost unmitigated gloom, with large tumours presenting late, and
removed with such dreadful results that even to consider
operating on a fit person with a small tumour was
regarded as inappropriate. Thefirstgreat advance came
with Cushing's intracapsular operation in the thirties but
total tumour removal with minimal mortality and morbidity only became a realistic goal much later with the
advent of microsurgical techniques, improvements in
neuroanaesthetic practice and advances in our understanding of the anatomy, physiology and pathophysiology of the brain stem and surrounding structures. The
next quest was for preservation of the facial nerve, previously regarded as expendable, and in the last ten years,
maintainence of useful hearing in certain cases has
become possible. Sadly, however, early diagnosis, which
has always appeared to be imminent with each new
audiological or radiological advance, remains elusive.
One of the greatest advances in the management of
these tumours has been the co-operation between the
neurosurgeon and the otologist, leading to the emergence of the neuro-otologist or otoneurologist depending on one's point of view, and to the team approach to
tumour surgery. During the last twelve years (March
1978-March 1990), the Departments of Otolaryngology
and Neurosurgery at Manchester Royal Infirmary have
collaborated in the management of a series of patients
with acoustic neuromas. This article describes the results
of these efforts.

One hundred and fifty one patients (89 females: 62
males. F: M ratio = 1.5:1) with 154 operated tumours
have undergone a total of 161 operations for histologically proven acoustic schwannomas Age at presentation
ranged from 15 to 79 years; the distribution is shown in
Figures 1 and 2. Total removal was achieved in 142
tumours, including six for which staged surgery was performed. Staging was not a premeditated philosophy in
any instance but was decided upon during primary surgery if changes in vital signs led to concern on the part of
the anaesthetist. This situation usually arose when dealing with the residual part of a large tumour firmly adherent to the brain stem. If the primary procedure had to be
abandoned prematurely, at least three months were
allowed to elapse before a second operation was performed. During this time, the distorted brain stem was
found to resume its normal position and the tumour remnant to become less adherent to it, thus facilitating
eventual removal (Figs. 3 a,b,c).
"Unplanned" subtotal removal was performed in six
patients because of anaesthetic problems. Of these one
died perioperatively (see below) and one two years later
of unrelated disease. The remainder are under regular
TABLE I
patient data
Acoustic neuromas March 1978-March 1990
Total operated patients
(Males = 62: Females = 89)
Total operated tumours
Operations (TL = 89: SO = 72)
No surgery patients

Accepted for publication: 16 November 1990.
165

151
154
161
9

166

J. E. M. DUTTON, R. T. RAMSDEN, R. H. LYE, K. MORRIS, A. O. KEITH, R. PAGE, J. VAFIDIS

10-19

20-29

30-39

4M9

50-59

6049

70-79

Age in Decades: Males
FIG. 1
Decade distribution of male patients.

review. Planned subtotal removal was carried out in a
further six patients with large tumours, five because they
were old or infirm, and one because it was hoped to preserve hearing in an only hearing ear. These patients continue to be kept under review with annual CT
surveillance. Further surgery has not yet been necessary
in any of these six patients although one has died of unrelated disease. The "total tumour removal" number of
142 includes 21 cases in whom it was necessary to leave
on the brain stem or facial nerve a tiny portion of capsule
manipulation of which caused major vital sign changes.
These tumour rests were coagulated with the bipolar
diathermy. These patients have been followed up with
regular CT scans for periods up to 12 years without evidence of regrowth; however the significance of these
cases, in the light of recent MR : Gd DTPA studies
(Pace-Balzan et al. in preparation), will be discussed
later in this paper. Of the 151 patients only one case (0.6
per cent) of true tumour recurrence has been observed,
and that was not in one of the patients in whom such a
tiny fragment was known to have been left behind. It is
important to define what one means by the term "recurrence", and the authors regard it as the reappearance of
tumour when in the opinion of the surgeons total re30-/

