Isolated cases of spontaneous cerebrospinal fluid (CSF) leakage with and without middle ear encephalocele have been reported. These leaks are usually accompanied by episodes of recurrent meningitis, hearing loss, or chronic headache. In this article, we report seven new cases of spontaneous CSF leakage. Six of these patients had conductive hearing loss and serous otitis media, and three had recurrent meningitis. Prior to a definitive diagnosis, six patients had received myringotomy tubes, which produced profuse clear otorrhea. Three patients had positive beta-2 transferrin assays. Computed tomography and magnetic resonance imaging confirmed a defect in the temporal bone tegmen. A combined transmastoid and middle fossa surgical approach with a three-layer closure was used to repair the tegmen defect. All patients had a lumbar drain placed prior to surgery. In addition to describing the seven new cases, we review the history of CSF leakage and discuss diagnostic methods, surgical findings, and our recommendatio ns for management.
Cerebrospinal fluid (CSF) leaks in the temporal bone are rare. They usually occur in patients who have a history of trauma, congenital defects, otologic surgery, infection, or neoplasm.  CSF leaks in patients who have a history of trauma or surgery occur when the pressure and pulsation of the brain causes a tear in the dura on the edges of a bony defect. Congenital defects, such as Mondini's dysplasia, are associated with an abnormal communication between the subarachnoid and perilymphatic spaces.  Similarly, patent cochlear aqueducts and defects in the cribriform plate of the lateral internal auditory canal have also been reported to cause CSF leakage.  Other congenital structural defects that cause CSF otorrhea include preformed pathways, such as an enlarged facial canal, petromastoid canal fistulae, and patent Hyrtl's fissures. 
The literature contains reports of 56 cases of spontaneous CSF leakage, which is defined as one that has no identifiable cause such as infection, neoplasm, trauma, congenital malformation, or previous surgery.  Surgical exploration is required to confirm the diagnosis. Most of these patients have a bony defect and dural dehiscence.  Several characteristic clinical findings point to a diagnosis of spontaneous CSF leakage. They include clear unilateral otorrhea, rhinorrhea, serous otitis media, recurrent meningitis, conductive hearing loss, aural fullness, the need for multiple ventilation tubes for serous otitis media, and persistent clear otorrhea following myringotomy. [5,6]
Escat reported the first case of spontaneous temporal bone CSF leakage in 1897.  In 1933, Kline identified recurrent meningitis and clear otorrhea as the initial symptoms of spontaneous CSF leakage.  In 1959, Dysart described the tegmen defect and the utility of surgical management with a middle cranial fossa approach.  In 1986, Ferguson et al analyzed 30 cases (26 previously reported and four new cases) in their review of the literature.  In 1992, Pappas et al reported eight new cases and reviewed 33 previously published cases.  At about the same time, Gacek reported three more cases.  In 1995, May et al performed another literature review and found a total of 44 cases.  They also reported 12 new cases, bringing the total to 56.
This article describes our retrospective study of seven previously unreported cases of spontaneous temporal bone CSF leakage. We discuss the initial signs and symptoms, the preoperative evaluation, and the different approaches to surgical management, including our recommended approach.
We retrospectively studied the cases of three men and four women, aged 42 to 67 years (mean: 55), whose followup ranged from 1 to 5 years (table). All patients met the criteria for spontaneous CSF leakage. Three patients had multiple episodes of meningitis. Six patients required myringotomy tube placement for conductive hearing loss, and each subsequently produced profuse clear otorrhea. Only one patient had clear rhinorrhea. All patients underwent computed tomography (CT) and magnetic resonance imaging (MRI) as part of their preoperative workup. Three patients underwent beta-2 transferrin testing, and two had a CT cisternogram. All patients had an identifiable middle ear defect, and one patient had a mastoid defect. Two patients had multiple defects. Three patients had dural herniation with encephalocele, and four patients had dural tears only.
All seven patients underwent surgery with a combined transmastoid and middle cranial fossa approach. Each defect was repaired with an autologous three-layer (fascia-bone-fascia) closure. Each patient also had a lumbar drain placed at the time of the surgery. No postoperative complications were noted. Symptoms resolved in all patients.
Case 1. A 59-year-old man came to our department with a history of intermittent hearing loss in the right ear. He reported no headache, rhinorrhea, or symptoms of meningitis. After a right myringotomy tube was placed, he produced a profuse clear otorrhea, which was positive for beta-2 transferrin. CT demonstrated a defect in the epitympanic tegmen. Combined mastoid and middle cranial fossa surgery was performed to repair the tegmen defect near the head of the malleus. The dural tear was repaired, and the bony defect was reconstructed with a fascia-bone-fascia graft.
