Advances in the medical and surgical treatment of spinal cord injuries and patients with spinal cord tumors and other spinal lesions have led to better outcomes and longer survivals. This good news has been accompanied by the bad news that an increasing number of these patients present to their physicians or health care providers with signs and symptoms of progressive neurologic dysfunction and increasing pain and disability. Differential diagnosis in these unfortunate patients must include progressive spinal column deformity or instability as well as new primary or metastatic spinal cord or column tumors or recurrence of previously treated lesions. Postradiation myelitis is another possibility. Transverse myelitis from Guillain Barré syndrome or other various inflammatory and metabolic causes of myelopathies must be ruled out. Unusual infections such as tuberculosis or fungal infections are increasingly seen in our increasingly large immunosuppressed population (e.g., individuals suffering from cancer, diabetes, AIDS). Iatrogenic causes include arachnoiditis associated with intraoperative dural leaks or intradural procedures or subarachnoid hemorrhages from aneurysms or vascular malformations or from inadvertent penetration of the dura during epidural blocks for pain relief or anesthetic administration, including during childbirth. As physicians, we also must be increasingly aware of the fact that patients who present to us are not protected from other diseases just because of their spinal cord injury or other spinal disorder. These patients may have a second or even third disease. It is not uncommon for patients with diabetes to have a peripheral neuropathy or a spinal cord infarct from small vessel disease. In the geriatric population we see an increasing number of patients with spinal cord injury or surgery who develop Parkinson’s disease, amyotrophic lateral sclerosis, or Alzheimer’s disease. These central nervous system disorders often become clinically exacerbated by the stress of the spinal injury or surgery. In the younger population, more commonly female, multiple sclerosis can mimic many of the signs and symptoms of progressive neurologic disability or dysfunction, and even pain, seen in spinal cord-injured or diseased patients. Congenital disorders such as Chiari I malformation, with or without syringomyelia, or tethered cord syndrome, with or without lipoma, may also present with a similar group of signs and symptoms. Many of these patients are treated with the wrong surgery because in addition to having an arteriovenous fistula, they may have cervical lumbar stenosis and present with dragging their leg. It is not uncommon to see patients who are surgically treated for abnormalities on their images but whose signs or symptoms have nothing to do with those findings.
Therefore, when someone with a spinal cord or spinal column injury presents with a clinical picture of progressive neurologic dysfunction, with or without pain, one must rule out the more common causes of their disabling signs and symptoms. Once this is done, progressive post-traumatic cystic myelopathy (PPCM) 6 or progressive post-traumatic myelomalacic myelopathy (PPMM), 6 both associated with what is sometimes called the tethered cord syndrome, must be highly suspect. These patients may present with the following group of signs or symptoms that may be unilateral or bilateral. They may present with one or more of these signs or symptoms, and any of them may be positionally related. In order of frequency, these are as follows: 1) pain local and/or radicular (this is in contrast to the distal deafferent pain suffered by victims of spinal cord injury or disease that is diffuse and often burning and dysesthetic; this local or radicular pain is more often associated with a tethered cord syndrome or can, for example, be seen in cases with progressive spinal deformity or instability), 2) increasing loss of motor function, 3) increasing loss of sensory function, 4) increasing spasticity, 5) autonomic dysreflexia, 6) hyperhydrosis, 7) increasing sphincter dysfunction, 8) increasing respiratory insufficiency, and 9) Horner’s syndrome. The clinician working up these patients must think of, and rule out, all of the possibilities listed above before focusing on the diagnosis of the tethered cord syndrome. For example, a urinary tract infection, which has occurred at least once in every victim of paralysis, can be associated with autonomic dysreflexia, hyperhydrosis, and increasing spasticity. Once these things are ruled out, it is time to proceed to an imaging workup to include radiograph of the spine with motion views and high resolution MRI, and if indicated, cine-MRI.
