Spontaneous Intracranial Hypotension Successfully Treated by Epidural Patching With Fibrin Glue


Address all correspondence to Dr. Yoshihisa Fujita, Department of Anesthesiology and Intensive Care Medicine, Kawasaki Medical School, 577 Matsushima, Kurashiki-city, Okayama 701-0192, Japan.


We report a case of spontaneous intracranial hypotension due to a cerebrospinal fluid leak at the C2 level, which was successfully treated by epidural fibrin glue patching. Epidural blood patching was performed twice, first with 6 mL of autologous blood and then with 10 mL, but the intracranial hypotension was unresponsive. Although successful treatment of postdural puncture headache and persistent leak after intrathecal catheterization by epidural patching with fibrin glue has been reported, fibrin glue has not been previously applied in spontaneous intracranial hypotension. Our observation suggests that epidural patching with fibrin glue should be considered in patients with spontaneous intracranial hypotension, if epidural blood patching fails to resolve the symptoms.

In spontaneous intracranial hypotension due to a cerebrospinal fluid (CSF) leak in which the site has been identified, targeted therapy is indicated when bed rest or medications do not relieve neurological symptoms.1 If targeted epidural patching with autologous blood is not effective, neurosurgical procedures are then considered.2,3 Recently, we evaluated and treated a patient with intracranial hypotension in whom epidural blood patching failed to resolve the symptoms, but success was achieved by epidural patching with fibrin glue.


A 38-year-old man was admitted for increasingly severe postural headache, dizziness, and nuchal soreness, which were relieved in the supine position. He had suffered from these symptoms for 10 months, and had received outpatient treatment with analgesics. During this time, bilateral chronic subdural fluid collections and diffuse pachymeningeal thickening had been diagnosed by gadolinium-enhanced magnetic resonance imaging (MRI) of the head. These conditions did not require surgical intervention, but he could not continue his job as a secretary because of the symptoms. In the last few months, he had spent much of his time in the recumbent position. Although he had no history of lumbar puncture or neurosurgical operations, a year and a half before, he had a traffic accident, possibly with neck hyperextension, which was treated medically.

On examination, a lumbar puncture showed a CSF opening pressure of 20 cm H2O. Cisternography performed after indium-111 radionuclide injection into the lumbar subarachnoid space disclosed early accumulation in the urinary bladder and direct evidence of a CSF leak at the upper cervical spinal level at 2 hours. Computed tomography (CT) myelography confirmed the CSF leak on the left side of the C2 level (Figure 1A). Although MRI studies no longer showed either the chronic subdural hematoma or the pachymeningeal thickening after the administration of gadolinium, the CSF leak was clearly visualized on a sagittal T2-weighted MRI image of the cervical spine (Figure 1B).

Figure 1.—.

A, Computed tomography myelography showing a cerebrospinal fluid (CSF) leak (arrows) into the epidural space and through the intervertebral foramen to the paravertebral space at the C2 level. B, Sagittal T2-weighted magnetic resonance imaging of the cervical spine showing a CSF leak (arrows) at the C1 and C2 levels and a CSF collection in the anterior epidural space.

He was treated with intravenous hydration and bed rest for a week, but his symptoms did not improve. We, thus, decided to perform targeted epidural blood patching. An epidural radiopaque, closed-end, multifocal catheter (Ref. 100/383/116, Portex, UK) was advanced to the C2 level using a 16-gauge Tuohy needle at the C5-6 interspinous space under fluoroscopic guidance (Figure 2). A proper epidural spread of contrast material near the suspected site of the CSF leak was confirmed. Epidural patching with 6 mL of autologous blood was subsequently performed through the epidural catheter, which was left in place. Although the patient remained at bed rest for 24 hours, his symptoms were not resolved. Furthermore, an abundant CSF leak was observed around the epidural catheter. A second epidural patching with 10 mL autologous blood performed on the third day after the catheter placement, followed by 24 hours of bed rest, was also ineffective.

Figure 2.—.

Anteroposterior cervical roentgenogram showing a radiopaque epidural catheter advanced to the left side of the C2 level (arrows) via C5-6.

