Comprehensive Algorithm for Skull Base

The Laryngoscope
Lippincott Williams & Wilkins, Inc.
© 2004 The American Laryngological,
Rhinological and Otological Society, Inc.
Comprehensive Algorithm for Skull Base
Dural Lesion and Cerebrospinal Fluid
Fistula Diagnosis
Cem Meco, MD; Gerhard Oberascher, MD
Objectives/Hypothesis: Skull base dural lesions
and cerebrospinal fluid (CSF) fistulas are potentially
vital conditions whose diagnosis can be challenging.
The authors’ aim was to compose a comprehensive
algorithm that combines the most modern diagnostic
tools in easily applicable patterns to indicate a possible dural lesion or CSF fistula. Study Design: Prospective clinical study. Methods: The authors collected the
data of all patients with suspicion of CSF fistula or
dural lesion, or both, between January 1999 and December 2002. Beta–trace protein, ␤2-transferrin, and
endoscopic and laboratory sodium fluorescein tests;
high-resolution computed tomography; and magnetic
resonance cisternography were used according to the
symptoms and etiological factors. The results of the
diagnostic tools that were used and intraoperative
findings (in case of an operative treatment) were reviewed. Results: From 1999 to 2002, 236 patients were
evaluated because of suspicion of dural lesion or CSF
fistula, mostly after head trauma. Pattern I of the
algorithm was applied for head trauma in dural lesion or CSF leak assessment, pattern II for postoperative CSF leaks, pattern III for evaluation of spontaneous CSF rhinorrhea, and pattern IV for the
assessment of recurrent pneumococcal meningitis related to dural lesions without CSF fistula. By applying the patterns of this algorithm, a dural lesion or
CSF leak that was also confirmed intraoperatively
was detected in 48 patients. Conclusion: The four patterns of the new diagnostic algorithm described in
the present study enable physicians to reliably clarify
suspicions of dural lesions and CSF fistulas and aim
to help them choose the best possible management.
Each pattern uses the optimal combination of CSF
tests and radiological imaging to reach a synergistic
effect for precisely detecting dural lesions or
CSF fistulas. Accordingly, this improves surgical
decision-making when necessary. Key Words:
From the Department of Otorhinolaryngology—Head and Neck Surgery, Salzburg University Paracelsus Medical School, Salzburg, Austria.
Editor’s Note: This Manuscript was accepted for publication December 12, 2003.
Send Correspondence to Cem Meco, MD, Department of Otorhinolaryngology—Head and Neck Surgery, Salzburg University Paracelsus Medical
School, Müllner Hauptstrasse 48, A-5020, Salzburg, Austria. E-mail:
[email protected]
Laryngoscope 114: June 2004
Skull base, dural lesion, cerebrospinal fluid fistula, diagnosis, algorithm.
Laryngoscope, 114:991–999, 2004
INTRODUCTION
Diagnosis and management of dural lesions and cerebrospinal fluid (CSF) fistulas can sometimes be challenging, even for the most experienced skull base surgeons. Dural lesions can pose a wide variety of clinical
presentations from clinically significant CSF fistulas to
dehiscence or scar on dura without CSF leak. Because
they can cause potentially vital complications, undiagnosed dural lesions and CSF fistulas (even if the leak is
occult) should be considered as life-threatening situations.
Bacterial meningitis is the major cause of morbidity and
mortality in these patients, independent of the etiological
factors, presenting type, or onset time.1–11 This aspect of
the problem make it imperative for clinicians to detect,
localize, and repair the dural defect and seal the CSF
fistula.
Dural lesions and CSF fistulas have been found to
occur mostly after trauma, specifically, after 10% to 30%
of skull base fractures.3–11 Other etiological factors such
as neoplasms, congenital malformations, spontaneous
CSF leaks, osteoradionecrosis, and prolactinoma therapy
are more rare.8 –11 Nevertheless, all patients with a suspicion of a dural lesion and CSF fistula require a reliable
state-of-the-art detection or exclusion. In this context, it is
essential to diagnose even hidden dural lesions or occult
fistulas to minimize the chance of secondary ascending
meningitis. For clinically apparent CSF fistulas, a definitive and reliable diagnosis is also essential. Other pathological conditions such as vasomotor rhinitis or rhinorrhea
attributable to unilateral autonomic dysfunction after injury might resemble and imitate CSF rhinorrhea, resulting in unnecessary procedures designed to repair presumptive dural defects.12 However, when to use which
diagnostic tool to reach the most reliable and accurate
result, while possibly avoiding the invasive methods at the
beginning, remains an important issue. Moreover, interpreting the combination of CSF tests and imaging studies
in conjunction with each other is another important issue,
which correlates directly with the diagnosis, localization,
Meco and Oberascher: Skull Base Dural Lesions
991
therapy, and, in case of surgery, approach selection and
timing.
The present study focuses on the development and
validation of a state-of-the-art diagnostic algorithm,
which allows physicians to identify and repair, possibly,
all skull base dural lesions and CSF fistulas by using the
most effective CSF diagnostic tests and imaging methods.