With the advent in MRI techniques, the syndrome of progressive post-traumatic myelopathy 6 has become an increasingly recognized entity, with a reported prevalence of 0.3–3.2%. 1,2,19–21,23,28 Etiologies include arachnoiditis, 16,19,26 spinal instability with cord compression, 17,19 spinal cord tethering, 17,26,27 and microcystic spinal cord degeneration or gliosis. 8,11 CT scan after intrathecal contrast injection, MRI, cine-MRI techniques, and intraoperative ultrasound 29 has made possible a better ability to identify and treat patients whose myelopathy could be attributed to post-traumatic syringomyelia or to post-traumatic myelomalacia. 5,8,18,19,28 Those with large, confluent cysts are classified under the syndrome of PPCM 6 or post-traumatic syringomyelia. Patients without demonstrable large cysts but with evidence of microcysts or myelomalacia and cord tethering are grouped under the syndrome of PPMM. 6 Tethering and cyst formation are frequently observed together. The treatment of post-traumatic syringomyelia without tethering consisted primarily of shunting of the syrinx. 3,4,21,25,30,32 Recent reports of syrinx treated by cyst fenestration and an expansile duraplasty have been encouraging. 13,24,25 Surgical treatment of PPMM is based on the release of the tethered cord, which could be achieved by transection of the cord at the tethered site 3 (a technique not recommended by these authors) or by the careful lysis of the adhesions. 11
The clinical outcome of both medical and surgical management of spinal cord cysts, however, has historically been disappointing. Despite shunting of the cyst to the pleural, peritoneal, or subarachnoid space, clinical improvement has not been long-lasting, and revision shunt surgery is rather frequent. 4,24,32,33 A shunt failure rate of 50% as well as high rates of revision surgeries have been reported. 4 Shunt placement is further associated with fibrotic reaction and arachnoiditis. 4 The authors present their experience of treating syringomyelia patients with the strategies outlined in Methods, in hope of developing a surgical treatment algorithm and improving surgical outcome.
Between July 1997 and July 2000, 53 patients with symptomatic progressive post-traumatic syringomyelia were diagnosed and underwent surgical treatment by the senior author (B.A.G.) at the Jackson Memorial/University of Miami Medical Center. The following inclusion criteria were used: (1) history of spinal cord injury or surgery, (2) evidence of progressive neurologic deficit, and/or pain syndrome, and (3) MRI demonstrating a spinal cord cyst with or without the suggestion of tethering (Figures 1A–C, 2A–D). Patients with only spinal cord tethering and myelomalacia (PPMM) were excluded from this review. All patients except one had a preoperative MRI with and without gadolinium injection. One patient underwent preoperative and postoperative CT myelogram because of previous surgical metal implants. A retrospective study was carried out to characterize these patients and to determine their surgical outcome.
The surgical procedure started with laminectomies to expose the dura. Intraoperative ultrasonography with a 7.5-MHz transducer was then performed to localize the cyst and site of tethering, if any (Figures 3and 4). A midline dural opening was made if possible. In cases of dense midline dorsal adhesions, an eccentric dural opening was made to avoid cord parenchymal damage. By leaving the adherent dura attached to the cord, the risk of further tissue injury was eliminated. The following treatment algorithm was used (Figure 5). For solitary cysts without focal tethering, a syringosubarachnoid shunt (stent) was placed with a small midline myelotomy (Figure 6). The only cases in which we still divert the spinal fluid flow or shunt outside of the subarachnoid space is in the rare patient with diffuse arachnoiditis and no possibility of re-establishing normal cerebrospinal fluid (CSF) flow dynamics even with untethering and duraplasty. In those cases we have used a shunt of the spinal cord cyst to the pleural or peritoneal space to palliate their symptoms rather than eliminate the cause of their problem, as in the case of the untethering and expansile duraplasty. For tethered spinal cords, release of adhesions was achieved with microdissection and with careful neurophysiological monitoring (sensory evoked potential, electromyography, and more recently, motor evoked potentials). The ultrasound image was then repeated to assess the adequate collapse of the cyst (defined as >50% collapse immediately) (Figure 7). If adequate untethering could not be achieved or if a persistent cyst was observed on the repeat ultrasound images, a short syringosubarachnoid shunt (or stent) was inserted. We fashion our stent using a Spetzler (Integra Lifesciences, Plainsboro, NJ) Silastic lumbar–peritoneal microcatheter passed into the cord for the length of the cyst following a small 2-mm midline myelotomy made with 11 blade. We then have approximately 1 cm of catheter protruding outside the cord into the newly created subarachnoid–subdural space under the circus tent-like dural allograft. We secure the tubing to the pia mater with a 6–0 Prolene suture. A piece of freeze-dried cadaveric dura is used to expand the subarachnoid space, after tacking up the dural edges to the paraspinal muscles. A water-tight dural closure can be accomplished using 5–0 Prolene running sutures placed close together and confirmed water-tight with repeated Valsalva maneuvers.