moval had been achieved. Regrowth of tumour when a
subtotal removal was knowingly performed should not
be regarded as recurrence, but as "regrowth".
During this period we have seen an additional nine
patients from whom surgery has been withheld. Six of
these are elderly with very small lesions, and three are
patients with tumours in their only hearing ears, e.g. the
second side of a bilateral case. These patients have not
been included in the analysis.
Eighty-nine translabyrinthine and 72 suboccipital
approaches have been performed. In the early days of
our collaboration there was a preponderence of suboccipital operations (26 out of the first 33) but as the neurosurgeons gained familiarity with the translabyrinthine
approach the situation reversed. The translabyrinthine
approach is at present employed for most tumours up to
4 cm in diameter, and the suboccipital approach larger
lesions or if hearing preservation is to be attempted. The
middle fossa approach has not been used.
Preoperative tumour size (Table II and Fig. 4): The
154 tumours were classified as intracanalicular (N = 9 : 5
per cent), up to and including 2.5 cm extracanalicular
diameter (N = 72:46 per cent),
2.5cm^cm
(N = 48:31 per cent), and greater than 4 cm
(N = 23 : 15 per cent). In addition there was one intravestibular schwannoma (Birzgalis and Ramsden, 1991)
and one acoustic schwannoma with intrapetrous but no
intracranial extension.
Results
Perioperative mortality
There have been five perioperative deaths in this
series (3 per cent). Scrutiny of these individually may
help to determine factors which might have been avoidable and so influence future practice.
Patient 1 This 76-year-old man with a history of cardiovascular disease presented with a 5 cm tumour which
was removed totally with preservation of the facial nerve
via a suboccipital approach, the operation taking
approximately 8 hours, at the end of which he died from
a supratentorial subarachnoid haemorrhage without
regaining consciousness. This case occurred early in the
series. The management of this patient would now be a
subtotal removal via a translabyrinthine approach.
Patient 2 This was a 54-year-old lady with hitherto well
controlled bronchial asthma who died in the respiratory
care unit in status asthmaticus ten days after the
uneventful removal of a 3 cm tumour.

20-

10-19

20-29

30-39

4M9

50-59

60-69

Age in Decades: Females
FIG. 2

Decade distribution of female patients.

70-79

Patient 3 This 65-year-old man developed bronchopneumonia and died approximately one week after the total
removal of a 3.5 cm tumour. At the time of surgery some
difficulty was experienced in separating the tumour from
the lower cranial nerves, and vagal damage with aspiration could have predisposed to his pneumonia. Early
tracheostomy might have provided protection for the
airway.
Patient 4 This 24-year-old woman presented with a
4.5 cm acoustic schwannoma associated with an ipsilateral petrous ridge meningioma, a contralateral cerebel-

167

ACOUSTIC NEUROMA (SCHWANNOMA) SURGERY 1978-1990

(c)

(a)

FIG. 3
Staged removal of large acoustic neuroma. The large tumour
shown in (a) was partially removed via the suboccipital route and
the residual fragment is seen on the brain stem (b). Three months
later (c), the brain stem has shifted and the tumour remnant has
come into contact with the petrous bone. It was subsequently
removed by a translabyrinthine approach.

TABLE II
tumour size data
Tumour size (N = 154
IC
€2.5 cm
2 . 5 ^ cm
>4cm
Intravestibular
Intrapetrous

9
72
48
23
1
1

lopontine angle tumour and raised intracranial pressure
when twenty-four weeks pregnant. Intracranial surgery
was delayed until after the thirty-fourth week when her
child was delivered by Caesarian section. Her acoustic
schwannoma and meningioma were removed via a suboccipital approach. At the end of a prolonged but
uneventful procedure she sustained a brain stem infarct
and died without regaining consciousness.
Patient 5 This 55-year-old man had an intrameatal
tumour removed via the suboccipital approach with total
preservation of his excellent pre-operative hearing.

168

J. E. M. DUTTON, R. T. RAMSDEN, R. H. LYE, K. MORRIS. A. O. KEITH. R. PAGE, J. VAFIDIS

14.94*

•

Intraanaltcular

•

'<2.5cm

•

2.5-4cm

0

>4cm

•

Intravestibular

•

Intrapetrous

31.17X

Tumour Size
FIG. 4
Tumour size at presentation.

Shortly before he was due to be discharged from hospital, he fell and struck the back of his head on the corner
of his bed, lacerated his lateral sinus and died of intracranial haemorrhage. It is unlikely that this would have
occurred if the tumour had been removed via the translabyrinthine route.
Brain stem damage
Major damage to the brain resulting from vascular
compromise was seen in two patients, one of whom suffered a fatal infarction five months after surgery to
remove a 5 cm tumour. A second patient with a 4 cm
lesion suffered an immediate post-operative lateral
medullary syndrome, from which he eventually made a
reasonable recovery being left with increased spasticity
in his limbs which prevented him from returning to
work. Two further patients suffered some residual ataxia
which prevented them from returning to their previous
level of recreational activity (e.g. one was unable to
dance).
CSFfistula
Eight patients developed a post-operative cerebrospinal fluid fistula requiring a further surgical procedure (5
per cent). Of these, six had had a translabyrinthine and
two a suboccipital removal. Four further patients who
had had a posterior fossa approach developed a pseudomeningocoele requiring CSF diversion procedures.
Meningitis occurred in four patients three of whom had
had a CSF leak, and in two instances ventriculitis developed. In all cases the infection responded promptly to
treatment with antibiotics and there was no residual
deficit.
Facial nerve function
In our assessment of facial nerve function, no patient
TABLE III
complications
Death
Brain stem syndrome
CSF fistula
Meningitis
Facial dysaesthcsia
Temporary neuropraxia
Donor site infection