Symptom-based and etiologically based (thus practical
and easy to apply) diagnostic patterns of this algorithm
are discussed, along with illustrative cases and possible
pitfalls in the diagnosis of dural lesions and CSF fistulas.
purpose, a new, faster, less expensive, and more sensitive quantitative CSF test, the ␤-trace protein (␤TP) test, was performed
simultaneously. The ␤TP values calculated from nasal secretions
were interpreted as described by Arrer et al.15
Moreover, in selected cases according to the need for further
evaluation, the invasive sodium fluorescein (Na-fluorescein) test
was used.8,16 After the lumbar intrathecal injection of 0.5 mL
fresh, prepared 5% Na-fluorescein, by using a blue light filter,
this dye was traced along the skull base with an endoscope as
described by Messerklinger.17 The chemical part of this test,
which remains the most sensitive test available, was performed
as described by Oberascher and Arrer.18
PATIENTS AND METHODS
Imaging Methods
The setting of the Department of Otolaryngology (Salzburg
University Medical School, Salzburg, Austria), which is a tertiary
health care center, has been used for the present study. All cases
with a suspicion of dural lesion and CSF fistula, as well as all
head trauma patients whose diagnostic evaluation has been performed in our department between January 1999 and December
2002 were included. The cases found to have had a neoplasm as
the etiological factor were excluded from the group because such
cases demand other diagnostic considerations, placing them outside the scope of the present study. Diagnostic workup for the rest
of the cases was performed according to etiological factors and
symptoms, which were determined by anamnesis and clinical
presentation.
Topographical Diagnosis
In case of a dural lesion or CSF fistula, defining its exact
localization is an important part of the diagnostic process. This
aspect is an obligatory part for the therapy planning, especially if
surgery is needed. For all diagnostic tools that are capable of
indicating localization, it is essential to interpret and report findings that are easily understandable for all persons involved in the
further steps of management. For the standardization of reporting topographical diagnosis, a clinically oriented anatomical classification of the anterior skull base fractures and defects was
used.8 This classification simplifies and constitutes an easily
understandable topographical diagnostic terminology that improves communication with other specialists. It is based on anatomical landmarks, which are the main interest for diagnosis and
therapy. To reach this aim, the anterior skull base was classified
into three (Table I) clinically relevant anatomical compartments.
Cerebrospinal Fluid Diagnostic Tests
As the components of the diagnostic workup, a number of
diagnostic tools were used. These consisted of CSF diagnostic
tests and imaging methods. For first-line screening purposes, the
gold standard of CSF diagnosis, the ␤2-transferrin (␤2-Tr) test,
was used as described by Oberascher and Arrer13,14 in 1986. This
qualitative test has shown its value for decades, and our laboratory has had a large experience with more than 2800 analyzed
samples. In addition, beginning in January 2000, for the same
TABLE I.
Clinically Oriented Anatomical Classification of the Anterior Skull
Base Fractures and Lesions.
Compartment
I
IIa
IIb
III
Laryngoscope 114: June 2004
992
Anatomical Region
Frontal sinus
Cribriform plate
Fovea
Sphenoid sinus
After the CSF diagnostic tests, imaging methods are the
second most important group of tools to diagnose dural lesions
and CSF fistulas. The most valuable of them is high-resolution
computed tomography (HR-CT), which has been shown to be
highly effective, especially in post-traumatic cases.19,20 Highresolution computed tomography can identify the smallest bony
defects along the skull base or indicate dural lesions through
indirect signs, such as pneumatocephalus or fluid level in the
sinuses. In our hospital, we use modern multislice CT systems
that enable us to perform the scan of the entire skull base in 4 ⫻
1- or 16 ⫻ 1-mm fashion in short time sequences of 40 or 10
seconds, respectively. Scanning is performed in the axial plane;
thus, no special positioning of the patient is needed. The saved
axial raw data are reconstructed in any desired plane. Fast,
only-axial-plane scanning also reduces the radiation exposure of
the patient. For the precise evaluation of the skull base, both
axial and reconstructed coronal planes are needed. The frontal
sinus posterior wall (compartment I) and sphenoid sinus lateral
and posterior walls (compartment III) can best be assessed on
axial plane scans, whereas cribriform plate (compartment IIa),
fovea (compartment IIb), and sphenoid roof (compartment III)
can be better visualized on coronal plane scans. Similarly, a
meticulous examination of the temporal bone (high-resolution
positron computed tomography [HR-PCT]) also requires images
in axial and coronal planes. All patients in the present study have
had either HR-CT of the anterior skull base or temporal bone, or
both, according to their clinical findings, except in cases that were
evaluated for a possible spontaneous CSF leak but had a negative
finding on ␤2-Tr or ␤TP test.
Another imaging method that we use in our diagnostic
workup is magnetic resonance cisternography (MR-Cis).20 –22 Although MR-Cis is an imaging examination, it can detect CSF
leaks without the need of using contrast material or spinal puncture, by emphasizing and enhancing the CSF signal with suppression and subtraction of the adjacent background tissue signal. Tracing this high-level signal obtained from CSF in coronal
and sagittal planes, especially in clinically active CSF leaks, also
enables topographical diagnosis of one dural fistula or multilocal
dural fistulas along the entire skull base, including the middle
fossa and the petrous bone, with additional information such as
brain herniation. Because it is a noninvasive and radiation-free
examination, it can be safely repeated in cases with intermittent
or subclinical occult leaks. In our study, MR-Cis was used in
selected cases, mainly if there was clinically suspected rhinorrhea, to detect single or multiple dural lesions, but also to diagnose a possible meningoencephalocele or meningocele.