After surgery the patients were kept at bedrest for 48 hours and log rolled every 2 hours side to side and resting in between turns in a semiprone to semiprone position to prevent possible positional retethering. Clinical follow-up (combination of medical record review and phone interviews) of an average of 23.9 months was obtained (range 12–102 months). Significant improvement or resolution of one or more of the major presenting symptoms or signs was considered a good result. A postoperative MRI was obtained from 2 months to 5 years after surgery on all patients except five.
A total of 53 patients underwent a surgical procedure for the symptomatic syrinx. Eight patients were lost to follow-up by 1 year after surgery (three each from the shunt only and untethering only groups and two from the untethering shunt group), leaving 45 patients with at least 1 year of clinical follow-up (range 12–84 months, mean 22.5 months). Of these 45 patients, 17 patients underwent a shunting procedure alone. Nineteen patients underwent an untethering procedure only. Nine patients underwent concurrent untethering and cyst shunting, as the cyst did not collapse >50% after the untethering portion of the procedure. The mean age of the 30 male and 15 female patients was 45.6 years (range 21–72 years). The etiologic factors leading to syringomyelia are detailed in Table 1. Blunt trauma was the most common cause (40 of 45 cases) with motor vehicle accidents accounting for 34. The other five cases encompassed four previously operated intramedullary mass lesions (two ependymomas, one cavernous angioma, and one arteriovenous malformation) and one with gunshot wounds. The level of the previous injury was listed in Table 2.
The neurologic deterioration and/or pain syndrome developed an average of 11 years after the initial trauma (range 0.4–41 years). Eighty percent of the patients presented with a pain syndrome (axial or radicular in nature), 60% with motor function level deterioration, and 58% with increased spasticity. Sensory function deterioration and progressive paresthesia were noted in 49% and 42% of the patients, respectively (Table 3). Other signs and symptoms occurred in <40% of the patients. There was no statistically significant difference in terms of clinical presentation between patients with or without cord tethering (χ2 analysis).
A total of 45 patients were observed for a mean of 22.5 months (range 12–84 months); eight patients were lost to follow-up before the minimum 12-month follow-up period. Thirty-three patients (73%) had satisfactory results (resolution of one or more of the presenting signs/symptoms without deterioration) after surgical treatment, and nine remained unchanged. Three patients (7%) had worsening motor functions, and one of these three patients also had increased spasticity. Surgical intervention improved the motor symptoms in 56% (15 of 27), and spasticity in 46% (12 of 26) of the patients. Sensory loss was halted and improved in 45% (10 of 22) and gait disturbance improved in 47% (8 of 17) of the patients. Axial and/or radicular pain syndrome was improved in 36% (13 of 36). Minimal symptomatic relief resulted from the operation in terms of paresthesia, sphincter dysfunction, or autonomic dysreflexia. The various operative procedures did not show a difference in terms of improvement of a particular sign/symptom (P > 0.05, χ2 analysis).
Preoperative MRI in 44 patients revealed either a confluent spinal cord cyst or a large multiseptated cyst in all patients. One patient underwent only a preoperative CT myelogram, which demonstrated a syrinx on delayed CT images. Twenty-eight of 45 patients were shown to have cord tethering in addition to the syrinx on the preoperative imaging. Of the 19 patients undergoing untethering only, 16 patients underwent postoperative imaging. The cyst was decompressed in 15 of 16 patients (94%) (Figures 8 and 9). Of the 17 patients undergoing cyst shunting only, 15 patients underwent a postoperative MRI scan, demonstrating initial cyst reduction in 14 patients. However, symptomatic deterioration in three patients prompted further radiographic investigation, which demonstrated recurrent cyst in all three patients. One of the three patients was shown to have cord tethering, in addition to the recurrent syrinx. For the nine patients who underwent both an untethering and shunt placement, the cyst size was reduced in eight of nine patients, although four patients still have visible cyst formation on follow-up MRI examination.