5
2
8
4
12
3
2

has been included who was operated on after March
1989. This was to allow a period of one year for recovery
in function in those nerves in which degeneration had
occurred. It is our experience that little further improvement will occur in such a nerve after that interval. One
hundred and thirty three tumours are therefore included
in this part of the analysis. Many reports over the years
have quoted figures for facial nerve preservation, by
which has been implied anatomical preservation of the
nerve. In this series 118 out of 133 facial nerves were
thought to have been preserved in anatomical continuity
(89 per cent) . It is however much more meaningful to
report facial nerve results in terms of function (Moffat et
al., 1989a), and several grading systems have been proposed to facilitate the recording of this information,
though none has gained universal acceptance. The most
widely used scheme is that suggested by the House group
(House and Brackmann 1985), comprising six grades.
Prior to that in 1982, we proposed a system based on five
grades of function (Lye etal, 1982), and a broadly similar
scheme has been more recently suggested by Bentivoglio et al. (1988). Our classification is listed in Table
IV. It differs from that of House and Brackmann in that
those authors divide our Group 3 into two subgroups,
defined as Moderate Dysfunction and Moderately
Severe Dysfunction.
Figures 5 and 6 indicate the results in this series. It will
be seen that the largest single group of patients fall into
Grade 1, i.e. have normal facial function (N = 56:43 per
cent), and that the majority of patients have facial function of Grade 3 or better, i.e. acceptable to normal
(N = 104: 79 per cent). Five patients were unassessable;
two were lost to follow-up (LTFU), of whom one emigrated shortly after her operation, and the other refused
to attend for further outpatient appointments. Three
further patients died in the early post-operative period
(cases 1,2 and 4 above). Eight patients had a nerve graft.
Four had an immediate intracranial repair (2 cable graft,
1 rerouting, and 1 reimplantation into the brain stem).
Three of them achieved Grade 3 function, but the fourth
died before function could be assessed (case 2 above).
Four patients underwent a delayed hypoglossal or accessory anastomosis and none attained better than grade 4.
Delayed transient facial weakness was seen in nine
cases. In these patients an immediate good to normal
face deteriorated at between the seventh and tenth days,
often to grade 4, only to recover without treatment in
about a week. It was originally thought that this might be
due to a physical or chemical effect from the abdominal
TABLE IV
facial nerve grading
(Lye et al. 1982)
Grade 1
Grade 2

Normal facial function.
Full facial movements but delayed blink, or
crocodile tearing, or dry eye, or subjective taste
disturbance.
Full facial movements with minimal synkinesis.
Minimal facial weakness.

Grade 3
Grade 4

Moderate facial weakness.
Moderate weakness but permanent tarsorrhaphy
necessary.
Severe weakness, but muscle tone present.
Total paralysis.

3%

1.2%
5%
2.4%
7%

1.9%
1.2%

Grade 5

169

ACOUSTIC NEUROMA (SCHWANNOMA) SURGERY 1978-1990

residual sensory changes. Of the twelve patients, five
continue to experience troublesome symptoms: two
have anaesthesia dolorosa with a coexisting total facial
paralysis, and three patients with Grade 4 or Grade 5
facial nerve function have markedly reduced corneal
sensation and trophic changes in the cornea. One of
these patients also has a tendency to trophic changes in
the skin of the face and nasal vestibule. The coexistence
of a severe degree of facial paralysis and reduced corneal
sensation presents one of the most serious and difficult
complications of acoustic neuroma surgery.

60 V
SO-

N=I33

2

3

4

5

LTFU

Facial Nerve Grading (See Table4)
FIG. 5
Overall facial nerve function (FU>1 year).

fat graft used in closure of the translabyrinthine defect,
but the fact that two of the patients had been operated by
the suboccipital route makes this suggestion untenable.
One patient who achieved a Grade 3 result displayed
marked wasting of the temporalis muscle which failed to
recover, suggesting an anomalous innervation of that
muscle by the facial rather than the trigeminal nerve.
If facial nerve function is analysed in relation to
tumour size, it is clear that the smaller tumours are associated with a more favourable outcome. (Fig 7). In
patients with tumours of 2.5 cm or smaller, facial nerve
function was Grade 1 in 60 per cent, Grade 1 or 2 in 89
per cent and Grade 3 or better in 96 per cent of cases.
When the tumour was larger than 2.5 cm these figures
fell to 24 per cent, 44 per cent and 66 per cent
respectively.
Facial dysaesthesia

Twelve patients (7 per cent) had a demonstrable but
slight alteration in facial sensation following tumour
removal. Ten had had pre-operative evidence of trigeminal involvement with tumours of 2.5 cm-6 cm diameter.
Two patients with small tumours (<2.5cm) and no preoperative signs of trigeminal involvement have slight
8ra*5