Diagnostic Algorithm
According to the presenting symptoms and etiological factors of each patient, the diagnostic tools have been systematically
used to reach a diagnosis. Various decisive factors such as sensitivity, specificity, noninvasiveness, ease of use, applicability, and
cost have been taken as criteria for their use. Our systematic
Meco and Oberascher: Skull Base Dural Lesions
approach to reaching a reliable diagnosis based on those criteria
has been formatted and is shown in Figure 1 as a diagnostic
algorithm. Clinical behaviors for different clinical presentations
have been formatted in symptom-based and etiologically based
patterns for easy applicability. The algorithm is based on the
experience gained over decades in our department concerning
diagnosis of dural lesions and CSF fistulas. In the present study,
we prospectively applied this diagnostic algorithm, which is easy
to use in clinical practice, for skull base dural lesions and CSF
fistulas. The collected data were analyzed to check the validity of
this new diagnostic algorithm.
RESULTS
Between January 1999 and December 2002, 236 consecutive cases in all were evaluated in our department for
a possible dural lesion and CSF fistula. The majority of
the cases occurred in patients who were evaluated after a
head trauma, followed by patients assessed for spontaneous CSF rhinorrhea and postoperative CSF rhinorrhea
and patients evaluated after recurrent pneumococcus
meningitis (Fig. 2). We were not confronted with any other
major etiological group. The four different groups had
Fig. 1. Symptom-based and
etiologically based diagnostic algorithm for dural lesions
and cerebrospinal fluid fistulas.
Laryngoscope 114: June 2004
Meco and Oberascher: Skull Base Dural Lesions
993
Fig. 2. Patients evaluated with a suspicion of dural lesions and cerebrospinal fluid (CSF) fistulas.
their specific characteristics and symptoms at the time of
presentation to our clinic. They were classified according
to the possible underlying etiological factors by means of
anamnesis and clinical presentation. This easy but necessary step helped in nearly all cases in choosing the appropriate diagnostic pattern to apply. Each diagnostic pattern had the best possible diagnostic tools, considering
that these groups required different diagnostic approaches and interpretations concerning the dural lesions
and CSF fistulas.
Head Trauma: Pattern I
Patients evaluated after a head trauma made up the
largest group with 177 patients. Overall male-to-female
ratio was 2.9:1. During the workup for head trauma evaluation, in addition to the ␤2-Tr and ␤TP determinations in
nasal secretions, all patients had HR-CT scans to rule out
possible skull base fractures. For head trauma cases, radiological evaluation plays an extremely important role
because a dural lesion or CSF fistula can occur only in the
existence of a skull base fracture or, randomly, through
the olfactory nerve bundles after microtears with the impact of trauma. In 177 head trauma patients, in 84 cases
anterior skull base fractures, in 66 cases lateral skull base
fractures, and in 27 cases no skull base fractures were
found on a radiological basis. As expected, cases showing
no fracture on HR-CT scan also had negative ␤2-Tr/␤TP
test results (pattern Ia).
When evaluated, anterior skull base fractures
yielded diverse characteristics, from massive dislocated
fractures with obvious CSF rhinorrhea to hardly detectable bony fissures. During the examinations 13 patients
having positive ␤2-Tr and ␤TP test results also had positive findings for fractures with dural lesion and CSF fistula on HR-CT scan (pattern Ib). Cerebrospinal fluid fistulas in those patients were validated and repaired by
surgery. Eight patients had successful surgery with an
endonasal approach. In the other five cases, transfrontal
extradural approaches were chosen because the fractures
either were in compartment I or were bilateral. From the
point of the localization, 10 patients had a fistula at compartment II either alone or together with other compartments. Two patients had fistula only at compartment I,
and one patient in compartment III.
Of the remaining 71 cases, 67 had a negative ␤2-Tr or
␤TP test result, and 4 had no statement for a CSF fistula
Laryngoscope 114: June 2004
994
with ␤2-Tr or ␤TP tests. All of the 71 cases had some
degree of fracture reported by HR-CT examination. Moreover, some of them also had direct or indirect diagnostic
signs for a dural lesion or CSF fistula indicated by the
HR-CT. As the next diagnostic step, the invasive Nafluorescein test was performed in those cases. After the
lumbar intrathecal application of 0.5 mL 5% Nafluorescein, only in three patients was it possible to find
an active CSF fistula [pattern Ic]. Similarly, those patients also had successful surgery, all through an endonasal approach, which showed two fistulas at compartment
II and one fistula at compartments II and III. More than
900 Na-fluorescein tests have been performed in our clinic
to date, and no complications have been observed after the
intrathecal application of Na-fluorescein.
