Further Explanations for the Formation of Syringomyelia: Back to the Drawing Table
==================================================================================

* Mauricio Castillo

The etiology of nontraumatic spinal cord cysts remains unclear. Many explanations have been offered, though most have never been validated. Most of the proposed hypotheses, however, have something in common. They all invoke the possibility of abnormalities in the subarachnoid space that surrounds the spinal cord where a syrinx is present. Lesions such as extramedullary tumors, arachnoiditis, and vascular compromise have been found in conjunction with spinal cord cysts. In this issue of the *AJNR* (page 1785), the article by Brugières et al entitled *CSF Flow Measurement in Syringomyelia* provides some new insights regarding the abnormal flow of CSF, both inside and outside a syrinx.

The authors of another recent article proposed that the status of the spinal cord's central canal was ultimately responsible for the features of syrinxes (1). That is, a patent canal may lead to an extensive syrinx, a canal that is only segmentally patent may lead to a focal syrinx, and a nonpatent canal may lead to cord edema. Regardless of the end result (extensive or focal cysts), if the patients are symptomatic, then treatment is needed.

Brugières et al showed that greater alterations in intracystic fluid velocity are seen in large cysts. They found that the pericystic CSF velocity was not significantly different in small (and mostly asymptomatic) cysts than in large (and mostly symptomatic) cysts. This last observation is at odds with the currently favored explanation of abnormal CSF flow in the subarachnoid space as the factor leading to syrinx formation. In addition, Brugieres et al found a significant lack of association between CSF velocities (both inside and outside of the cysts) and factors such as cyst size and symptoms. The only factors that correlated were a higher diastolic cyst velocity with more severe symptoms. I have always entertained the idea that intracystic fluid motion is mostly turbulent and disorganized, and that this may result in expansion of some cysts. Brugières et al found that the intracystic fluid clearly has systolic and diastolic velocity peaks. This suggests a type of “organized and sequential” motion of the intracystic fluid. Although this pattern of motion mimics that of the CSF in the pericystic subarachnoid space, it may not be induced by it. This is reflected by the findings of Brugiéres et al, who identified an earlier peak systole inside the cyst rather than outside of it, which implies that intracystic fluid motion may be independent from fluid motion in the subarachnoid space. In addition, there were no differences in subarachnoid space fluid motion between patients with a syrinx and healthy control subjects.

We can probably assume that fluid pulsatility originates in the cord secondary to its blood flow (the same phenomenon drives CSF out of the lateral ventricles in the brain). Because a syrinx has already expanded the cord and narrowed the subarachnoid spaces, the pulsations will be transmitted first and preferentially to the cyst. Since water is noncompressible, the cyst then transmits its pulsations to the subarachnoid space. This may not happen under normal circumstances in which most of the spinal cord pulsations are absorbed by the fluid in the subarachnoid space. Based on these data, it seems possible that some spinal cord cysts are a primary spinal cord disorder, not originating as a consequence of an abnormal subarachnoid space.

How does fluid accumulate initially inside the cord? Fischbein et al (1) thought that the initial entrance of fluid into the cord reflected an abnormality of the subarachnoid space, driving CSF into the cord through the perivascular spaces. In view of the data shown by Brugières et al, I would like to reconsider that explanation. Some believe that the Virchow Robin spaces may contain a small amount of CSF, and that it is possible the CSF enters these spaces from the subarachnoid space. Subsequently, expansion of the cord (as a result of the normal, alternating arterial and venous blood filling) normally “milks” this fluid from the cord into the subarachnoid space. An abnormal subarachnoid space prevents this type of flow, leading to accumulation of CSF within the cord. This fact is, however, at odds with current evidence that the perivascular spaces of the CNS are sealed from the subarachnoid space (for example, they do not contain blood in cases of subarachnoid hemorrhage, regardless of their size) (2). If this is true, invoking abnormalities of the subarachnoid space as the etiology for syringomyelia in some cases is probably wrong. Again, these data can only lead to a proposal that at least some syrinxes are caused by intrinsic cord abnormalities, which at this time are not clear. The fact that diastolic velocities are higher in symptomatic large cysts is also important. Peak systole is transient, but the pressure exerted during diastole is omnipresent. Thus, it is not surprising that larger and more symptomatic cysts have a higher diastolic velocity.

Consider what happened to Brugières et al's patients after surgery. After decompression of remote lesions, such as a Chiari type 1 malformation, the intracystic fluid velocities (including diastolic velocity) decreased. After syrinx decompression, the systolic peak was first observed in the subarachnoid space and then within the cord (a reversal to normal). Again, it seems as if the status of the spinal cord determined the neighboring fluid dynamics and not vice versa.

I am not discounting the possibility that some syrinxes are due to an abnormal subarachnoid space. I have seen syrinxes associated with concurrent or remote spinal canal masses, following meningitis, and as sequelae of surgery. In many of these situations, an abnormal subarachnoid space was probably the fundamental etiology. As for many other disease processes, the etiology of syringomyelia is probably multifactorial. The importance of the article by Brugières et al lies not in showing abnormal fluid flow in the cysts and surrounding spaces, or in showing a reversal of these abnormalities after therapy. The importance of this article is that it forces us back to the “drawing table” to try to come up with other explanations for the formation of the syringomyelia.

## References

1.  Fischbein NJ, Dillon WP, Cobbs C, Weinstein PR. **The “presyrinx” state: a reversible myelopathic condition that may precede syringomyelia.** AJNR Am J Neuroradiol 1999;20:7-20
    
    [Abstract/FREE Full Text](http://www.ajnr.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiYWpuciI7czo1OiJyZXNpZCI7czo2OiIyMC8xLzciO3M6NDoiYXRvbSI7czoyMToiL2FqbnIvMjEvMTAvMTc3OC5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

2.  Gray F, Nordmann P. **Bacterial infections.** In: Graham DI, Lantos PL, eds. *Greenfield's Neuropathology* 6th ed. London: Arnold; 1997;114 
    
    

*   Copyright © American Society of Neuroradiology