Hearing preservation
Hearing preservation was attempted in five patients
with intrameatal tumours. In three there was immediate
post-operative loss of hearing. In two cases there was
immediate preservation of hearing so that the patients
were able to converse on the telephone using the operated ear. In one of these, however, the hearing was suddenly lost on about the tenth post-operative day, and the
other was one of the perioperative deaths (see case 5
above). In a sixth patient with a large second side
tumour and good hearing, a portion of the tumour capsule was left in the hope of retaining useful hearing.
Some perception of low frequency sound did remain but
speech discrimination was lost.
Problems related to operative positioning
Two patients operated upon in the parkbench position
developed temporary lateral popliteal nerve damage
with weakness of dorsiflexion of the foot, and one
patient sustained transient neuropraxia of the ulnar
nerve which also recovered fully. One patient suffered a
small area of scarring on the forehead related to the
three pin headrest. The translabyrinthine approach was
associated with a small area of pressure necrosis on the
contralateral side of the scalp in a few patients. This was
related to the use of the rather firm neurosurgical horseshoe headrest, and has not been seen since the introduction of a softer foam support.
Miscellaneous
There were two instances of delayed healing of the leg
incision during the first year of the series when muscle

8.21X

50-/
Qrafe 4

8.96X

30-

Grade 3

14.93X
.'< 2.5cm
> 2.5 cm

N-133
Grade 2

20.90%

Facial Nerve Grades: Minimum 1 Year FU (See Table 4)
FIG. 6

Overall facial nerve function as percentage. (FU>1 year).

Facial Nerve Grading (See Table 4)
FIG. 7
Facial nerve grading as function of tumour size. (FU>1 year).

170

J. E. M. DUTTON, R. T. RAMSDEN, R. H. LYE, K. MORRIS, A. O. KEITH, R. PAGE, J. VAFIDIS

from the vastus lateralis was used to obliterate the translabyrinthine defect. Another patient developed a haematoma in the abdominal fat donor wound but required
no treatment. One patient temporarily developed the
syndrome of inappropriate ADH secretion, and another
had his latent diabetes mellitus unmasked.
Dreams
We have noted that several of the patients in this series
have experienced vivid, frightening and recurrent
illusions which appear to be hypnagogic hallucinations,
both before and more frequently for a period of several
days following surgery. These often took the form of
brightly coloured and changing abstract shapes which
strangely might convey to the patient a tactile sensation
of texture, or of threatening anthropomorphic forms.
Unlike normal dreams they were usually readily recalled
even weeks after they had ceased to occur. The frequency of these experiences was studied by comparing a
small group of thirty-two acoustic neuroma patients,
most of whom had had a suboccipital operation with
eleven patients undergoing posterior fossa exploration
for other reasons (nearly all for removal of other
tumours). There was no obvious difference between the
groups with respect to age, sex distribution, duration of
surgery or anaesthetic agents used. Seventeen of the
acoustic neuroma group (53 per cent) experienced hallucinations compared with two patients in the nonacoustic group (18 per cent). The possible mechanism of
this phenomenon is explored later.
Discussion
Population characteristics
Age and sex distribution
In this series of 151 patients there is a female:male
ratio of approximately 1.5:1, which corresponds with
many other published series. The age distribution by
decades indicates that the greatest incidence is between
the age of 40 and 60. If however the decade distributions
are analysed by sex it will be seen that there are differences and that whereas only 11 per cent of males present
under the age of 40, the figure is 33 per cent for females
(Figs. 1 & 2). these figures, together, with the observation that tumour growth in females seems to be influenced by pregnancy and puberty (Wiet et al., 1989) raise
the possibility that acoustic schwannomas may be hormone dependent. Siglock et al. (1990) found a higher
incidence of oestrogen receptors in tumours removed
from women than from men. A further study (Curley et
al., 1990) did not reveal a significant incidence of hormone receptors in such tumour specimens.
Tumour size
Accurate estimation of tumour size is a remarkably
difficult exercise and there often seems to be a difference
of opinion between the neuroradiologist, measuring
from a CT image and the surgeon at the time of the operation. For that reason any attempt to define tumour size
more accurately than we have seems fruitless. 2.5 cm
was suggested by King and Morrison (1980) as a cut off
between medium and large tumours, and has some prac-

tical value in that a tumour of that diameter is beginning
to involve the trigeminal nerve and to make contact with
the brain stem (May, 1986). To date all tumours larger
than 4 cm diameter have been removed by the suboccipital route. It will be seen (Fig. 4) that although just
over half the tumours are 2.5 cm or less, the number of
wholly intracanalicular lesions is still small and is considerably exceeded by the very large tumours (>4 cm).
Little can be done to overcome the delay between the
onset of symptoms and the patients' seeking professional advice, but even when they do it is still too often a
case of delayed referral to hospital and sadly missed
diagnosis by otolaryngologists. Thomsen and Tos (1990)
highlight the magnitude of the problem of diagnostic
delay, and quote a mean delay of seven years in 233 out
of a series of 300 patients. The reason why this should be
so is discussed elsewhere but owes much to undue
dependence on traditional but insufficiently reliable
audiometric techniques (Ramsden et al. in preparation).
It is to be hoped that as MR imaging becomes more
widely available and surgeons are prepared to refer
patients at an early stage for this investigation, the ideal
of prompt diagnosis may be attainable (Stack et al.,
1988). The economic as well as human sense of such a
policy has been highlighted recently by Moffat et al.
(1989b).