There remained 68 patients with negative results or
results that were not interpretable (no statement) on
␤2-Tr and ␤TP tests and negative Na-fluorescein test results, but with a positive HR-CT finding; this was the
most challenging group to diagnose (pattern Id). A number of patients had, because of a blood cloth, mucosal
edema, bone fragment or brain herniation, temporary closure of the dural lesion, consequently, hiding the fistula
and resulting in negative findings on CSF testing. In those
cases, the HR-CT findings such as pneumatocephalus or
the degree of fracture dislocation played an extraordinary
role. Of 68 patients in this group, 7 had pneumatocephalus with dislocations at compartment II (n ⫽ 4), compartment III (n ⫽ 1), and compartments II and III
(n ⫽ 2). Pneumatocephalus alone is a direct sign of a dural
lesion; thus we performed exploration endonasal endoscopic surgery in these patients, finding a dural lesion in
each case, consistent with the HR-CT findings. Similarly,
12 patients had only evident dislocations (⬎3 mm) of the
fracture with negative CSF test results and were graded
as a having high suspicion for dural lesion in our clinic
(Fig. 3). Exploration surgery through an endonasal endoscopic approach also confirmed a dural lesion in eight of
those patients at compartment II (four cases), compartment III (two cases), and compartments II and III (two
cases). In the other four patients found to have no dural
lesion during surgery, the dislocated bone has been moved
and replaced by a mucoperiosteal free flap to allow natural
healing (compartment II, two cases; compartments I and
II, one case; and compartments II and III, one case). EnMeco and Oberascher: Skull Base Dural Lesions
testing (Fig. 4). However, nine other patients had surgery
for conductive hearing loss caused by ossicula dislocation.
Suspicion of Postoperative Cerebrospinal Fluid
Fistula: Pattern II
Fig. 3. Head trauma (pattern Id) in a 17-year-old male patient after
head trauma. All cerebrospinal fluid tests were negative. However,
high-resolution computed tomography (CT) indicated (arrow) a
right-sided, 4-mm bony defect of the ethmoid roof with dislocation.
Endoscopic endonasal exploration revealed a 1-cm dura tear.
donasal approach was chosen in three cases, and one case
had surgery through a transfrontal extradural approach.
The remaining 49 cases had little or no dislocation of
the fracture (⬍3 mm). We rated these findings as having
a low suspicion for a dural lesion and as an indication for
follow-up but not for an exploration surgery. However, in
many of these cases, middle face fractures or frontal sinus
anterior wall impression fractures accompanied the anterior skull base fracture. Nearly half of our 49 cases (24
cases) had an indication for reduction of these fractures
and thus for a functional and reconstructive surgery. The
indication for surgery was impression of the frontal sinus
anterior wall in 16 cases and different types of Le Fort
fractures in 8 of them. In these patients, during the same
procedure, an endoscopic examination of the anterior skull
base fracture was also performed by applying preoperative
Na-fluorescein to exclude a dural lesion. Their fractures
were distributed to compartment I (6 cases), compartments I and II (9 cases), and compartment II (9 cases). In
none of the 24 cases was a dural lesion identified. In
general, if we decide to perform surgery for dural lesions
and CSF fistula, we routinely apply Na-fluorescein preoperatively so that we can better detect CSF leak or multiple
CSF leaks or control water-tight sealing intraoperatively.
The remaining 66 patients had a temporal bone fracture. When compared, patients with anterior and lateral
skull base fractures showed different profiles. First, maleto-female ratios were 5:1 and 1.8:1, respectively. Second,
␤2-Tr and ␤TP test results from nasal secretions were
positive in 13 of 84 patients with anterior skull base
fractures compared with 4 of 64 with temporal bone fractures indicating an otogenic CSF rhinorrhea (pattern Ie).
In all four cases, localization of CSF fistula was detected
by HR-PCT and MR-Cis. Middle fossa approach (2 cases),
mastoidectomy (one case), and both of these combined (one
case) were used to close the dural lesions. Other than
these four patients, none of the 62 patients with temporal
bone fracture had a surgical indication for a dural lesion
based on the HR-PCT scan and findings on ␤2-Tr and ␤TP
Laryngoscope 114: June 2004
Another group that requires dural lesion and CSF
fistula diagnosis is made up of patients who have postoperative suspected, generally unilateral rhinorrhea, especially after skull base or endonasal surgery. To rule out a
postoperative CSF fistula (pattern II), we use, first, ␤2-Tr
and ␤TP tests and, second, MR-Cis or HR-CT (or both)
according to the case. Since 1999, we have had three
patients who had a postoperative CSF fistula. All of these
cases occurred after a dural repair of the anterior skull
base following either trauma or removal of a neoplasm. In
all of them, the ␤2-Tr and ␤TP test results and MR-Cis
findings were positive, whereas HR-CT scan indicated a
bony lesion. All cases were in compartments II and III
(two cases in compartment II and one case in compartments II and III) and had revision surgery confirming the
fistula through endonasal approach.
Suspicion of Spontaneous Cerebrospinal Fluid
Fistula: Pattern III
In contrast to presentation after a trauma or surgery,
sometimes patients present with clear rhinorrhea with
suspicion of CSF without a definitive cause. These patients might be having spontaneous CSF fistulas, allergies, vasomotor rhinitis or other conditions, but a definitive evaluation is required so that a CSF fistula is not
missed. In the time interval of the present study, we
investigated 52 patients with a suspicion of spontaneous
CSF fistula (pattern III). They all had clear nasal discharge. Twenty-three of them were beyond all suspicion
Fig. 4. Head trauma (pattern Ie [otogenic cerebrospinal fluid rhinorrhea]) in a 64-year-old man with clinical cerebrospinal fluid rhinorrhea from the right-side nasal cavity. The ␤–trace protein test result
was positive. High-resolution computed tomography (CT) scan of
the anterior skull base showed no fracture. Axial CT scan of the
temporal bone demonstrated (arrows) a dislocated fracture of the
posterior fossa accompanied with pneumatocephalus. The coronal
planes indicated no fracture of the tegmen. Transmastoid revision
and duraplasty were performed.
