| Message: https://bjo.bmj.com/content/bjophthalmol/70/8/570.full.pdf
British Journal of Ophthalmology, 1986, 70, 570-574
Pseudoepiphora from cerebrospinal fluid leak:
case report
ROBERT M DRYDEN,' AND ALLAN E WULC2
From the 'Department of Ophthalmology, University ofArizona, Tucson, Arizona, and
2University of Pennsylvania Scheie Eye Institute, Department of Ophthalmology, Philadelphia,
Pennsylvania, USA
SUMMARY A 4-year-old tearing child with obstruction of the nasolacrimal duct was treated with
dacryocystorhinostomy three years after naso-orbital injury. However, what appeared to be
tearing peristed, and meningitis developed. Coronal CT scans demonstrated traumatic encephalocele of the posterior superior orbital roof. A chronic orbital cerebrospinal fluid (CSF) leak was
diagnosed. To our knowledge no case of chronic CSF leak has been reported that simulated tearing
in an otherwise asymptomatic child. In the tearing patient who has a naso-orbital fracture the
possibility of chronic CSF leak should be considered.
Lacrimal outflow obstruction is a well recognised
complication of midfacial fractures involving the
maxilla, the lacrimal bones, or the ethmoids.18
Common manifestations of obstruction to the
lacrimal outflow system include epiphora, dacryocystitis, or mucocele.
A case of epiphora thought to be secondary to
nasolacnmal duct obstruction following midfacial
trauma is herein described. A 4-year-old child seen
two years after naso-orbital injury had traumatic
lacrimal duct obstruction and underwent uncomplicated dacryocystorhinostomy (DCR).
The aetiology of tearing in this child was chronic
post-traumatic cerebrospinal fluid (CSF) leakage
that occurred simultaneously with lacrimal duct
obstruction. Despite a successful, DCR, chronic
copious CSF leakage simulating tearing continued.
To our knowledge this is the first reported case of
pseudoepiphora caused by chronic CSF leakage in an
otherwise asymptomatic child.
After midfacial fracture chronic CSF leak should
be considered in the differential diagnosis of tearing,
and the appropriate laboratory and radiographic
tests should be performed.
Case report
A 4-year-old boy was seen with epiphora of the right
eye of two years' duration. A motor-car accident two
Correspondence to Robert M Dryden, MD, 601 N Wilmot Road,
Suite 58, Tucson, Arizona 8571 1, USA.
years previously resulted in hospital admission,
where lethargy progressing to coma and blood in the
middle ear led to the diagnosis of basilar skull
fracture. Computerised axial tomography (CT) was
done three times while he was in hospital and the
scans were reported as normal.
Immediately following hospital discharge persistent right-sided tearing was noted. There was no
irrigation or pain, and tearing was not related to
eating or salivation. Ocular examination showed
visual acuities of 20/30 in both eyes. Right inferior
oblique overaction was noted with a V-pattern
exotropia. Palpation of the orbit revealed no anterior
defects and a normally positioned, non-pulsatile
globe with no rim defects.
Lacrimal examination with a Jones primary dye
test showed no passage of dye after 40 minutes.
Traumatic lacrimal drainage obstruction was inferred.
An examination under anaesthesia revealed an
obstructed nasolacrimal duct. No fluid could be
irrigated through either the upper or lower canaliculus into the nose. Probing revealed a tight internal
common punctum and an obstructed nasolacrimal
duct.
A dacryocystorhinostomy with canalicular intubation with silicone tubing was performed. The bone
was thickened, presumably from past naso-orbital
fracture. The patient received 500 mg of intravenous
cephazolin (Kefzol) and oral cephalosporins were
continued for seven days.
570
Pseudoepiphora from cerebrospinalfluid leak: case report
On the fifth postoperative day mild erythema of
the scar site was noted. Tearing had not ceased.