Tumour removal
'Total' tumour removal was achieved in 142 instances,
but as pointed out above this figure includes 21 cases in
which a tiny fragment of tumour capsule was left on the
brain stem, because manipulation and attempted removal brought about dramatic alterations in pulse or
blood pressure. These vestiges were never larger than
1-5 mm and were vigorously diathermized. These
patients have been regularly followed up for up to 12
years, and no clinical or convincing CT evidence of
recurrence has been detected. The single tumour recurrence in the series as a whole was not one of the cases in
which a fragment was knowingly left behind. A recent
MR:Gd DTPA study by Pace-Balzan et al (in preparation), however, indicates that although approximately 50% of such remnants do disappear, some do
persist and may be capable of slow regrowth. Despite
the paucity of references to this problem in the literature, we cannot believe that we are the only workers to
have encountered it. Our experience suggests that these
cases can usually be regarded as having had total tumour
removal, but regular follow-up with MR:Gd DTPA is
advisable.
Total removal is the usual aim but there have been
occasions in which subtotal excision has seemed the
safer alternative. Such instances have almost always
involved elderly or medically unfit patients in whom
there was a pressing neurosurgical indication for surgery. Such patients had translabyrinthine subtotal
excisions, in the interests of speed, with tumour residue
left on both the facial nerve and on the brain stem. The
behaviour of the tumour remnant has been monitored
by annual CT scan and as yet no significant enlargement
of the residual portions have been observed.
As in most series, there are a number of elderly
patients with small (<1 cm) tumours from whom surgery

ACOUSTIC NEUROMA (SCHWANNOMA) SURGERY 1978-1990

has been withheld. In only one of the nine has the
tumour grown significantly (doubled in size in two
years), and that patient is now scheduled for surgery.
Three young patients with small tumours in only hearing
ears are also under annual review.
The factors which influence growth in acoustic
tumours are poorly understood. Wiet et al. (1989) could
identify no factor, clinical (including age), audiological
or histological that predicts tumour growth. Valvassori
and Guzman (1989) in a radiological follow-up study of
35 patients aged 16-81 who for a variety of reasons did
not have their acoustic tumours removed, were unable
to establish a correlation between rate of tumour growth
and either age or clinical findings. They were able to conclude that the growth rate though unpredictable was
usually minimal to moderate, and felt that if the patient
is over 50 and the tumour small (1 cm or less) surgery
should not be performed but the patient should be
scanned by MRI every two years. The present authors
feel that 65 is a more realistic age for such a "wait and
scan" policy. The rate of growth of tumour fragments
left behind at primary surgery may well be different from
that of the original tumour (Pace-Balzan et al., in preparation). Beatty et al. (1987) point out that unless one
adopts a policy of routine CT scanning in all patients for
up to seven years after primary surgery, recurrences may
be missed, and in the absence of a functioning eighth
nerve, clinical clues to recurrence will remain absent
until the tumour makes contact with the brain stem.

Mortality and morbidity

The findings of this study confirm those of other
authors that the death and complication rate following
acoustic neuroma surgery is higher with large tumours
(>2.5 cm) than small, the only exception being CSF fistula which is related to surgical approach. Of the five
perioperative deaths and two major brain stem syndromes in this series, all but one (case 5 above) were
associated with tumours of 3 cm or greater. It is also significant that of these seven cases, six were seen during
the first 100 and only one subsequently. It is certain that
increasing experience leads to better results, and this
strengthens the argument that acoustic neuroma surgery
is best performed in centres specializing in such work, by
an otologist and a neurosurgeon working as a team. This
has certainly been the philosophy in this series. As
regards the choice of surgical approach we have tended
not to be rigidly committed to one route, but have developed a flexibility which has allowed us to acquire equal
familiarity both with the translabyrinthine and the suboccipital operations. Obviously as experience is gained
with the translabyrinthine route so the surgeons, and in
particular the neurosurgeons for whom it is a less familiar approach, become happier with the size of tumour
that they feel they can remove through it. It is to be
hoped that the experience gained in the removal of large
tumours through the suboccipital route will pay dividends with small lesions in which hearing preservation is
attempted. The advantages and disadvantages of both
techniques are objectively reviewed by King (1988), and
we subscribe strongly to the view expressed by him that
the ideal is to have both methods available, and to tailor
the approach to the demands of the individual case. The