Meco and Oberascher: Skull Base Dural Lesions
995
Fig. 5. Spontaneous cerebrospinal fluid (CSF) fistula (pattern IIIb) in
a 55-year-old woman (body mass index, ⬎30 kg/m2). For 6 months,
she had had intermittent rhinorrhea from the right side. After referral
to our clinic she collected the secretion at home in a test tube. The
␤–trace protein test result was positive. The high-resolution computed tomography (CT) scan (A) of the anterior skull base revealed
(arrow) bony defect at the lateral wall of the fossa olfactoria. At the
same site, magnetic resonance cisternography (B) showed (arrow)
an active CSF fistula. Successful endoscopic repair was performed.
because they had unilateral rhinorrhea. We performed
screening ␤2-Tr and ␤TP tests in the first line. Forty-six of
the cases were negative (pattern IIIa), and six of them
were positive (pattern IIIb). In cases with negative findings, there was no need to conduct other diagnostic evaluation because ␤2-Tr and ␤TP tests can detect CSF with
high sensitivity and specificity, especially concerning
draining CSF fistulas. On the other hand, all cases with
positive findings on ␤2-Tr or ␤TP tests also had positive
findings on HR-CT scan and MR-Cis testing (Fig. 5). They
all presented with a unilateral rhinorrhea. Male-to-female
ratio in this group was 1:5, and with the exception of one
patient, all patients had a body mass index greater than
30 kg/m2. All six cases had endonasal endoscopic surgery
to close the spontaneous CSF fistula, which was also confirmed during the operation. In four cases the dural lesions were located in compartment III, and in two cases in
compartment II.
Recurrent Pneumococcal Meningitis Without
Cerebrospinal Fluid Fistula: Pattern IV
The fourth group of patients that we evaluated for a
possible dural lesion or CSF fistula comprised patients
having recurrent pneumococcus meningitis with or without previous trauma. During the study period, we had four
such patients. At the time of diagnostic workup, none of
them had a positive finding on ␤2-Tr and ␤TP testing
(pattern IV). In every case, we found a suspected site at
the HR-CT examination. To evaluate a possible meningocele or meningoencephalocele, we also performed MR-Cis,
which showed meningoceles with CSF accumulation in
two cases and a meningoencephalocele in one case (Fig. 6).
All of these cases had exploration surgery verifying the
site of dural thinning that was causing the meningitis.
Their localization was compartment II in two cases and
compartment III in two cases. Surgery was performed
through endonasal approach in three cases and through a
transfrontal intradural approach in one case.
Overall Outcome
In general, the symptom-based and etiologically
based algorithm helped us to reach a reliable diagnose in
Laryngoscope 114: June 2004
996
Fig. 6. Recurrent pneumococcal meningitis without cerebrospinal
fluid (CSF) fistula (pattern IV): a 30-year-old man had meningitis 18
years earlier, with no trauma history. The patient was unconscious
when admitted. Lumbar puncture showed purulent meningitis.
High-resolution computed tomography (CT) scan (A) of the skull
base showed a normal finding of the temporal bone but a bony
defect of the left-side sphenoid sinus (arrow) with suspicion of brain
herniation. Magnetic resonance cisternography (B) revealed (arrow)
a meningoencephalocele. All CSF test results were negative.
every case we evaluated. All patients who had surgery in
each diagnostic group, as well as the patients who did not
have surgery but had skull base fracture and negative
results on CSF tests, were followed for at least 6 months.
The follow-up relied on ␤2-Tr and ␤TP tests, anamnesis,
and nasal endoscopy in selected cases. Other than three
cases of postoperative CSF fistula, all ␤2-Tr and ␤TP test
results were negative. Among the patients who were followed, no complication (eg, meningitis) has become known
to us to date. The average age of 236 examined patients
was 44.8 years (age range, 19 mo–75 y). With the stateof-the-art techniques it was possible to detect a dural
lesion in 48 patients in a 4-year period. At the time of
diagnosis, 33 patients had clinical or occult CSF fistula
detected by ␤2-Tr and ␤TP or Na-fluorescein tests. Head
trauma was not only the most common cause for CSF
fistula (20 cases) and for dural lesions (35 cases), but also
the largest referral group for suspected dural lesions and
CSF fistulas. Table II shows the topographic localization
of the anterior skull base fractures in the entire head
trauma group. Compartment II was the most common
localization for anterior skull base fractures, as well as for
dural lesions and CSF fistulas. Furthermore, it was also
the most commonly affected region in other etiological
groups, together with compartment III. Thus, the endonasal endoscopic microscopic approach was the most commonly used approach (79%) to repair dural lesions, which
were indicated by the diagnostic algorithm.
TABLE II.
