The changes of syrinx volume after posterior reduction and fixation of basilar invagination and atlantoaxial dislocation with syringomyelia

Although many theories have been proposed [11‐15], the pathogenesis of syringomyelia is still unclear. It is generally accepted that syringomyelia is caused by disturbances of normal cerebrospinal fluid (CSF) flow circulation [7, 11‐13, 16, 17]. Several studies have revealed that BI/AAD are associated with syringomyelia [4‐7, 12], but there is a paucity of reports focusing on mechanism of BI/AAD combined with syringomyelia. This study found that there are three factors which may be the reasons for syringomyelia in patients with BI/AAD: (1) ventral compression by dislocated odontoid process. (2) Dorsal compression by Chiari I malformation. (3) Concomitant Klippel–Feil syndrome and occipitalization of atlas. Concrete analysis is as follows.

In Menezes’s study [6], 84 patients who had CM-I with CVJ abnormalities and syringomyelia underwent ventral decompression. The ventral decompression procedure was followed a week later by the dorsal decompression and fusion. During this time, interval postoperative MR imaging was performed and showed early improvement in the syrinx. A ventral decompression of the CVJ had been shown to allow for regression of the syrinx. In this series, evaluation of the radiological findings of 71 patients with BI/AAD, the dislocated odontoid process directly compressed the brainstem which decreased the dimension of the FM and obstructed CSF circulation. We speculate that ventral compression of the spinal cord caused by dislocated odontoid process is the primary reason of syringomyelia in patients with BI/AAD. Among 71 cases, 66 (93.0 %) cases experienced satisfied restoration of AAD/BI, and 65/66 (98.5 %) cases had syrinx shrinkage. It suggested that the relief of ventral compression was the main reason of regression of syrinx.

The pathogenesis of Chiari malformation (CM) with or without associated BI and/or syringomyelia is very complex. One of the current concepts is that there is a reduction in posterior fossa volume which obstructs the free communication of CSF flow from the cranial to the spinal subarachnoid space [5, 6, 18]. Posterior fossa decompression with or without duraplasty as treatment for symptomatic CM-I with syringomyelia has shown satisfactory results except in the presence of ventral bone abnormalities [12, 13, 16]. In Klekamp’s study [19], craniospinal stabilizations with the foramen magnum decompressions were performed in 31 patients with BI and CM-I and ventral compression, nine patients underwent transoral decompressions, and a favorable clinical result was achieved. Their studies indicated that ventral and dorsal compressions were commonly associated in patients with BI and CM-I, the ventral decompression and stabilization were very important. Recently, Goel [20] pointed out that the pathogenesis of CM with or without associated basilar invagination and/or syringomyelia is primarily related to atlantoaxial instability. His theory needed more prospective studies and research to be proved. Previous literatures show that Syringomyelia occurs in conjunction with CM in 50–70 % of patients [12, 21]. The rate of CM combining with BI/AAD/syringomyelia is still unclear. In this series, 59 (83.1 %) cases with BI/AAD/syringomyelia had concomitant CM-I. We speculate the dislocated odontoid shifts posteriorly, which compresses the CVJ and decreases the dimension of the FM. The concomitant hindbrain herniation makes the overcrowded FM worse and interrupts the CSF circulation.

Congenital fusion of C2 and C3 is most commonly seen in KFS [22]. Atlas occipitalization and congenital C2–3 fusion often lead to AAD and BI, eventually resulting in hypermobility or symptomatic stenosis of the CVJ [22]. These patients often have cervical myelopathy or concomitant anomalies such as BI, CM and syringomyelia [7, 23]. Patients with AAD and KFS had a significantly higher incidence of assimilated posterior arch of atlas compared to the patients with AAD as there was an increase in the Ishihara curvature index [24]. In our series, 68 (95.8 %) cases were combined with assimilation of atlas and 29 (40.8 %) cases were combined with KFS. The patients with assimilation of atlas and KFS frequently had cervical lordosis leading to odontoid process and greater inclination towards the spinal cord, and the normal CSF circulations were obstructed.

Various surgical treatments of BI/AAD have been reported [1, 2, 4‐6]. Preoperative cervical traction had been introduced in patients with BI/AAD to assess neurologic status and reducibility [4]. Several reports have detailed successful application of transoral approaches for BI/AAD [5, 6]. However, there continues to be several controversies regarding the optimal indications and approaches for anterior decompression [25]. In recent years, posterior approaches have become the primary methods for the treatment of BI/AAD [1, 2, 7, 8]. The direct posterior distraction technique seems to be an effective method.