One treatment failure (5%) and two complications (11%) occurred as the result of surgical untethering. During the follow-up period one patient in the untethering only group experienced symptomatic recurrence. Initial 3-month postoperative MRI demonstrated cyst decompression. The MRI obtained 1 year after surgery showed minimal enhancement on the cranial portion of the cyst as well as cyst reaccumulation. The cyst was shunted, with some improvement of her symptoms. One patient experienced worsened motor function in her upper extremities associated with increased spasm. Her MRI examination showed some myelomalacia but no cyst reaccumulation or tethering. One patient developed motor weakness in one upper extremity that partially improved with time. Another patient developed a CSF leak with pseudomeningocele requiring a course of lumbar CSF drainage.
One complication (6%) and three treatment failures (18%) occurred in the shunt only group. Two patients from the shunt only group needed a shunt revision. One of the two patients needed two additional shunt revisions over a 24-month period. One patient from this group developed ventral cord tethering and worsening motor function but declined further surgical intervention.
Three of the nine patients (33%) who underwent both an untethering and shunting procedure had clinical recurrence and MRI evidence of ventral cord retethering. Six of nine patients developed transient weakness for 3–6 months after surgery, although five recovered motor strength to near baseline. One complication (11%) occurred in one patient who developed a CSF leak with pseudomeningocele requiring CSF drainage and eventually surgical re-exploration and repair.
A feature that is commonly associated with both post-traumatic cystic and noncystic myelomalacic myelopathy is tethering of the spinal cord. 9,11,14,20,27,28 MRI showed that in 27 of our 45 cases of syringomyelia, the spinal cord was locally tethered by scar tissue. This tethering and scar formation of the spinal cord may play an important role in the pathophysiology after trauma or surgery. Scarring and tethering along with local hemorrhagic environment, recumbent position of the patient, chronic ischemia, and changes in local CSF flow dynamics are likely interrelated factors contributing to changes in the cord parenchyma. Valsalva maneuvers create a tissue-shearing force associated with pain and/or progressive neurologic dysfunction. Post-traumatic syringomyelia and post-traumatic myelomalacia (tethered cord syndrome) may be a continuum of the same pathologic process. 28,33 Another suggested mechanism for the myelopathy is traction of the tethered cord leading to mechanical distortion and ischemia. 2,3,12,14,21 It has been demonstrated that movement of the cervical spinal cord is considerable during flexion–extension neck movements and that, with adhesions, the segments rostral to the point of fixation undergo relatively more stretching. 17
A treatment aiming at altering the pathophysiology of cyst formation should be more efficacious than one that simply drains the cyst. 13,22,24,25,33 This is evident in the many reported long-term follow-up series of shunted spinal cord cysts, which generally only had a 50% shunt patency rate. 4,24,25 Clinical improvement has been minimal, although some authors have reported significant improvement with motor dysfunction or dysesthetic pain after shunting of the syrinx. 15,30 A cyst shunting procedure also tends to increase the incidence of arachnoiditis and cord tethering. 4,32 Previous reports have described the use of lysis of adhesions, cyst fenestration, and duraplasty, without the use of a shunt, as the successful treatment for a syrinx. 13 The authors believe that adequate release of the adhesions causing cord traction and CSF flow blockage may obviate the need for cyst shunting in a significant number of cases. In this retrospective review the majority of patients who developed progressive post-traumatic syringomyelia associated with cord tethering benefited from the surgical release of the adhesions. Many patients with syringomyelia treated with cyst shunting have frequent treatment failures. 1,24,33 The possible explanations include residual tethering or retethering with alteration of normal CSF flow dynamics. Scar tissue may also block a shunt tube at either end (i.e., in the cord cyst or in the subarachnoid or pleural or peritoneal space). The nine patients in this series undergoing concurrent untethering and shunting had a high rate of persistent–recurrent cyst formation. In general, a longer segment of tethering, or existence of extensive ventral tethering, was observed in these patients. Accordingly, the clinical outcome is slightly less favorable for this population.