Topographic Localization of Anterior Skull Base Fractures.
Pattern
Compartment
Ia
Ib
Ic
Id
Ie
Total No.
I
I ⫹ II
II
II ⫹ III
III
Total No.
—
—
—
—
—
0
2
3
6
1
1
13
—
—
2
1
—
3
10
15
33
6
4
68
—
—
—
—
—
0
12
18
41
8
5
84
Meco and Oberascher: Skull Base Dural Lesions
DISCUSSION
When the etiological factors of dural lesions and CSF
leaks are considered, head trauma has the highest incidence. Dural lesions and CSF leaks have been found to
occur in 10% to 30% of skull base fractures.7 We had
similar outcomes from our case material; 21.3% of all skull
base fractures caused a dural lesion with or without CSF
fistula. In more detail, anterior skull base fractures
(33.3%) had an incidence of dural lesion that was five to
six times higher than for temporal bone fractures (6.1%).
Whatever the cause, diagnosis of CSF ensures a dural
lesion. In our prospective study, the screening ␤2-Tr and
␤TP tests diagnosed CSF from nasal secretions in 26 cases
(54.2%), and Na-fluorescein testing was needed in only
three trauma cases (6.3%) in 48 patients with dural lesions. On the other hand, in the remaining 19 cases, 15
cases (31.3%) with anterior skull base fracture did not
have active CSF fistulas at the time of sample collection,
which resulted in all CSF test findings being negative
although there was a dural lesion. Temporary closure of
CSF fistula could happen because of a blood clot, bony
fragment, mucosal edema, prolapsed brain, brain edema,
or the like.8,14 Similarly, in four cases (8.3%) with recurrent pneumococcal meningitis, all CSF test results were
also negative. All but one of the four cases had meningoencephaloceles, either congenital or related to previous
trauma. Inadequate integrity or healing of dura at its
weakest site allowed bacterial migration from the skull
base; the cases did not have a CSF fistula.8,14,23 Lack of
CSF fistula differentiates this group from cases with meningoencephaloceles having CSF fistula, possibly because
of elevated intracranial pressure forcing intracranial contents through inherently weakened areas such as widely
pneumatized sinuses, which causes attenuated skull base
areas with bony defects in the latter group.24
Independent from the sensitivity and specificity levels or limits of CSF diagnostic tests that may play a role in
the diagnosis of occult CSF fistulas, even most modern
CSF tests cannot help us if there is no CSF fistula. In such
cases, imaging methods play a more important role; however, as has also been shown in our results, they are not
able to diagnose every dural lesion or CSF fistula alone.
Nonetheless, we ought to arrive at the correct diagnosis in
these patients as well. Therefore, to overcome this problem, we need to combine diagnostic tools in a logical way
that makes sense for every single case. Only by so doing
can we decide to use the next costly, time-intensive, or,
more important, invasive examination that might carry
risks for the patient. During the decision-making process
the valuable hints are in fact given by symptoms and
etiological factors of patients. For that reason, it is our
opinion that etiological factors play an important role
during the planning of an efficient diagnostic process.
Moreover, an efficient, reliable diagnosis for dural lesions
and CSF fistulas can be reached only through a wellplanned diagnostic algorithm. As explained earlier in the
present study, the core rationale is that none of the CSF
diagnostic tests and imaging studies alone can identify
dural lesions and CSF fistulas in all cases. All of them
have their specific limits, which the other examination
methods may or may not cover. Therefore, it is of utmost
Laryngoscope 114: June 2004
importance to know which method to use when, and also
for which patients, to clarify each case in the most specific,
sensitive, noninvasive, and cost-effective way. As shown
in our results and illustrative cases, our patients with
different etiological factors required different examination
methods for the definitive diagnosis. Nevertheless, it is
also important to mention that, sometimes, even the combination of the most modern diagnostic tools leaves us a
“gray zone,” in which we can only discuss a high or low
level of suspicion of a dural lesion. In our series, we had
such a “gray zone” in pattern Id. Concerning this pattern,
within which patients had negative CSF test results but,
somehow, had positive HR-CT findings after head trauma,
all HR-CT findings showing a pneumatocephalus indicated a dural lesion. In the same way, in two-thirds of
patients having bony dislocation of 3 mm or greater at the
site of anterior skull base fractures a dural lesion was
revealed intraoperatively. In contrast, bony dislocations of
less than 3 mm were not an indication for surgery for us in
patients in pattern Id who had a low level of suspicion for
dural lesion; hence, these patients had only follow-up.