One week after operation the patient developed
fever with headache and vomiting. He was admitted
to a community hospital with a presumptive diagnosis
of periorbital cellulitis and begun on intravenous
cephazolin. Two days later he complained of neck
pain. A lumbar puncture was done which showed
cloudy CSF, raising the possibility of bacterial meningitis. An additional antibiotic was instituted. An axial
CT scan was performed which showed air and soft
tissue swelling of the right orbit consistent with the
operative appearances as well as an opacity of the
right ethmoids consistent with the postoperative
haemorrhage he had had.
In the ensuing two weeks the patient developed
recurrent spiking fevers accompanied by meningeal
signs. Intravenous antibiotics were altered to methicillin and chloramphenicol, and he remained afebrile
until three weeks after operation, at which time
another temperature spike to 101 degrees was noted.
The patient was transferred to Tucson Medical
Center for further evaluation.
On admission, the temperature was 38 3°C and the
vital signs were otherwise normal. No periorbital
cellulitis was noted. The dacryocystorhinostomy scar
was healing well. The silicone stent was in place.
Tearing was persistent and the results of the eye
examination were otherwise normal. The medical
examination, including a neurological examination,
gave results within normal limits. All laboratory tests
including examination of the CSF gave normal
results.
On the second day in hospital an ophthalmic plastic
surgery opinion was sought. The child continued to
Fig. 1 Coronal CTdemonstrating DCR ostium and
thickening ofnasal mucosa on the right (arrow).
'tear' profusely in the right eye. A Jones primary dye
test showed dye in the oropharynx at 2 minutes. It
was concluded that the dacryocystorhinostomy surgery had been successful and that causes of pseudoepiphora should be sought to explain the 'tearing'.
Our examination included a Dextrostix test of the
tear fluid. The sugar content of the tear fluid was
observed. The Dextrostix test of the right eye showed
a glucose level of 100 to 180 mg/dl (5-6-10.0 mmol/l).
The glucose level from the left eye was less than 100
mg/dl (5.6 mmol/l). A specimen of the patient's tears
was collected and sent for laboratory analysis. A
discrepancy of 2 to 1 in the glucose concentration
existed in the right and left eyes. The data strongly
suggested CSF leakage into the orbit. A CT scan was
performed (Figs. 1 and 2). A gas lucency in the region
of the right orbit with thickening of the right medial
rectus and opacification of the right ethmoid was
observed. Anteriorly the DCR ostium was visible
along with thickening of the nasal mucosa (Fig. 1). A
bony defect was noted in the posterior orbit that
did not communicate with the DCR ostium (Fig.
2).
On the tenth day in hospital a frontal craniotomy
was performed. An encephalocele wasfound into a
fracture in the posterior medial orbital roof. It was
retracted from the fracture site, and a Silastic implant
was placed over the defect. The dural tear was noted
and repaired with a free pericranial graft and closed
with a single running suture. Postoperatively the
patient did well and was discharged six days later.
Tearing ceased immediately after surgery. Measured
glucose levels in both eyes were the same and less
than 100 mg/dl (5.6 mmol/l).
'demonstrating defect in orbital roofon
the rignt (arrows).
571
Robert MDryden and Allan E Wulc
Discussion
CSF leak has been reported to occur in 6% to 35% of
patients with severe non-penetrating intracranial
injury.'59 It occurs in approximately 20% of patients
with midfacial fracture or fractures through the
paranasal sinuses.:
CSF leaks are rare in children below the age of 2
because of the flexibility of the cranial base and the
relative immaturity of the paranasal sinuses. However, in children older than 2, CSF leakage is by no
means uncommon."'