171
middle fossa approach has not been employed because
we feel it does not provide sufficient flexibility of access
in the event of the unpredictable being encountered, for
example a tumour larger than anticipated, or bleeding in
the posterior cranial fossa. Furthermore the facial nerve
is more likely to be damaged because it is situated
between the surgeon and the tumour.
Functional preservation of the facial nerve is now one
of the surgeon's main objectives when removing an
acoustic tumour, and we suggest that reports which refer
only to anatomical preservation of the nerve are worthless. Microsurgical technique, confident identification of
the facial nerve in the internal meatus and at the brain
stem, and adherence to the correct surgical plane
between tumour and arachnoid all contribute, but the
most important factor in preservation of the facial nerve
is the single minded desire to do so, and to take as long
over doing so as is necessary. In this respect the anaesthetic techniques which free the surgeon from the tyranny of the clock must be acknowledged. As yet we have
been reluctant to place too much reliance on per-operative monitoring of facial nerve funcion, and support the
view expressed by Luetje (1989) that the surgeon should
rely more on skill, experience and clinical judgement
than "bells, beeps and whistles" to keep him out of
trouble. The grading system which we have used has one
fewer category than that of House and Brackmann. The
problems with any system is that it seeks to impose a finite number of groups on a continuum from total paralysis
to normality. Nevertheless it does allow one to make
comparisons and it is quite clear that tumours of 2.5 cm
and smaller carry a very much better prognosis than
those greater than 2.5 cm (Fig. 7). Apart from tumour
size, the other important factor in post-operative facial
nerve function is the presence or absence of weakness on
presentation. Although no patient in this series had a
major degree of pre-operative weakness there were
nevertheless a number with minor degrees of facial
asymmetry, usually of blink. Those tended to have a less
favourable outcome (Lye et al., 1982). The assignation
of a grade to a facial nerve is not however immutable. In
an admittedly small study of a group of our patients
Vafadis et al. (1987) found that a number of patients with
grade 1 face at one year, subsequently slipped back into
grade 2 at between two and five years, usually by
developing slight synkinesis or a minor degree of dryness
of the eye. Our small numbers of patients who underwent facial nerve grafting procedures suggest that in
every case a better functional result may be expected
from a primary facial nerve graft (cable, rerouting with
end to end anastomosis or direct reimplantation into the
brain stem) than from facio-hypoglossal or facio-accessory anastomosis (grade 3 as against grade 4).
One of the main criticisms of the translabyrinthine
operation has been the high incidence of CSF fistula, and
it is certainly true that the technique of closure is of great
importance if problems are to be minimized. The middle
ear is packed with small pieces of muscle in order to
occlude the opening of the Eustachian tube. To do this
adequately the incus is removed, but in doing so great
care must be taken not to displace the stapes from the
oval window. In the most recalcitrant of the fistulae in
this series, the source of the leak was eventually identified as the oval window after two re-explorations. In the

172

J. E. M. DUTTON. R. T. RAMSDEN, R. H. LYE, K. MORRIS. A. O. KEITH. R. PAGE. J. VAFIDIS

early cases the cavity in the temporal bone wasfilledwith
a free muscle graft, but in the last six years strips of
abdominal fat have been used. A flap of periosteum
from the mastoid process, based superiorly on the temporalis muscle, is sutured back over the fat graft. We feel
that a firm pressure dressing, maintained in position for
four days is of the greatest importance. CSF leakage
after a posterior fossa exploration results from entry of
fluid into the mastoid through air cells that have been
opened at the time of surgery. These are closed with
bone wax and muscle. If a leak persists, it may be sealed
by re-exploring the posterior fossa and identifying and
sealing the fistula, or by performing a mastoidectomy
and inserting a fat graft in the manner outlined above.
Robson et al. (1989) have suggested a technique using a
pedicled sternomastoid flap but we have not employed it
as yet. There have been no donor site problems since the
introduction of abdominal fat, but there were two cases
of infection of the thigh wound in the early part of the
series when muscle was used.
Hearing preservation
Our enthusiasm for hearing preservation has been
tempered by the realization that there were very few
patients in the series in whom the hearing on presentation was good enough to justify the attempt. Itis essential to differentiate between hearing which is preserved
at its pre-operative level, a result which may satisfy the
surgeon, and hearing which is preserved at a level which
satisfies the patient. In his analysis of the differences
between the patients' and surgeons' view of "successful"
tympanoplastic surgery, Gatehouse (1990) states that it
is necessary to bring the hearing level in the operated ear
to within at worst 40 dB and preferably to within 20 dB
of the good side for the patient to appreciate the
improvement. With these predominantly conductive
hearing losses the deafness is purely a result of attenuation. If the same criteria are applied to acoustic
schwannoma patients in whom there is often the additional consideration of impaired speech discrimination,
it will be realized that there are in fact a very small
number in whom useful hearing can be preserved.
Nevertheless, we have encountered a few patients in
whom we have attempted hearing preservation, but as
yet we have failed to achieve a result that satisfied the
patient. The patients for whom hearing preservation
would of course be of crucial importance, are those with
bilateral tumours and useful pre-operative thresholds,
and it is in the hope of being able to help those patients
that surgeons should attempt to perfect the necessary
techniques.