Case 2. A 48-year-old woman reported a left conductive hearing loss of 8 months' duration and rhinorrhea, but no other signs or symptoms. Myringotomy revealed the presence of a mass in the middle ear and yielded a profuse clear otorrhea, which was positive for beta-2 transferrin. Radiography detected a defect in the epitympanic tegmen. As in case 1, the defect was repaired with a combined surgical approach, and the encephalocele was amputated from the middle ear side of the defect.
Case 3. A 45-year-old man complained that he had experienced decreased hearing in the left ear for the previous 2 months. A middle ear mass and clear fluid were noted upon ventilation tube placement. CT detected a defect in the epitympanic tegmen, which was repaired with the three-layer closure. During exploration, the encephalocele was amputated from the middle ear. Even though the defect was found around the head of the malleus and the incus, there was no ossicular deformity.
Case 4. A 65-year-old woman with recurrent meningitis denied any history of otorrhea or conductive hearing loss. During her workup for meningitis, CT and MRI had revealed an encephalocele over the malleus and incus. The defect was repaired with the three-layer closure, and the encephalocele was amputated from below.
Case 5. A 67-year-old man had a history of recurrent meningitis and left conductive hearing loss for more than 2 years. He produced clear otorrhea after placement of a ventilation tube. Surgical management included repair of multiple dural tears and bony defects.
Case 6. A 62-year-old woman had a 2-year history of significant recurrent meningitis. Ventilation tube placement produced clear otorrhea. CT, MRI, and CT cisternography showed a direct communication between the CSF space and the left mastoid. A bony defect overlying the incus and malleus was repaired.
Case 7. A 42-year-old woman had recurrent left serous otitis media, conductive hearing loss, and multiple episodes of upper respiratory infection. She had been treated with antibiotics, antihistamines, and decongestants. Her clear otorrhea was positive for beta-2 transferrin. The CT cisternogram revealed a communication between the CSF space and the mastoid. She had two bony defects, one in the area of the sinodural angle and one in the epitympanic tegmen. There was granulation around the defect at the sinodural angle. Both defects were repaired with the three-layer closure.
The pathophysiology of spontaneous CSF leakage in temporal bone is not well understood. Investigators have attributed these leaks to the aging process, decades of CSF pulsation and pressure, and deficiencies in temporal bone remodeling. [10,12-15] Ahren and Thulin reported that the incidence of multiple dural plate defects on autopsy was as high as 6%.  Gacek concluded that herniation of aberrant arachnoid granulations over the middle and posterior fossa might be responsible. 
The most common clinical findings are serous otitis media and recurrent meningitis, and a high index of suspicion when faced with these clues helps make the diagnosis. Also, most of these patients have a history of ear fullness and conductive hearing loss, and most have been treated with antibiotics, decongestants, and steroids for allergies. Clear, watery otorrhea is almost always pathognomonic for CSF. The amount of otorrhea can be increased by compressing the internal jugular vein, having the patient perform Valsalva's maneuver, or placing the patient in a head-down position.
Diagnostic tests. Routine tests for CSF chemistry in middle ear fluid, such as tests for glucose (limit of normal: 60% of serum glucose) and protein (limit of normal: [less than]200 mg/ml), have relatively poor sensitivity and specificity, and, therefore, they should not be relied on. Likewise, the sedimentation test has been used to differentiate CSF from mucus (after 12 hours, sediment usually remains in samples of mucus, while CSF will be clear), but because its false-negative rate is high, it is not clinically useful. 
On the other hand, immunologic identification of beta-2 transferrin is a reliable test, even in samples that are contaminated with mucus, blood, or saliva.  This rapid, noninvasive test is highly sensitive and selective in detecting CSF.  Skedros reported that the sensitivity of the beta-2 transferrin assay is near 100%, and its specificity is approximately 95%. 
When combined with CT and MRI, radionuclide cisternography is useful, but it is not necessary for every patient. We do not use intrathecal agents (fluorescein and methylene) because of their adverse side effects and their inability to precisely localize the leak. 
CT with contrast and MRI are the most important ancillary tests to localize the site of a temporal bone CSF leak. As many as 89% of encephaloceles can be accurately diagnosed this way.  High-resolution CT localizes the bony defect, and MRI defines soft-tissue densities. Because these defects are so small, they are best evaluated with two radiologic studies.
