The Resolution of Syringohydromyelia Without Hindbrain Herniation After Posterior Fossa Decompression

Bermans J. Iskandar, M.D., Gary L. Hedlund, D.O., Paul A. Grabb, M.D., W. Jerry Oakes, M.D., Division of Pediatric Neurosurgery and Department of Pediatric Imaging, Children's Hospital, Birmingham, Alabama

Neurosurg Focus. 2000;8(3) 

In This Article


In this report, we describe five patients with presumably idiopathic syringohydromyelia who improved after a craniocervical decompression. The improvement was documented both clinically and radiographically, and there was a dramatic reduction in the size of the syrinx in all patients. Although the average follow-up period was relatively short (a little more than 1 year), the initial data are encouraging.

We infer from these results that, on occasion, the disturbance of CSF flow at the foramen magnum does not have to be associated with downward herniation of the cerebellar tonsils. Such a disturbance in fluid dynamics may be explained in one of our patients by the presence of a veil at the fourth ventricular outlet, and by arachnoidal adhesions in two other patients. In fact, within seconds of surgical lysis of these adhesions and veil, CSF flow from the fourth ventricle into the subarachnoid space seemed to improve. Of further importance is the observation that even though the tonsils were not located below the foramen magnum on the preoperative MR images, intraoperative exploration revealed crowding of posterior fossa structures at the level of the foramen magnum in two patients (Cases 2 and 3). Both of these patients underwent intradural decompression, and one of them underwent shrinking of the tonsil by means of electrocautery. These findings were coupled with evidence of flow compromise on preoperative cine MR imaging (Fig. 3), with improvement in flow postoperatively. However, although all four of the preoperative flow studies were found to be abnormal when evaluated retrospectively, the prospective readings of two of these studies did not document the abnormality. It was only when these studies were compared with the postoperative ones, which had all shown improvement, that the disturbance became suspect. However, of more significance than the patients with crowding of the posterior fossa contents are those with very adequate CSF ventral to the brainstem. These patients would seem to be the least likely to improve after posterior fossa decompression, yet they fared as well as the others.

Figure 3. Sagittal view of the craniocervical junction obtained by using the 2-D phase-contrast cine MR technique. Note the compromised CSF flow dorsal to the cerebellar tonsils.

In an older study of syringohydromyelia, Newton[10] describes the presence of fourth ventricular drainage anomalies. In some cases, hindbrain herniation was also absent; however, because such cases were investigated in the pre-CT/MR imaging era, radiological studies were limited to myelography and air ventriculography. Surgical treatment in these patients usually consisted of posterior fossa decompression in addition to syrinx drainage, with good results. In our series, draining the syrinx was deemed unnecessary.

The size (volume) of the posterior fossa is difficult to assess retrospectively because of the differences in slice acquisition on MR imaging among the patients, as well as between these patients and any standardized control group. A small posterior fossa could explain the phenomenon that we describe in some of the patients. However, that hypothesis cannot be confirmed with the available data. A prospective evaluation of such patients with standardized MR imaging protocols may help shed light on this issue in the future.

The use of cine MR imaging was not particularly helpful in this study. Although it has been attempted in some studies to compare the results of quantitative cine MR imaging among patients with normal brains and those with Chiari I malformations,[2,4] we have not found any studies in which the use of this technique was standardized in a large control group. This makes our cine MR imaging data difficult to interpret. For instance, the improvement in flow noted after surgery in two cases may reflect the normal expansion of the subarachnoid space postoperatively.

Several pathogenetic theories have been proposed to explain the occurrence of Chiari malformations and the development of syringohydromyelia. Williams'[14] modification of Gardner's theory[8,12] is probably the most appropriate when attempting to explain the findings in this study. Williams suggested that normally occurring valsalva maneuvers (coughing, straining, and sneezing) resulted in epidural venous congestion and intracranial as well as intraspinal pressure rises, causing fluid to flow both cranially and caudally. In the presence of craniocervical abnormalities, flow into the cranial compartment meets no resistance, whereas caudal flow may be delayed by hind-brain adhesions and outlet obstruction, thus creating a pressure differential between the cranial and spinal compartments. This pressure differential may last from a few seconds to several minutes and cause worsening hindbrain impaction. When the fluid cannot exit through the fourth ventricular outlet foramina because of this impaction or blockage, it may be rerouted into the central canal of the spinal cord, causing syringohydromyelia. In an attempt to give more credence to his theory, Williams measured the pressures in the caudal and spinal compartments before and after craniocervical decompression, showing equilibration of the pressures in the two compartments, which, in turn, correlated with clinical improvement. In our series, we were unable to detect a patent central canal rostral to the syrinx on MR imaging. In addition, two patients (Cases 1 and 3) demonstrated long skip areas between the foramen magnum and the rostral aspect of the syrinx (Figs. 1 and 3). These observations cast some doubt on the applicability of Williams' theory in this group of patients, and present the same controversies seen with the Chiari I-associated syringohydromyelia.[12] Authors who have attempted to resolve such controversies include Ball and Dayan,[3] Aboulker,[1] and Pillay, et al.[12] Ball and Dayan hypothesized that CSF under pressure may enter the spinal cord through Virchow-Robin spaces, and gave indirect evidence of their theory by demonstrating dilated vessels and perivascular spaces in spinal cord autopsy specimens. Albouker noted that a large amount of CSF is produced in the spinal cord. When associated with an obstruction at the level of the foramen magnum, such fluid accumulates, causing the formation of a syrinx. Finally, Pillay, et al., attempted to combine these hypotheses into one unified theory of syringohydromyelia. However, to this date no theory of formation of syringohydromyelia has accounted for all the features of such anomalies.

Although idiopathic cervical syringohydromyelia may respond to posterior fossa decompression, we caution that the presence of syringohydromyelia in patients with no Chiari I malformation is much more likely to be caused by serious diseases, such as spinal cord tumors. It is imperative that a thorough investigation for tumor, posttraumatic arachnoiditis, and spinal dysraphism be conducted prior to any operative intervention. Furthermore, we encourage surgeons to use confirmatory specialized studies such as cine MR imaging, which measure the flow of CSF at the craniocervical junction, keeping in mind that such studies have not yet been optimized and standardized in large numbers of normal patients.

Manuscript received August 4, 1997.

Accepted in final form March 24, 1998.

This manuscript was presented at the American Association of Neurological Surgeons/Congress of Neurological Surgeons Joint Section of Pediatric Neurosurgery, New Orleans, Louisiana in December 1997 under the title "The Chiari Zero Malformation."

This manuscript was previously published in the Journal of Neurosurgery, J Neurosurg 89:212-216, 1998.