With our concept, even in this challenging group, all lesions that most modern diagnostic tools are capable of
diagnosing were diagnosed. Consequently, our diagnostic
algorithm targets reduction in the number of undiagnosed
cases of dural lesion and CSF fistula. Diagnosed cases
must be adequately treated, to minimize the chance of
secondary ascending meningitis.4 For example, 10% to
25% of cases of untreated head trauma with a CSF fistula
proceed to meningitis with a mortality of approximately
10%.6,9
In every patient and with any etiological factor, if we
ever have a suspicion of dural lesion or CSF fistula, any
further step taken without state-of-the-art detection or
exclusion of dural lesion and CSF fistula can, in principle,
result in unpleasant consequences. Bateman and Jones12
have shown that lack of reliable diagnosis before surgery
can result in unnecessary procedures designed to repair
presumptive dural defects. On the other hand, again, to
take the necessary measures in trying to prevent the
potentially vital complications of dural lesions and CSF
leaks, a reliable diagnosis is needed.1– 8 Equilibrium could
be achieved by using a new diagnostic algorithm that
indicates the definitive diagnosis. It is our strong belief
that the definitive, reliable diagnosis obtained by use of
our structured algorithm adds important advantages to
the surgical planning, to surgery itself, and, as a consequence, to its outcome. In our opinion, these aspects have
sometimes been partially neglected in previous studies
focusing on surgical decision-making, techniques, and outcomes.9,10,25–27 Furthermore, other proposed diagnostic
algorithms have generally targeted diagnosis of CSF fistulas regardless of the underlying etiological factors.11,28,29 Also, they have not sufficiently addressed the
cases with a dural lesion but with no active CSF leak.
They have concentrated mainly on the CSF diagnostic
tests and imaging modalities in a single pattern for all
cases. We think that, although these approaches seem to
be practical, they could easily reach their limits and be
inefficient when dealing with the problem cases. Not only
has our experience from the past shown us, but also the
Meco and Oberascher: Skull Base Dural Lesions
997
present study has verified, that it is more practical and
efficient to use custom-made patterns for different etiological groups of patients. For example, the patient shown in
Figure 5 had a clear spontaneous nasal discharge identified as CSF by ␤2-Tr and ␤TP tests, and HR-CT showed a
bony defect on the skull base requiring MR-Cis to localize
the point of the leak, as well as to rule out a possible
meningoencephalocele or multiple fistulas that might
have been overlooked otherwise. Magnetic resonance cisternography has been reported as an efficacious and costeffective imaging technique, with sensitivity, specificity,
and accuracy as high as 87%, 57%, and 78%, respectively,
along with the recommendation of combining HR-CT
scans.21 On the other hand, as in the patient shown in
Figure 3, after a head trauma resulting in skull base
fracture and epistaxis, the CSF test results were negative
but the HR-CT showed a severe bony dislocation in compartment II. In such cases, MR-Cis would not bring any
additional information because there is no active CSF
leak. Although there are dural lesions, temporary closure
of the CSF fistula can happen, as mentioned earlier in the
present study. There is no doubt that the most challenging
patients are those who have a dural lesion but at the time
of presentation are having a temporary closure of the CSF
fistula. Nevertheless, MR-Cis can give valuable information even if the CSF test results are negative. This is true
when it is used for the indicated group, as in the patient
shown in Figure 6, who was admitted to our clinic with
recurrent pneumococcal meningitis without a CSF leak.
Magnetic resonance cisternography revealed a meningoencephalocele at compartment III, which had inadequate
dural integrity as observed during endoscopic repair. Because of such cases and also, more challenging cases, there
is a need for a comprehensive, symptom-based and etiologically based diagnostic algorithm to emphasize the advantages and minimize the disadvantages of diagnostic
tools for reaching reliable and definitive diagnosis.
CONCLUSION
The present prospective study has focused on a practical approach to dural lesions and CSF fistulas. Using our
anamnesis-guided, etiologically based, and symptombased algorithm, patients were divided into four groups.
Through the combination of the most modern CSF tests
and imaging methods available, patients in each group
were evaluated for the most favorable and precise diagnosis. While developing the diagnostic patterns of our algorithm, ease of use, effectiveness, and noninvasiveness
were the main considerations. These measures enable a
broader base of use among different specialties. In general, the aim of this algorithm is to help surgeons indicate
an enhanced diagnosis, whether or not there is a dural
lesion or CSF fistula. In case of a dural lesion, the algorithm extensively targets detection of the site of the lesion
(or lesions). Although there are controversies among different centers about dural lesion and CSF fistula management, especially regarding possible surgery timing and
approach, the enhanced diagnosis reached by the algorithm in the present study would help us achieve the
optimal management for each case. Each pattern of this
comprehensive, symptom-based and etiologically based alLaryngoscope 114: June 2004
998
gorithm uses the optimal combination of CSF tests and
radiological imaging to reach a synergistic effect for precisely detecting dural lesions and CSF fistulas that has
not been reached previously. Accordingly, this would help
us to improve our decision-making for the most favorable
therapy, especially concerning whether or not surgery is
indicated. Furthermore, the algorithm enables surgeons
to apply the most appropriate surgical approach for each
patient from a spectrum of minimally invasive endoscopic
techniques to neurosurgical approaches.
BIBLIOGRAPHY
1. Pappas DG Jr, Hammerschlag PE, Hammerschlag M. Cerebrospinal fluid rhinorrhea and recurrent meningitis. Clin
Infect Dis 1993;17:364 –368.
2. Crawford C, Kennedy N, Weir WRC. Cerebrospinal fluid rhinorrhea and Haemophilus influenza meningitis 37 years
after a head injury. J Infect 1994;28:93–97.
3. Hilary AB. Prophylactic antibiotics for posttraumatic cerebrospinal fluid fistula: a meta-analysis. Arch Otolaryngol
Head Neck Surg 1997;123:749 –752.