Finally, there are several issues of strategic importance. The first is prevention. Can spinal column realignment and spinal cord decompression and early mobilization of acute spinal cord-injured patients help avoid tethering of the spinal cord and the resultant in formation of a syrinx? We use the Rotorest (Kinetic Concepts, San Antonio, TX) treatment table to mobilize our patients from the conventional supine position immediately after injury or surgery. This continuous turning or kinetic 10 therapy is applied until the patient can be mobilized to sitting in a wheelchair or standing. For 30 days postoperation or postinjury, patients are constantly rotated on the Rotorest or turned every 2 hours from side to side while maintaining a semiprone position to prevent positional retethering. Another issue of strategic importance might be the routine untethering of the spinal cord and performing a duraplasty after intradural surgery for tumor, arterial venous malformation or fistula, or arachnoid cyst, etc. Also, the use of a physical barrier interposed between the cord and meninges to avoid retethering may be a reasonable strategy. These are issues that we are looking at both clinically and in the laboratory in a prospective fashion. We are presently evaluating Surgicel (Johnson & Johnson, New Brunswick, NJ) as one such potential material. Others have advocated for Gortex (W.L. Gore, Flagstaff, AZ) or other synthetic graft materials as being “nonstick.” Biotechnology being developed includes creation of molecules that could modulate scar formation, i.e., prevent intradural adhesions without compromising dural and wound healing. Medications (both systemic and topical) and various biomaterials to serve as suitable graft materials and as sealants to prevent CSF leak are being evaluated and tested. Certain hypotheses need to be tested clinically, such as: what is the value of decompression and realignment to avoid postinjury or postsurgical tethering not only by scar, but also by spinal canal compression or deformity? In our experience, failure of surgical untethering has been more commonly observed in patients with spinal column deformity, especially kyphosis. Technical advances are also necessary to allow for safe and effective anterior untethering procedures, especially for difficult to approach cases with ventral cord adhesions. Residual ventral spinal cord tethering seems to be associated with a significant number of the patients who failed with what we considered our “ideal” surgical treatment protocol. More definitive diagnosis and understanding of the pathophysiology of post-traumatic syringomyelia can be, in large part, attributed to the sophistication of MRI and cine-MRI technology. This technology has brought this field a long way in the last 20 years and promises to open up even more therapeutic possibilities in the near future. 7,31,34,35
- Long-term outcome of cyst shunting has historically been discouraging.
- The underlying pathophysiology of PPCM includes cord tethering and alteration of normal CSF flow dynamics.
- Intraoperative ultrasonography can detect subtle cord tethering, monitor and guide surgical management, and evaluate the cyst size after untethering of the spinal cord.
- A more logical approach to post-traumatic cystic myelopathy may lead to more favorable long-term outcomes.
1. Asano M, Fujiwara K, Yonenobu K, et al. Post-traumatic syringomyelia. Spine 1996; 12: 1446–53.
2. Barnett HJM, Jousse AT. Syringomyelia as a late sequel to traumatic paraplegia and quadriplegia. In Barnett HMJ, Foster JB, Hudgson P, eds. Syringomyelia: Major Problems in Neurology, vol. 1. London: Saunders, 1973: 129–53.
3. Barnett HJM, Jousse AT. Nature, prognosis and management of post-traumatic syringomyelia. In: Barnett HJM, Foster JB, Hudgson P, eds. Syringomyelia: Major Problems in Neurology, vol. 1. London: Saunders, 1973: 154–64.
4. Batzdorf U, Klekamp J, Johnson JP. A critical appraisal of syrinx cavity shunting procedures. J Neurosurg 1998; 89: 382–8.
5. Donaldson I, Gibson R. Spinal cord atrophy associated with arachnoiditis as demonstrated by computed tomography. Neuroradiology 1982; 24: 101–5.
6. Edgar R, Quail P. Progressive post-traumatic cystic and non-cystic myelopathy. Br J Neurosurg 1994; 8: 7–22.
7. Eismont FJ, Green BA, Quencer RM. Post traumatic spinal cord cysts: a case report. J Bone Joint Surg Am 1984; 66: 614–8.
8. Falcone S, Quencer RM, Green BA, et al. Progressive post-traumatic myelomalacic myelopathy: imaging and clinical features. AJNR Am J Neuroradiol 1994; 15: 747–54.
9. Gebarski JS, Maynard FW, Gabrielsen TO, et al. Post-traumatic progressive myelopathy. Radiology 1985; 1157: 379–85.
10. Green BA, Green KL, Klose KJ. Kinetic nursing of acute spinal cord injury patients. Paraplegia 1980; 18: 181–6.
11. Green BA, Lee TT, Madsen PW, et al. Management of post-traumatic cystic myelopathy. Top Spinal Cord Inj Rehabil 1997; 2: 36–46.