Spontaneous CSF leakage in the middle fossa is usually associated with encephalocele.  In contrast, leakage from the posterior fossa is associated with exuberant arachnoid granulation tissue.  In our study, all patients with encephaloceles had them in the middle ear (floor of the middle fossa). The clinical presentation is almost identical to that of a posterior fossa defect except that the latter has a higher incidence of accompanying meningitis.  We did not observe this in our study.
Surgical management. Both transmastoid and middle cranial fossa surgical approaches, individually and in combination, have been advocated. We routinely employ the combined approach because we feel it provides the best exposure of any tegmen or posterior fossa defect, and it provides a good view of the floor of the middle and posterior cranial fossae. The middle cranial fossa approach greatly optimizes the exposure of the tegmen plate and medial petrous roof (in the event of multiple defects) with a low risk of causing hearing loss. 
Many authors have advocated different craniotomy approaches. Adkins and Osguthorpe  promoted the minicraniotomy, while May et al  preferred a keyhole craniotomy. It is important to remember that a defect anterior to the epitympanic space is not accessible with a keyhole craniotomy. Because the true size and the location of the defect are not always definitively known prior to surgery, even after using all diagnostic modalities, it is important to take advantage of the optimum exposure afforded by the combined approach.
A bony defect can be corrected with different materials and methods. Other authors have recommended Silastic, fascia, stainless steel plates, and mesh. We do not recommend any of these materials because they are difficult to insert and they can interfere with imaging studies. But when a surgeon does elect to use a mesh material, titanium is best because it does not interfere with MRI scanning. Many combinations of autologous tissue (i.e., fat, muscle, fascia, bone, and cartilage) are described in the literature.  Langman et al used adipose tissue to obliterate the mastoid without exposing the defect from above in selected cases.  Atrophy of the adipose tissue could lead to a recurrence of CSF leakage.
We prefer the autologous fascia-bone-fascia graft for a three-layer reconstruction. We use temporal fascia and split calvarial bone grafts harvested from the adjacent outer cortex.  This method provides good structural support without long-term sequelae. The dural defect is usually repaired with 4-0 Neurolon. The encephalocele in the middle ear or mastoid is usually devitalized and can be safely amputated. We do not try to reduce the contaminated tissue. We routinely use a lumbar drain when elevation of the temporal lobe is necessary. Draining off CSF prior to elevation eases the pressure on the temporal lobe. We feel this decreases the small but real risk of venous infarction of the temporal lobe. Allowing the lumbar drain to stay in place following surgery also allows for sealing of the defect, relaxation of the temporal lobe, and diversion of CSF flow. After 48 hours, the lumbar drain is removed and antibiotics are discontinued.
Our management algorithm (figure), which is based on information gathered from our study and others in the literature, is rather direct. After taking a careful history and performing a physical examination, the surgeon should look for fluid emanating from the middle ear. If it is present, a specimen should be sent for a beta-2 transferrin assay. If the test is positive, the patient should undergo both CT and MRI scanning of the temporal bone to localize the source of the leak. If the assay is negative, a CT cistemogram should be performed to detect the egress of CSF into the temporal bone. For patients who have experienced recurrent meningitis without evidence of middle ear effusion, CT and MRI scanning should be ordered to localize any defects. Once the presence of CSF is confirmed by beta-2 transferrin assay or imaging and the defect is localized, the leak can be repaired.
In conclusion, spontaneous CSF leakage should be suspected in patients who experience profuse clear otorrhea after the placement of a ventilation tube for unilateral serous otitis media or unexplained episodes of recurrent meningitis. The most useful confirmatory diagnostic test for fluid is a beta-2 transferrin assay. CT and MRI are important in surgical planning. We feel that the combined transmastoid and middle cranial fossa surgical approach provides the best exposure of the defect, and that the defect is best repaired with an autologous three-layer closure.
From the Section of Otolaryngology, the New Jersey Medical School, the University of Medicine and Dentistry of New Jersey, Newark (Dr. Patel, Dr. Kwartler, and Dr. Baredes), and the Department of Neurosurgery, Overlook Hospital, Summit, NJ (Dr. Hodosh).
Reprint requests: Jed A. Kwartler, MD, 55 Morris Ave., Ste. 304, Springfield, NJ 07081. Phone: (973) 379-3330; fax: (973) 379-3337; e-mail. Kwartlja@umdnj.edu
(1.) Jahrsdoerfer RA, Richtsmeier WJ, Cantrell RW. Spontaneous CSF otorrhea. Arch Otolaryngol 1981;107:257-61.