4. Bernal-Sprekelsen M, Bleda-Vazquez C, Carrau RL. Ascending meningitis secondary to traumatic cerebrospinal fluid
leaks. Am J Rhinol 2000;14:257–259.
5. Choi D, Spann B. Traumatic cerebrospinal fluid leakage: risk
factors and the use of prophylactic antibiotics. Br J Neurosurg 1996;10:571–575.
6. Eljamel MS, Foy PM. Acute traumatic CSF fistulae: the risk
of intracranial infection. Br J Neurosurg 1990;4:381–385.
7. Marentette LJ, Valentino J. Traumatic anterior fossa cerebrospinal fluid fistulae and craniofacial considerations.
Otolaryngol Clin North Am 1991;24:151–163.
8. Oberascher G. Rhinoliquorrhoe-otoliquorrhoe. HNO 1998;3:
197–200.
9. Hegazy HM, Carrau RL, Synderman CH, Kassam A, Zweig J.
Transnasal endoscopic repair of cerebrospinal fluid rhinorrhea: a meta-analysis. Laryngoscope 2000;110:1166 –1172.
10. Gassner HG, Ponikau JU, Sherris DA, Kern EB. CSF rhinorrhea: 95 consecutive surgical cases with long term
follow-up at the Mayo Clinic. Am J Rhinol 1999;13:
439 – 447.
11. Simmen D, Bischoff T, Schucknecht B. Erfahrungen mit der
Abklärung von Frontobasisdefekten, ein diagnostisches
Konzept. Laryngorhinootologie 1997;76:583–587.
12. Bateman N, Jones NS. Rhinorrhoea feigning cerebrospinal
fluid leak: nine illustrative cases. J Laryngol Otol 2000;
114:462– 464.
13. Oberascher G, Arrer E. First clinical experience with ␤2transferrin in cerebrospinal fluid oto- and rhinoliquorrhea.
HNO 1986;34:151–155.
14. Oberascher G, Arrer E. Efficiency of various methods of identifying cerebrospinal fluid in oto- and rhinorrhea. Otorhinolaryngology 1986;48:320 –325.
15. Arrer E, Meco C, Oberascher G, Piotrowski W, Albegger K,
Patsch W. ␤-trace protein as a marker for cerebrospinal
fluid rhinorrhea. Clin Chem 2002;48:939 –941.
16. Stammberger H, Greistorfer K, Wolf G, Luxenberger W. Operativer Verschlu␤ von Liquorfisteln der vorderen Schädelbasis unter intrathekaler Natriumfluorezeinanwendung.
Laryngothinootologie 1997;76:595– 607.
17. Messerklinger W. Nasenendoskopie: Nachweis, Lokalization
und Differentialdiagnose der nasalen Liquorrhoe. HNO
1972;20:268 –271.
18. Oberascher G, Arrer E. A new method for using fluorescein to
demonstrate oto- and rhinoliquorrhea. Arch Otorhinolaryngol 1996;243:117–120.
19. Lloyd MNH, Kimber PM, Burrows EH. Post-traumatic cerebrospinal fluid rhinorrhea: modern high definition computed tomography is all that is required for the effective
demonstration of the site of leakage. Clin Radiol 1994;49:
100 –103.
Meco and Oberascher: Skull Base Dural Lesions
20. Shetty PG, Schroff MM, Sahani DV, Kirtane MV. Evaluation
of high-resolution CT and MR cisternography in the diagnosis of cerebrospinal fluid fistula. AJNR Am J Neuroradiol 1998;19:633– 639.
21. El Gammal T, Sobol W, Wadlington VR, et al. Cerebrospinal
fluid fistula: detection with MR cisternography. AJNR
Am J Neuroradiol 1998;19:627– 631.
22. El Jamel MS, Pidgeon CN, Toland J, Phillips JB, O’Dwyer
AJ. MRI cisternography and the localization of CSF fistulae. Br J Neurosurg 1994;8:433– 437.
23. Schick B, Weber R, Mosler P, Keerl R, Draf W. Langzeitergebnisse frontobasaler Duraplastiken. HNO 1997;45:117–122.
24. Schlosser RJ, Bolger WE. Management of multiple spontaneous nasal meningoencephaloceles. Laryngoscope 2002;112:
980 –985.
Laryngoscope 114: June 2004
25. Lanza DC, O’Brien DA, Kennedy DW. Endoscopic repair of
cerebrospinal fluid fistulae and encephaloceles. Laryngoscope 1996;106:1119 –1125.
26. Anand VK, Murali RK, Glasgold MJ. Surgical decisions in the
management of cerebrospinal fluid rhinorrhea. Rhinology
1995;33:212–218.
27. Park JI, Strelzow VV, Friedman WH. Current management
of cerebrospinal fluid rhinorrhea. Laryngoscope 1983;93:
1294 –1300.
28. Zapalac JS, Marple BF, Schwade ND. Skull base cerebrospinal fluid fistulas: a comprehensive diagnostic algorithm.
Otolaryngol Head Neck Surg 2002;126:669 – 676.
29. Marshall AH, Jones NS, Robertson IJA. An algorithm for the
management of CSF rhinorrhea illustrated by 36 cases.
Rhinology 1999;37:182–185.
Meco and Oberascher: Skull Base Dural Lesions
999