Subacute posttraumatic ascending myelopathy (SPAM): A potential complication of subarachnoid shunt for syringomyelia?

Abstract

Context: Treatment of primary spinal syringomyelia is still controversial. Among others, shunting syrinx fluid to the subarachnoid, peritoneal or pleural space has been utilized with varying success. Shunt obstruction, migration, and infection represent the most common complications of these procedures.

Findings: The authors present the case of an 81-year-old woman who developed an unusual neurological deterioration resembling a subacute posttraumatic ascending myelopathy (SPAM) after the insertion of a syringosubarachnoid shunt for the treatment of slow-growing D10 syringomyelia.

Conclusion/Clinical Relevance: To date, no cases of SPAM secondary to the insertion of a syringosubarachnoid shunt for the treatment of syringomyelia have been reported. The potential pathogenesis related to this phenomenon is discussed.

Abbreviation list

SPAM:

Subacute posttraumatic ascending myelopathy

MRI:

Magnetic resonance imaging

MRC:

Medical Research Council

CT:

Computed tomography

SCI:

Spinal cord injury

Introduction

While craniovertebral decompression plus duroplasty is a well-established treatment in case of syringomyelia associated with Chiari malformations, surgical management of primary spinal syringomyelia is still controversial.^1,2 Shunting syrinx fluid to the subarachnoid, peritoneal, or pleural space is a potential surgical treatment that has been utilized with varying success. Shunt obstruction, migration, and infection represent the most common complications of these procedures.^2–7

Here we describe the case of an 81-year-old woman who developed an unusual neurological deterioration resembling a subacute posttraumatic ascending myelopathy (SPAM) after the insertion of a syringosubarachnoid shunt for the treatment of slow-growing D10 syringomyelia.

Case presentation

This 81-year-old woman underwent 7 years before D10–L1 laminectomy for the evacuation of a spontaneous subdural dorso-lumbar hematoma presenting with an acute paraparesis, urinary retention and complete sensory loss of lower limbs. Following the surgical procedure sphincter control and anesthesia partially recovered. Conversely, the paraparesis progressively worsened over time despite physical therapy. Follow-up MRI imaging showed a posttraumatic thoracic myelopathy and the presence of a slow-growing D10 syringomyelia (Fig. 1).

Figure 1.

(Panels A and B) Pre-operative T2-weighted sagittal lumbar MRIs showing lumbar syrinx increasing over time. (Panel C) Pre-operative whole spine T2-weighted sagittal MRI showing no cervical myelopathy prior to surgery.

On admission, neurological examination revealed spastic paraparesis (3/5 MRC) and moderate impairment of proprioceptive, tactile, thermal and pain sensation in both lower limbs. Considering the paraparesis progressive worsening, surgical treatment was then proposed. Under ultrasound guidance, the syrinx was partially opened and a small catheter was introduced, creating a communication between the intrasyrinx compartment and the subarachnoid space. No changes during intraoperative somatosensory and motor evoked potential monitoring were noticed. The immediate post-surgical course was uneventful except for a transient mild worsening of the touch hypoesthesia in the right inferior limb. A CT scan confirmed shunt correct positioning.

Two days later she suddenly complained of a tingling sensation in the right hand followed by a burning pain involving the right hemisoma. A few hours later, she reported she could not move both her right hand and lower limbs. Neurological examination demonstrated flaccid paraplegia, bilateral anesthesia with a pinprick level at D5–D6, urinary retention, and right-side superior arm weakness (3/5 MRC), involving most notably distal muscle groups. Urgent MRI of the whole spine revealed an intramedullary T2 signal abnormality extending from C7 to the conus medullaris, with no gadolinium enhancement. No shunt migration was observed, and the syrinx was no longer visible (Fig. 2). Spinal angiography showed no signs of medullary ischemia and documented the patency of Adamkiewicz’s artery, but an unusual delayed venous filling was noticed (Fig. 2). The patient was immediately started on intravenous methylprednisolone 30 mg/kg bolus followed by 5.4 mg/kg/h for 23 hours, and urgent surgical removal of the shunt was performed. Postoperatively a mild improvement of left-sided lower limb distal muscle groups was observed, while urinary retention, anesthesia, and right hemisoma weakness remained unaffected. An MRI scan performed 1 week later showed an increased longitudinal eccentrically T2-high and T1-low signal from C3 to the conus medullaris (Fig. 2).

Figure 2.

Early postoperative spinal angiogram (Panel A) showing the patency of Adamkiewicz’s artery and the delayed spinal venous filling. Early (Panel B) and late (Panels C and D) T2-weighted sagittal spinal MRIs documenting the upward myelopathy occurring after surgery.

At the discharge, the patient was enrolled in an intensive rehabilitation program. At 1-year follow-up, the MRI shows nearly complete resolution of the cervical myelopathy and the absence of the syrinx, while the patient recovered her neurological pre-operative status (Fig. 3).

Figure 3.

One-year follow-up MRI showing the resolution of the cervical myelopathy.

Discussion

SPAM is defined as a progressive, delayed, upward migration of loss of spinal cord or nerve root function days to weeks after an initial spinal trauma.^8,9 It is a rare, devastating complication developing mostly after traumatic spinal cord injury, such as a vertebral fracture-dislocation or a gunshot wound.^8–15 In severe cases, brainstem involvement with ailments such as dysphagia and increased ventilator requirement may be possible.^8,16 Frankel was the first to identify this pathological entity in 1968, describing eight patients harboring this condition.¹⁰ Since then, only small series described SPAM clinical and radiological features (Table 1). Burning pain or tingling paraesthesia in the superior limbs, shoulder, chest, or trunk may precede the neurological status worsening.¹⁶ SPAM MRI appearance usually consists of a central area of high-intensity signal on T2-weighted sequences ascending at least four segments above the initially injured segment.^8,9,16

Table 1. Summary of all cases of SPAM in the literature (adapted from Miller BA et al.¹⁴).

