Absent pachymeningeal gadolinium enhancement on cranial MRI despite symptomatic CSF leak

Case report.

Patient 1.

A 34-year-old woman developed severe throbbing orthostatic headaches and posterior neck pain and nausea in July 1997. Her neurologic examination was normal. Head MRI showed that the tips of the cerebellar tonsils were slightly below the foramen magnum (figure, A), without DPMGE (figure, B). Cervical spine MRI showed an extensive extra-arachnoid and extradural CSF collection extending through multiple neural foramina (figure, C). The site of the leak could not be identified. Four head and spine MRIs over the next 16 months showed no change. CT myelogram in November 1998 showed extradural contrast from C5 to L3, most prominent at C6 on the right. CSF opening pressure was 10 mm of water. Epidural blood patch at C6-7 markedly decreased the symptoms.

• Download figure
• Open in new tab
• Download powerpoint

Figure. Patient 1. (A) Sagittal MRI shows subtle evidence of brain sagging manifested by mild obliteration of suprasellar cistern, while the tips of cerebellar tonsils extend slightly below the posterior margin of the foramen magnum. (B) Gadolinium-enhanced coronal MRI is normal without evidence of abnormal pachymeningeal gadolinium enhancement or any subdural fluid collections. (C) MRI of cervical spine shows extra-arachnoid CSF collection within the cervical spinal canal, which appears to extend out through neural foramina (arrow).

Discussion.

Initially coined as “spontaneous aliquorrhea” by Schaltenbrand and as the cause of postural headaches,6 this entity later became known as spontaneous intracranial hypotension. On head MRI, it has several fairly characteristic MRI features, including DPMGE, subdural fluid collections, and imaging evidence of descent of the brain or “sagging brain.”3,4,7,8 MRI of the spine helps to identify not only the level of the leak but also the meningeal diverticulae and presence of extra-arachnoid/subdural or epidural and paraspinal fluid. Meningeal diverticulae are not rare, and detection of a diverticulum, even when large, may not necessarily indicate that the site of CSF leak has been identified (as brought out in Patient 4). Extravasation of CSF (or contrast) must be demonstrated. CT myelography is the best test for locating the site of the leak. Sometimes, when an initial study fails to show the site of the leak, a second CT myelogram after a few to several weeks may reveal it.

A broader clinical and imaging spectrum of CSF leaks is emerging. Sometimes the headache may lose much of its orthostatic features, evolving into a lingering chronic daily headache with fluctuating intensity.3,9 Some of the patients have CSF pressures that are consistently within normal limits.5 Our Patient 4 had an opening CSF pressure of 120 mm of water. After the initial blood patch, she partially improved, but remained symptomatic. She had documented CSF leak while her CSF opening pressure had increased to a higher normal level at 170 mm of water. The independent factor for the development and maintenance of symptoms appears to be CSF hypovolemia, which leads to the descent of the brain, traction on the pain-sensitive anchoring structures, and development of headaches. It is likely that if the MRI was obtained in the upright position, the descent of the brain would be detected more frequently and more prominently.

A decrease in CSF volume is compensated by intracranial venous hypervolemia and subdural fluid collections. The pachymeningeal venous engorgement is responsible for the DPMGE, which is reversed when the CSF volume is entirely or partially restored. In our first two patients, DPMGE never developed. It is probable that the loss of volume and hydrostatic pressure changes were not quantitatively sufficient to result in venous congestion significant enough to lead to the appearance of DPMGE.9 The next three patients initially had DPMGE, which subsequently resolved while they were still symptomatic with a CSF leak, albeit to a lesser degree. Our last patient initially had normal meninges, but DPMGE developed later.

A head MRI with normal meninges does not exclude the possibility of CSF leak. With suggestive clinical manifestations, additional diagnostic studies should be pursued. MRI of the spine may show not only the presence but also sometimes the site of a CSF leak, even with normal head MRI.