12. Hoffman HJ, Hendrick EB, Humphreys RP. The tethered spinal cord: its protean manifestations, diagnosis and surgical correction. Childs Brain 1976; 2: 145–55.
13. Lee TT, Alameda GJ, Gromelski EB, et al. Outcome after surgical treatment of progressive post-traumatic cystic myelopathy. J Neurosurg Spine 2000; 92: 149–54.
14. Lee TT, Arias JA, Andrus HL, et al. Progressive post-traumatic myelopathy: treatment with untethering and duraplasty. J Neurosurg 1997; 86: 624–8.
15. Levi ADO, Sonntag VKH. Management of post-traumatic syringomyelia using an expansile duraplasty. Spine 1998; 23: 128–32.
16. MacDonald RL, Findlay JM, Tator CH. Microcystic spinal cord degeneration causing post-traumatic myelopathy: report of two cases. J Neurosurg 1988; 68: 466–71.
17. Milhorat TH, Kotzen RM, Mu HTM, et al. Dysesthetic pain in patients with syringomyelia. Neurosurgery 1996; 38: 940–7.
18. Pang D, Wilberger JE Jr. Tethered cord syndrome in adults. J Neurosurg 1982; 57: 32–47.
19. Perrouin-Verbe B, Lenne-Aurier K, Robert R, et al. Post-traumatic syringomyelia and post-traumatic spinal canal stenosis: a direct relationship. Review of 75 patients with a spinal cord injury. Spinal Cord 1998; 36: 137–43.
20. Quencer RM, Green BA, Eismont FJ. Post traumatic spinal cord cysts (PTSCC): clinical features and characterization with metrizamide computed tomography. Radiology 1983; 146: 415–23.
21. Quencer RM, Green BA, Eismont FJ. The clinical and radiographic features of post-traumatic spinal cord cysts. In: Post MDJ, ed. Computed Tomography of the Spine. Baltimore: Williams & Wilkins, 1984: 809–15.
22. Ragnarsson TS, Durward QJ, Nordgren RE. Spinal cord tethering after traumatic paraplegia with late neurological deterioration. J Neurosurg 1986; 64: 379–401.
23. Rossier AB, Foo D, Shillito J, et al. Post-traumatic cervical syringomyelia: incidence, clinical presentation, electrophysiological studies, syrinx protein and results of conservative and operative treatment. Brain 1985; 108: 439–61.
24. Schaller B, Mindermann T, Gratzl O. Treatment of syringomyelia after post-traumatic paraparesis or tetraparesis. J Spinal Disord 1999; 12: 485–8.
25. Schneider RC, Knighton R. Chronic neurological sequelae of acute trauma to the spine and spinal cord: III. The syndrome of chronic injury to the cervical spinal cord in the region of the central canal. J Bone Joint Surg Am 1959; 41: 905–19.
26. Sgouros S, Williams B. A critical appraisal of drainage of syringomyelia. J Neurosurg 1995; 82: 1–10.
27. Sgouros S, Williams B. Management and outcome of post-traumatic syringomyelia. J Neurosurg 1996; 85: 197–205.
28. Shaw MDM, Russell JA, Grossart KW. The changing pattern of spinal arachnoiditis. J Neurol Neurosurg Psychiatry 1978; 41: 97–107.
29. Sklar E, Quencer RM, Green BA, et al. Acquired spinal subarachnoid cysts: evaluation with MR, CT myelography, and intraoperative sonography. AJNR Am J Neuroradiol 1989; 10: 1097–104.
30. Smith KA, Rekate HL. Delayed postoperative tethering of the cervical spinal cord. J Neurosurg 1994; 81: 196–201.
31. Stevens JM, Olney JS, Kendall BE. Post-traumatic cystic and non-cystic myelopathy. Neuroradiology 1985; 27: 48–56.
32. Tator CH, Briceno C. Treatment of syringomyelia with a syringosubarachnoid shunt. Can J Neurol Sci 1988; 15: 48–57.
33. Umbach I, Heilporn A. Review article: post-spinal cord injury syringomyelia. Paraplegia 1991; 29: 219–21.
34. Wiart L, Dautheribes M, Pointillart V, et al. Mean term follow-up of a series of post-traumatic syringomyelia patients after syringo-peritoneal shunting. Paraplegia 1995; 33: 241–5.
35. Williams B. Post-traumatic syringomyelia, an update. Paraplegia 1990; 28: 296–313.