In our series, a direct posterior reduction and fixation approach was used for all patients. Four pediatric patients between 15 and 17 years old received the same surgical treatment as adults. Pre- and intraoperative cervical tractions were never used. An AAD that is “irreducible” under cervical traction can become reducible by a direct intraoperative distraction technique. Only two patients had further transoral decompressions because of unsolved ventral compressions and no clinically improvements. The technique decompresses the spinal cord and medulla oblongata and expands the volume of FM, which alleviates the obstruction of CSF circulation and shrinks the syrinx. In this series, the direct posterior distraction technique showed a satisfactory result.

As a treatment for symptomatic CM with syringomyelia, posterior fossa decompression (PFD) has shown satisfactory results [12, 13, 16], but PFD for the treatment of CM with BI is controversial [7, 20, 26]. Some studies suggested that the surgical treatment of CM with BI and syringomyelia should be directed toward atlantoaxial stabilization and PFD is not necessary [7, 20]. Klekamp suggested that posterior fusion should be performed along with the foramen magnum decompression in the patients of CM with BI, regardless of pre- or intraoperative signs of instability [19]. A significant controversy also exists concerning whether the posterior decompression approach should be intradural or extradural [12, 13, 16, 19, 27]. In our series, all patients underwent resection of a small part of the posterior margin of FM. We consider that reduction and fusion is necessary, and PFD is a supplement of the expansion in posterior fossa volume. When the reduction is unsatisfied or tonsillar herniation is severe, PFD will play a partial role. Duroplasty was not used in our series; a good clinical efficacy was achieved and the complications correlated with duroplasty were avoided.

Robert [14] performed a retrospective study of patients who developed worsening syringomyelia after Chiari decompression was performed. Each case in the study had a blockage of CSF pathways at the FM and most cases were found to have intradural scarring or an arachnoid web. Reports of worsening syringomyelia in the patients with BI/AAD after surgery were absent. In our series, postoperative follow-up in 71 cases with syringomyelia showed that 5 cases had no image changes of the syrinx. All of the five patients did not improve clinically. This phenomenon signified that the blockage of CSF flow at the CVJ was still unresolved after the surgery. The possible reason of why there was no change in syrinx is summarized as follows.

The primary reason is due to an unsatisfied reduction of the dislocated odontoid process. Under such circumstances, although the screw fixation and fusion improve stability of CVJ, the dislocated odontoid process still directly compressed the brainstem and obstructed CSF circulation (Fig. 5). In the five cases with unsatisfied reduction of the BI/AAD, only 1 (20 %) syrinx shrank.

Another reason is severe tonsillar herniation. Usually, if AAD restoration is satisfied, the concomitant syringomyelia will regress. However, when ventral compression is persistent, the unchanged severe tonsillar herniation will aggravate obstruction of CSF circulation. In this series, a patient with no image change of the syrinx had severe tonsillar herniation. The tonsillar tonsil herniated downward to C2 vertebral body level. After the surgery, ventral compression was not resolved completely and the herniated tonsillar aggravated the crowding of FM. Under such circumstances, the two-stage ventral decompression and/or tonsillectomy with duraplasty may be required.

Local excessive kyphosis is the third reason. The CCA can be used to quantify the amount of BI; ventral spinal cord compression may occur with angles less than 150° [9]. In the patients with BI and AAD, superior migration of the odontoid and/or the horizontal clivus can lead to direct compression of the brainstem and upper cervical spinal cord. These changes place the cervicomedullary junction in excessive kyphosis and decrease the CCA [9, 28]. In the five cases which had no image change of the syrinx of this series, the pre- and postoperative CAA was 120.0° and 118.9°. Although it was not statistically significant, it implied more severe local kyphosis after the surgery. In the case of a patient with severe local kyphosis, the pre- and postoperative CAA was measured to be 98° and 95.3°.

Although the patient’s postoperative odontoid apex was below Wackenheim line, anterior compression by local kyphosis was unresolved. The postoperative MR image revealed no syrinx change.

The pathology of syringomyelia combined with CVJ anomalies is only partially understood and its treatment is still a matter of debate; especially involving CM-I. This study will help with the recognition and awareness of this complicated pathology. The results suggested that the cranial vertebral realignment and internal fixation is the more appropriate technique in effectively managing this subgroup of patients. However, several limitations of this study still exist. Though 86 patients were matched with the inclusion criteria, only 71 could be documented in this series. The shortest follow-up period was 6 months which may not be long enough for this complex disease. After 2012, the distraction technique was modified. Thus, the effect of surgery had a chance to be affected by the use of different surgical methods.