(2.) Wetmore SJ, Herrmann P, Fisch U. Spontaneous cerebrospinal fluid otorrhea. Am J Otol 1987;8:96-102.
(3.) Schuknecht HF. Mondini dysplasia: A clinical and pathological study. Ann Otol Rhinol Laryngol Suppl 1980;89:1-23.
(4.) Gacek RR, Leipzig B. Congenital cerebrospinal otorrhea. Ann Otol Rhinol Laryngol 1979;88:358-65.
(5.) May JS, Mikus JL, Matthews BL, Browne JD. Spontaneous cerebrospinal fluid otorrhea from defects of the temporal hone: A rare entity? Am J Otol 1995;16:765-71.
(6.) Pappas DG Jr., Hoffman RA, Cohen NL, Pappas DG Sr. Spontaneous temporal bone cerebrospinal fluid leak. Am J Otol 1992;13:534-9.
(7.) Escat E. Ecoulement spontane de liquide cephalorachidien par le conduit auditif externe, fistule congenitale probable. Arch IntLaryngol 1897;10:653-9.
(8.) Kline OR. Spontaneous cerebrospinal otorrhea. Arch Otolaryngol 1933;18:34-9.
(9.) Dysart BR. Spontaneous cerebrospinal otorrhea: Report of a case with successful surgical repair. Laryngoscope 1959;69:935-9.
(10.) Ferguson BJ, Wilkins RH, Hudson W, Farmer J Jr. Spontaneous CSF otorrhea from tegmen and posterior fossa defects. Laryngoscope 1986;96:635-44.
(11.) Gacek RR. Evaluation and management of temporal bone arachnoid granulations. Arch Otolaryngol Head Neck Surg 1992;l 18:327-32.
(12.) Kaufman B, Yonas H, White RJ, Miller CF. Acquired middle cranial fossa fistulas: Normal pressure and nontraumatic in origin. Neurosurgery 1979:5:466-72.
(13.) Rotilio A, Andrioli GC, Scanarini M, et al. Concurrent spontaneous CSF otorrhea and rhinorrhea. Eur Neurol 1982;21:77-83.
(14.) Kaufman B, Nulsen FE, Weiss MH, et al. Acquired spontaneous, nontraumatic normal-pressure cerebrospinal fluid fistulas originating from the middle fossa. Radiology 1977;122:379-87.
(15.) Jackson CG, Pappas DG Jr., Manolidis S, et al. Brain herniation into the middle ear and mastoid: Concepts in diagnosis and surgical management. Am J Otol 1997;18:198-205.
(16.) Ahren C, Thulin CA. Lethal intracranial complications following inflation of the extemal auditory canal in treatment of serous otitis media and due to defects in the petrous bone. Acta Otolaryngol (Stockh) 1965;60:407-21.
(17.) Gacek RR. Arachnoid granulation cerebrospinal fluid otorrhea. Ann Otol Rhinol Laryngol 1990;99:854-62.
(18.) Montgomery WW. Dural defects of the temporal bone. Am J Otol 1993;14:548-51.
(19.) lrjala K, Suonpaa J, Laurent B. Identification of CSF leakage by immunofixation. Arch Otolaryngol 1979;105:447-8.
(20.) Ryall RG, Peacock MK, Simpson DA. Usefulness of beta 2-transferrin assay in the detection of cerebrospinal fluid leaks following head injury. J Neurosurg 1992;77:737-9.
(21.) Skedros DO, Cass SP, Hirsch BE, Kelly RH. Beta-2 transferrin assay in clinical management of cerebral spinal fluid and perilymphatic fluid leaks. J Otolaryngol 1993;22:341-4.
(22.) Hubbard JL, McDonald TJ, Pearson BW, Laws ER Jr. Spontaneous cerebrospinal fluid rhinorrhea: Evolving concepts in diagnosis and surgical management based on the Mayo Clinic experience from 1970 through 1981. Neurosurgery 1985;16:314-21.
(23.) Langman AW, Chu FW, Lindeman RC. Spontaneous posterior fossa cerebrospinal fluid leak. Otolaryngol Head Neck Surg 1993;108:192-5.
(24.) Adkins WY, Osguthorpe JD. Mini-craniotomy for management of CSF otorrhea from tegmen defects. Laryngoscope 1983 ;93: 1038-40.
(25.) Lundy LB, Graham MD, Kartush JM, LaRouere MJ. Temporal bone encephalocele and cerebrospinal fluid leaks. Am J Otol 1996;17:46 1-9.
COPYRIGHT 2000 Medquest Communications, Inc.
COPYRIGHT 2000 Gale Group