Because fibrin glue is widely used to achieve watertight dural closure in neurosurgical and orthopedic operations and because successful treatment with fibrin glue injection has been reported in postdural puncture headache4 and in persistent CSF leaks during long-term intrathecal catheterization,5 we decided to perform epidural patching with fibrin glue (Beriplast P, Aventis Pharma) before undertaking neurosurgical repair of the leak. Fibrin glue consists of two solutions. Each of the two solutions was divided into two doses, and a total of 2.8 mL of fibrin glue was injected through the epidural catheter in an alternate fashion without premixing the solutions. The catheter was removed after flushing with 0.5 mL of normal saline. His neurological symptoms disappeared after 36 hours of bed rest. The absence of CSF leak at the C2 level was confirmed by MRI and CT myelography, which were performed 2 weeks later (Figure 3). Although slight staining by contrast medium was noticed in the epidural space of C5 by CT myelography (not shown), further examination was not performed. No symptoms were reported by the patient on outpatient follow-up at 1 month.

Figure 3.—.

Computed tomography myelography at the C2 level (A) and sagittal T2-weighted magnetic resonance imaging of the cervical spine (B) 2 weeks after epidural patching with fibrin glue showing no cerebrospinal fluid leak.


This case illustrates that epidural patching with fibrin glue may be a good alternative to autologous blood, when epidural blood patching does not result in resolution of the symptoms in patients with spontaneous intracranial hypotension. Intracranial hypotension, which is characterized by orthostatic headache, dizziness, tinnitus, and improvement of the symptoms in the supine position, is often encountered after spinal anesthesia. It has thus been referred to as postspinal headache. In most cases, symptoms resolve with bed rest and adequate hydration. Even if the headache persists, epidural patching with autologous blood is effective in virtually all patients with postspinal headaches.6

Conversely, spontaneous intracranial hypotension, which occurs without previous dural puncture or neurosurgical procedures, is relatively rare.1,7 Tears in the nerve root sleeves are thought to be responsible for the CSF leak in spontaneous intracranial hypotension. Our patient had a car accident 18 months before the onset of his symptoms, at which time his neck might have been hyperextended. The relation between the accident and the CSF leak at C2 in this patient remains speculative.

In our patient, the CSF opening pressure was not low, and pachymeningeal thickening was not demonstrated on the gadolinium-enhanced MRI performed after admission. We postulate that the long-term CSF leak resulted in a compensatory increase in CSF production. In addition, because he had avoided the head-up position due to postural headache for a prolonged period, the normal intracranial pressure might have been maintained in the recumbent position. We do not believe that the normal CSF opening pressure and MRI findings in this patient contradict the diagnosis of spontaneous intracranial hypotension. In accordance with our notion, Mokri et al described cases of spontaneous intracranial hypotension with normal CSF opening pressures8 and with absent pachymeningeal gadolinium enhancement on MRI,9 and emphasized that these findings do not rule out CSF leakage.

The efficacy of epidural patching with autologous blood in this condition is unknown, although successful treatment with epidural blood patching has been reported.2,10 In our patient, epidural patching with 6 mL and 10 mL of autologous blood resulted in no resolution of the symptoms. Moreover, the CSF continued to leak around the epidural catheter, indicating a persistent CSF leak into the epidural space. Encouraged by reports of the use of fibrin patch in postdural puncture headache,4,5 we applied it to our patient with spontaneous intracranial hypotension. To the best of our knowledge, there have been no previous reports of treatment with epidural fibrin glue for spontaneous intracranial hypotension.

Epidural blood patching is usually performed through a needle advanced into the epidural space. Thus, it may be noteworthy that we used a dwelling epidural catheter instead of a needle. Although the fibrin glue is a viscous fluid, it could be injected through the catheter without great difficulty. The catheter technique may have several advantages in that autologous blood or fibrin glue can be injected close to the site of the leak by adjusting the catheter tip, thus allowing a smaller injected volume, and blood patching can be repeated without the need for needle placement. Furthermore, accidental intrathecal injection of fibrin glue, the effects of which are not known, can be avoided with the catheter technique, when proper epidural spread of contrast material is fluoroscopically confirmed. This may be very important in patients with a CSF leak in the cervicothoracic region. In addition, the CSF leak around the catheter was suggestive of the ineffectiveness of epidural patching with autologous blood, as mentioned above.