Dreams
The realization that a number of our patients experienced bizarre disturbances of visual imagery in the first
few days following surgery came slowly because it was
not a complaint about which one would necessarily think
of questioning the patient. Once it had been drawn to
our attention that such phenomena were occurring, then
direct questioning revealed it with remarkable frequency. Certain features seemed to recur regularly. The
imagery was always vivid, almost luminous, and seemed

to convey a strong sense of texture. There were frequently strange anthropomorphic and geometric forms
and surreal conversion of everyday objects. The effect
on the patients was of disproportionate terror considering the innocuous subject matter. These episodes which
lasted for no more than a minute or two, occurred at any
time of the night or day, and in some subjects could be
activated almost at will by closing the eyes. Indeed one
subject, an artist, was able to reproduce the images on
paper by doing so. All of these features suggest that the
phenomena were similar to hypnagogic hallucinations,
the "sensory errors of half sleep" (Parkes, 1985). They
are described in normal subjects on sudden wakening
from REM sleep, and in between 20 per cent and 50 per
cent of narcoleptics. Benzodiazepines may cause hypnagogic hallucinations at sleep onset but the patients in
our series were not on these or indeed any psychotrophic
drugs. The phenomenon has also been described in sleep
apnoea.
Experiments in cats have revealed the existence of
phasic potentials in the pons which are conducted rapidly via the lateral geniculate body to the occipital cortex. These PGO waves (pons-geniculate body-occipital
cortex) have been seen to coincide with the onset of
REM sleep. They appear to originate at the level of the
dorsal pontine tegmentum, and they seem to have certain features in common with the cortical K complex
recordable in humans (Parkes, 1985). It would seem
possible that the initiating factors in the genesis of these
hallucinations might be minor degrees of vascular
embarrassment in that very discrete part of the brain
stem, supplied by branches of the anterior inferior cerebellar artery, or perhaps more plausibly by congestion in
the local veins which drain into the vein of the lateral
recess or the lateral medullary vein.
Conclusion
The results of over a decade of close collaboration
between otologist and neurosurgeon in the management
of acoustic neuromas indicate that the factors that influence the successful outcome of surgery both in terms of
mortality and of morbidity are tumour size and surgical
experience. These observations are particularly germaine in respect of facial nerve function. Tumours of
2.5 cm or less carry a very much better prognosis for
facial function than larger lesions. Tumour size is a
reflection of tumour diagnosis. It is still sadly true that
almost half the tumours in this series measured more
than 2.5 cm in diameter at diagnosis. MRI with gadolinium enhancement will allow the smallest of tumours to
be diagnosed, and should be the investigation of choice
of all patients suspected of harbouring an acoustic
schwannoma. If the facial nerve is known to have been
sacrificed, immediate grafting is to be preferred to faciohypoglossal or facio-accessory anastamosis.
The incidence of other serious complications is low
and these too are usually related to tumour size. The
coexistence of a total facial palsy and trigeminal dysaesthesia is to be regarded as potentially grave, with
important implications for the health of the eye. CSF fistula, whilst always to be taken seriously, has not presented any major problems in this series.
We have developed a surgical philosophy which has