Traumatic CSF leaks usually result from tears in
the dura at the skull base that allow CSF to escape
from the subarachnoid space into the nasopharynx,
the paranasal sinuses, or subcutaneously. The floor
of the anterior cranial fossa, particularly the cribriform plate area, is thin, and the dura at this site firmly
invests the olfactory fissure where the olfactory nerve
penetrates the skull. Midfacial trauma may result in
penetration of the anterior cranial fossa by ethmoidal
roof or wall fragments and dural tear. CSF escapes
immediately. If the edges of the dura are in apposition and do not gape, healing is relatively rapid and
CSF leak resolves in three to 10 days. Sinus mucosa
may bridge a gaping dural defect and cause leakage to
cease. However, if bony spicules project through the
tear, a portion of the brain herniates through the
defect. If a large dural defect is present, healing may
be incomplete. High CSF flow due to coughing or
straining may also result in late onset of CSF leakage
due to rupture of a partially healed tear. Dural tears
may be enlarged by brain and high amplitude CSF
pulsation with resultant erosion of bone.9 "
The diagnosis, aetiology, and management of
fractures of the orbital roof have been well described by McClachan et al. 2 Rhinorrhea is the
most common presenting symptom of CSF leakage
and presents in 25% of patients with midfacial
fracture.' Hypo-osmia or anosmia may occur from
damage to olfactory nerve fibres in 5% of patients.9
Smell may be intact in patients who have anteriorly
placed frontal sinus fractures or sphenoidal sinus
fractures. If fracture of the petrous portion of the
temporal bone has occurred, CSF from the middle or
posterior cranial fossa may escape into the mastoids
and produce middle ear fluid mimicking serous otitis
media. Rarely CSF tracks subcutaneously, producing
a subepicranial hydroma.'3 Headache occurs in 15%
of patients with CSF leakage and is therefore thought
to be an inconstant symptom.9 Meningitis occurs in
between 3% and 50% of cases of CSF leakage and the
incidence is higher when the leakage is chronic."`116
The observant patient may describe a gush of salty
fluid, or choke on arising in the morning. In the
recumbent position leaking CSF collects dependently
in the sphenoid sinus. On the patient's standing, an
influx of CSF collected overnight rushes into the
oropharynx causing the so-called reservoir sign.'7
Other signs of CSF leakage include frequent
awakening from sleep with coughing fits and soaking
of the bed sheets with clear nasal discharge each
night. Rhinitis has also been described from chronic
CSF leakage.9"
CSF leakage following midfacial fracture in an
otherwise asymptomatic patient may be confused
with traumatic epiphora. In a series of 300 patients
with midfacial fracture where 35% had CSF leakage,
5 3% had concomitant epiphora or dacryocystitis.5 In
a series of 100 patients reported by Campbell with
midfacial fracture 12% had complaints of tearing.'8
Naso-orbital trauma may injure the medical palpebral tendon, lacerate or compress the nasolacrimal
sac or duct, or cause cicatricial ectropion of the lower
lid or punctal malposition.478 While CSF leakage
usually resolves, dacryocystitis and epiphora commonly persist until surgically treated.
Few cases of chronic CSF leakage into the orbit
have been reported in the neurosurgical or ophthalmic literature. In a review of giant epidermoid
and dermoid tumors of the orbit Carey described two
cases of chronic CSF leakage into the orbit thought to
be from direct communication between the subarachnoid space and the inner portion of the tumour. 9 In
his series other symptoms including pulsating proptosis, and extraocular movement limitation were also
observed. 9
A case of traumatic CSF fistula simulating epiphora
occurred in a child 8 months of age and was reported
in the neurosurgical literature.2" Epiphora occurred
immediately following head trauma. The patient
presented with coma, hemiparesis, and facial weakness and developed a pulsating exophthalmos accompanied by bruit and chronic epiphora. Chemosis and
epiphora resulted from a leak through the ethmoidal
sinuses directly to the conjunctiva. The child underwent surgical repair of the fistula soon afterwards.
A case of CSF fistula in a 22-year-old victim of a
motor vehicle accident with a cranial nerve palsy has
been reported. This patient had a pulsatile right
upper lid and chemosis from direct subarachnoid
communication with the upper lid. No tearing was
observed at any time.2'
The present case of CSF leakage differs from the
above-mentioned cases and has many interesting and
hitherto undescribed features. It occurred chronically and continuously in an awake and alert child for
a period of three years following injury. Proptosis,
chemosis, or eyelid swelling were not noted at any
time, and neurological deficit did not develop. No
pulsations were observed. Finally, tearing was not
suspected to be secondary to orbital CSF leakage
572
Pseudoepiphorafrom cerebrospinalfluid leak: case report
until after successful lacrimal drainage surgery had
been performed.
CSF leakage occurred as a result of dacryocystorhinostomy surgery in two cases reported by Neuhaus
and Baylis and may occur if bone removal extends to
the level of the cribriform plate.22 The same authors
noted an average of 5 mm between the nasal ostium
and the floor of the anterior cranial fossa.22 While it is
possible that in the reported cases DCR surgery
could have created a CSF leak, it is unlikely. Tearing
was chronic prior to surgery and persisted after
surgery, and it promptly resolved following neurosurgical repair of the posterior orbital encephalocele.
Other unlikely hypotheses that might explain the
findings in the presented case include the possibility
that CSF leaked from the canaliculi in retrograde
fashion from the abnormality of the cranial vault,
simulating tearing. The CT evidence for this theory is
lacking, as the encephalocele was well posterior to
the lacrimal drainage system. While remanipulation
of a traumatically altered orbit and cranial base
undoubtedly introduced meningeal infection, the
pathway of CSF leakage from posterior orbit to the
tear film is a matter of speculation.
Reagent strips such as Dextrostix and Labstix
commonly used for urine glucose determination,
while not helpful in distinguishing CSF from nasal
discharge,'223-26 may be helpful in diagnosing CSF
leakage into the orbit. CSF glucose is approximately
70% of that of serum glucose. Normal levels of
glucose in tears are insignificant.26 27 For the purposes
of the clinical examination in the reported case, tear
concentration in 20 normal individuals without diabetes mellitus was measured with a Labstix reagent
strip by placing it directly on the tear meniscus and
waiting until wetting had occurred. After 30 seconds
the glucose level was measured by matching the
colour of the paper with the colour of the glucose as
determined on the table found on the bottle. In all
patients the glucose level was found to be within 0
and 100 mg/dl (0 and 5 6 mmol/l). In the case
described here the glucose level on the uninvolved
side was from 0 to 100 mg/dl, corresponding to the
normal value. On the involved side the tear glucose
level was between 100 and 180 mg/dl (5.6 and 10-0
mmol/l). The higher glucose level on the involved
side in this patient was confirmed by laboratory
analysis. A large sample needs to be collected for
glucose concentration to be analysed in a laboratory.
The reagent strips are inexpensive, immediate, and
require substantially less tear volume and were
effective in diagnosing the leakage.
The site of CSF leakage may be localised radiographically with or without the aid of contrast
material. Skull films may show an air fluid level in the
sinuses.9 Materials can be instilled intrathecally to
demonstrate leakage of CSF. Methylene blue and
fluorescein were at one time injected but have been
abandoned owing to numerous complications."2"29
Radioiodide labelled serum albumin and technetium
are used to directly visualise CSF leak with cisternography and are diagnostic in 86% of cases."'
Metrizamide encephalocistemography with high
resolution CT is mandatory in evaluation of the
patient with a suspected chronic orbital CSF leak. In
our patient CT scans with axial views were obtained
at the time of the original trauma and repeatedly
thereafter to evaluate brain, orbits, and skull base
and were all within normal limits. One week following DCR axial CTs again did not show any abnormality of the skull base or orbital roof. However,
coronal CTs with 1-5 mm cuts clearly show a posterior bony defect and suggested meningoencephalocele and chronic CSF leak. The different perspective
offered by directly formed coronal views may be an
asset in considering the possibility of chronic CSF
leakage into the orbit.