173

ACOUSTIC NEUROMA (SCHWANNOMA) SURGERY 1978-1990

allowed both members of the team to become familiar
with both the translabyrinthine and suboccipital
approaches, and feel that the results of the surgery have
less to do with the approach than with the experience of
the surgeons. Hearing preservation has remained an elusive goal, because of the small numbers who appear to
justify the attempt. Nevertheless it does seem probable
that early diagnosis will become a more readily attainable ideal in the next decade, and so the numbers of
patients for whom hearing preservation will be
attempted must increase. At the same time surgeons
must not lose sight of the concept of "useful" hearing
preservation.
Whilst total tumour removal has been our objective,
there are occasions on which a subtotal removal may be
one's deliberate choice, or may be forced upon one by
circumstances. Such patients should be followed up with
regular CT scanning or MRI. There is also a group of
patients in whom a tiny fragment of capsule firmly
adherent to the brain stem, may be diathermized and left
in situ. Although in none of the 21 patients in this group
has there been any convincing evidence of regrowth
either clinically or on CT imaging, it may be that MR:
Gd DTPA studies will indicate that some of these small
fragments are capable of survival and indeed regrowth.
Finally, attention is drawn to the occurrence of hypnagogic hallucinations an unusual manifestation of transient post-operative brain stem dysfunction, not as far as
we can establish previously described.
Acknowledgements
The authors have been supported in this work by the
Brenda and Geoffrey Thornley Foundation (B and GT
Foundation), and gratefully acknowledge its continuing
help in the clinical, educational and research aspects of
acoustic neuroma and related conditions.
The senior authors, JEMD and RTR were TWJ
travelling fellows to North America in 1978, and recognize with gratitude the value of the foundation's contribution to the establishment of the otoneurological
programme in Manchester.
References
Beatty, C. W., Ebersold, M. J., Harmer, S. G. (1987) Residual and
recurrent acoustic neuromas. Laryngoscope, 97: 1168-1171.
Bentivoglio, P., Cheesman, A. D., Symon, L. (1988) Surgical
management of acoustic neuromas during the lastfiveyears. Part
II: Results for facial and cochlear nerve function. Surgical Neurology, 29: 205-209.
Birzgalis, A., Ramsden, R. T. (1991) Intravestibular schwannoma
Journal of Laryngology and Otology (in press).
Key words: Acoustic schwannoma

Curley, J. W. A., Ramsden, R. T.,Lye, R. H., Howell, A., Healey,
K. (1990) Oestrogen and progesterone receptors in acoustic neuromas. Journal of Laryngol and Otology, 104: 865-867.
Gatehouse, S. (1990) Modelling of disability in unilateral hearing
loss, Audiology. In press.
House, J. W., Brackmann, D. E. (1985) Facial nerve grading
system, Otolaryngology—Head and Neck Surgery, 93: 146-147.
King, T. T. (1988) Surgical approaches to acoustic tumours. British
Journal of Neurosurgery, 2: 433-438.
King, T. T., Morrison, A. W. (1980) Translabyrinthine and transtentorial removal of acoustic tumours. Results in 150 cases.
Journal of Neurosurgery, 52: 210-216.
Luetje, C. M. (1989) Intraoperative monitoring: hardware addiction is no substitute for clinical judgement. Editorial. American
Journal of Otology 10(5).
Lye, R. H., Dutton, J., Ramsden, R. T., Occleshaw, J. V., Ferguson, I. T , Taylor, I. (1982) Facial nerve preservation during
surgery for removal of acoustic nerve tumours. Journal of Neurosurgery, 57: 739-746.
May, M. (1986) In The Facial Nerve (May M., ed), Thieme Inc;
New York, p 33.
Moffat, D. A., Croxson, G. R., Baguley, D. M., Hardy, D. G.
(1989a) Facial nerve recovery after acoustic neuroma tumour
removal. Journal of Laryngology and Otology. 103: 169—172.
Moffat, D. A., Hardy, D. G., Baguley, D. (1989b) Strategy and
benefits of acoustic neuroma searching. Journal of Laryngology
and Otology. 103: 51-59.
Parkes, J. D. (1985) In Sleep and its disorders, W. B. Saunders &
Co; London, p 23-24.
Robson, A. K., Clarke, P., Dilkes, M., Maw, A. R. (1989) Transmastoid extracranial repair of CSF leaks following acoustic neuroma resection. Journal of Laryngology and Otology, 103:
842-844.
Siglock, T. J., Rosenblatt, S. S., Finck, F., House, W. F., Hitzelberger, W. E. (1990) Sex Hormone Receptors in Acoustic
Neuromas. American Journal of Otology, 11: 237-239.
Stack, J. P. Ramsden, R. T., Antoun, N. M., Lye, R. H., Isherwood, I., Jenkins, J. P. R. (1988) Magnetic resonance imaging of
acoustic neuromas: the role of gadolinium-DTPA. British
Journal of Radiology 61: 800-805.
Thomsen, J., Tos M. (1990) Acoustic neuroma: clinical aspects,
audiovestibular assessment, diagnostic delay and growth rate.
American Journal of Otology, 11:12-19.
Vafadis, J. A., Page, R. D., Keith, A., Lye, R. H., Dutton, J.,
Ramsden, R. T. (1987). Diagnosis and management of acoustic
neuroma: a ten year experience. Scientific Exhibit. The Eighth
European Association of Neurological Surgeons Congress. Barcelona. September 1987.
Valvassori, G. E., Guzman, M. (1989) Growth rate of acoustic
neuromas. American Journal of Otology, 10:174-176.
Wiet, R. J., Young, N. M., Monsell, E. M., O'Connor, C. A.,
Kazan, R. (1989) Age considerations in acoustic neuroma surgery: the horns of a dilemma. American Journal of Otology, 10:
177-180.
Address for correspondence:
Mr Richard T. Ramsden F.R.C.S.,
Department of Otolaryngology,
Manchester Royal Infirmary,
Oxford Road,
Manchester M13 9WL.