When a CSF leak is noted in the early period post
midfacial injury, opinions are varied as to whether
direct and immediate reduction of facial fracture or
repair of dural tear is indicated. Most authors
recommend bed rest with the head of the bed raised.
Prophylactic antibiotics are administered intravenously. Bed rest, stool softeners, and medication
to prevent coughing or straining are recommended.
A lumbar drain, placed in the lumbar subarachnoid
space, may be useful in lessening CSF flow through
the dural rent, thus allowing repair to take place.
Likewise, acetazolamide may be administered to
decrease the production of CSF.
The majority of CSF leaks heal without the need
for surgery.9" When the neurological condition has
stabilised, fractures may be reduced with the
realisation that additional trauma to the dura may
result.
The treatment of chronic persistent CSF leak is
neurosurgical."I 1617 If meningitis has supervened, an
adequate period of time should elapse for the patient
to be on appropriate antibiotics. Chronic CSF leaks
usually do not subside spontaneously. The area of the
dural rent must be localised and repaired. In the
reported case a frontal craniotomy with exposure of
the fracture site, Silastic roof implant, and epicranial
duroplasty were performed. The techniques of this
neurosurgical procedure go beyond the scope of this
paper.
Patients who tear after naso-orbital trauma must
be suspected of having a CSF leak even if signs of
nasolacrimal duct obstruction are present. It may be
useful to ask patients specifically about rhinorrhoea
or anosmia, as well as to examine for increased
tearing during the Valsalva manoeuvre. CT scans
573
Robert MDryden and Allan E Wulc
should be reviewed if they have been performed and
coronal sections are preferred. Finally, Dextrostix
can be used to obtain an objective measurement of
tear sugar level in order to diagnose chronic orbital
CSF leak. If CSF leak is a possibility, the patient
should be referred for neurosurgical consultation
prior to any lacrimal surgical intervention.
The authors are grateful to Jack H Dunn, the neurosurgeon involved
in this case.
References
1 Dawson RLG, Fordyce GL. Complex fractures of the middle
third of the face and their early treatment. Br J Surg 1953; 41:
254-68.
2 Converse JM, Smith B. Naso-orbital fractures. Ophthalmology
(Rochester) 1963; 67: 622-34.
3 Converse JM, Smith B. Malunited fractures of the bones of the
orbit. In: Converse JM, ed. Reconstructive plastic surgery;
principles and procedures in correction, reconstruction, and
transplantation. Philadelphia: Saunders, 1964; 2: 645-61.
4 Converse JM, Smith B. Naso-orbital fractures and traumatic
deformities of the medial canthus. Plast Reconstr Surg 1966; 38:
147-62.
5 Morgan BDG, Madan DK, Bergerot JPC. Fractures of the
middle third of the face-a review of 300 cases. Br J Plast Surg
1972; 25: 147-51
6 Beyer CK, Smith B. Naso-orbital fractures: their complications
and treatment. In: Tessier P, Callahan A, Mustarde JC, Salyer
KE, eds. Symposium on plastic surgery in the orbital region.
St Louis: Mosby, 1976: 107-12.
7 Converse JM. Orbital and naso-orbital fractures. In: Tessier P,
Callahan A, Mustarde JC, Salyer KE, eds. Symposium on plastic
surgery in the orbital region. St Louis: Mosby, 1976: 79-106.
8 Beyer CK, Fabian RL, Smith B. Naso-orbital fractures, complications, and treatment. Ophthalmology (Rochester) 1982; 89:
456-63.
9 Lewin W. Cerebrospinal fluid rhinorrhea in closed head injuries.
Br J Surg 1954; 42: 1-18.
10 Caldicott WJH, North JB, Simpson DA. Traumatic cerebrospinal fluid fistula in children. J Neurosurg 1973; 38: 1-9.
11 Spetzler RJ, Wilson CB. Dural fistulae and their repair. In:
Youmans JR, ed. Neurological surgery. 2nd ed. Philadelphia:
Saunders, 1982; 4: 2209-27.
12 McClachan DL, Flanagan JC, Shannon GN. Complications of
orbital roof fractures. Ophthalmology (Rochester) 1982; 89:
1274-8.
13 Cullen JR. Unfamiliar swelling after head trauma. Clin Pediatr
1974; 13:378.
14 Brawley B, Kelly W. Treatment of skull fractures with and
without cerebrospinal fluid fistula. J Neurosurg 1967; 26: 57-61.
15 Mincy JE. Post-traumatic cerebrospinal fluid fistula of the frontal
fossa. J Trauma 1966; 6: 618-22.
16 Raskind R, Doria A. Cerebrospinal fluid rhinorrhea and otorrhea of traumatic origin. Int Surg 1966; 46: 223-6.
17 Dandy WE. Treatment of rhinorrhea and otorrhea. Arch Surg
1944; 49:75-85.
18 Campbell W. The radiology of the lacrimal system. Br J Radiol
1964; 37: 1-26.
19 Carey PC. Epidermoid and dermoid tumors of the orbit. Br J
Ophthalmol 1958; 42: 225-39.
20 Joshi KK, Crockard A. Traumatic cerebrospinal fluid fistula
simulating tears. J Neurosurg 1978; 49: 121-3.
21 Galzio RJ, Lucantoni D, Zenobii M, et al. Traumatic craniopalpebral cerebrospinal fluid fistula. J Neurosurg Sci 1981; 25:
105-7.
22 Neuhaus RW, Baylis HI. Cerebrospinal fluid leakage after
dacryocystorhinostomy. Ophthalmology (Rochester) 1983; 90:
1091-5.
23 Dadeholt H. The reaction of glucose oxidase test paper in normal
nasal secretion. Acta Otolaryngol 1964; 58: 271-2.
24 Kirsch AP. Diagnosis of cerebrospinal fluid rhinorrhea: lack of
specificity of the glucose oxidase test tape. J Pediatr 1967; 71:
718-9.
25 Healy CE. The significance of a positive reaction for glucose in
rhinorrhea. Clin Pediatr 1969; 8: 239.
26 Van Haeringen JH. Clinical biochemistry of tears. In: Milder B,
Weil B, eds. The lacrimal system. Norwalk, CT: AppletonCentury Crofts, 1983: 23-48.
27 Gasset AR, Braverman LE, Fleming MC, McArky RA, Alter
BR. Tear glucose detection of hyperglycemia. Am J Ophthalmol
1968; 65: 414-20.
28 Evans JP, Keegan HR. Danger in the use of intrathecal
methylene blue. JAMA 1960; 174: 856-9.
29 Wallace JD, Weintraub MI, Mattson RH, Rosenagle R. Status
epilepticus as a complication of intrathecal fluorescein, a case
report. J Neurosurg 1972; 36: 659-70.
30 DiChiro G, Reames PM, Matthews WB. RISA ventriculography
and RISA cisternography. Neurology (NY) 1964; 14: 185-91.
31 Allen MB, Jr, Gammal TE, Ihnen M, Cowan MA. Fistula
detection in cerebrospinal fluid leakage. J Neurol Neurosurg
Psychiatry 1972; 35: 664-8.
32 Ashburn WL, Harbert JC, Briner WH, DiChiro G. Cerebrospinal fluid rhinorrhea study with a gamma camera. J Nucl Med
1968; 9: 523-9.
33 DiChiro G, Ommaya AK, Ashburn WL, Briner WH. Isotope
cisternography in the diagnosis and follow-up of cerebrospinal
fluid rhinorrhea. J Neurosurg 1968; 28: 522-9.
Acceptedfor publication 3 December 1985.
574
In chest roentgenograms, the pulmonary meniscus sign is a crescent-shaped inclusion of air surrounded by consolidated lung tissue. The common cause is aspergilloma. A hydatid cyst is, however, the most common cause in endemic areas. |