	Age(y)/Sex	Initial injury	Initial spine level	SPAM superior extent	Symptoms	Time until decline(d)
Frankel10	40/U 28/U 20/U 43/U 23/U 45/U 24/U	Fracture Fracture/dis Fracture Fracture/dis Fracture/dis Fracture/dis Fracture/dis	D12 D12 L1 L1 L4 L2 D5	D5 D7 D5 D7 C8 C7 C7	Rising sensory level Rising sensory level Rising sensory level Rising sensory level Rising sensory level Rising sensory level Rising sensory level	6 2 6 11 2 8 5
Aito et al.15	30/F	MVA	T11	C4	UE numbness	14
Belanger et al.13	21/M 44/M 25/M	Fall Fall Fall	C7 L1 D6	Medulla C3 C3	Respiratory decline UE weakness UE weakness and numbness	7 8 13
Al-Ghatany et al.21	37/M	MVA	D4	C5	UE paresthesias, hand weakness	12
Schmidt17	35/M	Hyperflexion	D11	C5	D5 sensory level	14
Planner et al.8	43/F 18/M 49/M 39/M 24/M 23/M 47/M 30/M 32/M 21/F 23/M	MVA MVA MVA MVA MVA MVA GSW Fall Vascular inj. Fall	C4/5 C3/4 C6 C5 D4 C5 D4 C5 C5 C6 C4	Foramen magnum Medulla C2 Foramen Magnum C2 Foramen Magnum Medulla Foramen Magnum Foramen Magnum Medulla Foramen Magnum	C4 symptoms Dysphagia, respiratory decline Neck/shoulder pain None Shoulder pain Respiratory decline Respiratory decline C3 symptoms Shoulder pain, UE sensory symptoms C4 symptoms C4 symptoms	10 14 69 16 86 12 36 7 8 4 13
Meagher et al.18	38/M	MVA	D4	C4	UE weakness and numbness	23
Kovanda and Horn12	15/M	MVA	D4	C3	UE weakness	38
Okada et al.16	75/F	MVA	D8	C2	UE weakness, respiratory decline	4
Miller et al.14	18/M	GSW	C4	C2	Bilateral UE weakness, long-tract signs of myelopathy	42
Present case	83/F	Shunt for posttraumatic syringomyelia	D10	C3	Right hemisoma paresthesias and pain, D5 sensory level	2

SPAM, subacute posttraumatic ascending myelopathy; SCI, spinal cord injury; U, unknown; dis, dislocation; F, female; M, male; MVA, motor vehicle accident; UE, upper extremity; GSW, gunshot wound.

The pathogenesis of SPAM remains unclear, even though several etiologies have been proposed: vascular or hypotensive injury to the spinal cord, disruption of the CSF flow dynamics, infective and autoimmune myelitis, anterior spinal artery thrombosis, venous congestion, or delayed secondary injury.^8,14,17,18 In our case, a spinal infarction following the occlusion of the anterior spinal artery or the Adamkievicz’s artery was excluded by a negative spinal angiography and by the recognition on MRI that spinal cord involvement did not match with a specific arterial territory. Moreover, the resolution of the myelopathy 1 year after the removal of the shunt cast doubts on the arterial ischemic pathophysiology, which typical course is that of sudden onset and little clinical recover.¹⁹ In our opinion, a decrease of spinal cord perfusion pressure with consequent ischemic insult cannot explain the clinical picture here reported for several reasons as well. First of all, the patient did not have any hypotensive episodes during surgery and postoperative course, being always hemodynamically stable. Second, a severe but single hypotensive episode would not explain the progressive nature of the neurological deficit suffered from the patient.

As far as the CSF dynamic theory is concerned, it has been speculated that intramedullary hemorrhage, edema, and debris following traumatic or surgical injury could alter normal CSF flow, blocking spinal cord canal drainage over time and leading to neurological dysfunction.⁹ Fischbein et al. termed this condition as a “presyrinx” state, which would be still reversible if normal CSF flow is swiftly re-established.²⁰ Although the good response to shunt removal and the imaging findings of our case correlate well with the CSF dynamic theory, we are skeptical about this possibility because of the clinical onset of our case. The “presyrinx” state, in fact, has been reported to develop over time, being a chronic condition. The relatively acute onset of the motor dysfunction in our case seems to not match this characteristic. No clinical signs of infective myelitis were noticed as well. The patient, in fact, did not present pyrexia, nor high blood leucocyte count. Furthermore, after the surgical removal of the shunt and the concomitant beginning of the steroid treatment, a mild improvement of the neurological condition was noticed, which would have been unlikely in case of infective or inflammatory aetiology.⁹ Commonly reported complications such as shunt infection, dislodgment and obstruction are rarely observed in the immediate postoperative course and for this reason we ruled them out as potential causative factors. Furthermore, postoperative CT and urgent spine MRI showed no evidence of mechanic or infective shunt-related problems. One may argue, instead, that a spine local trauma due to the shunt insertion could have produced a secondary delayed injury similar to that observed after some spinal cord injuries. This theory, however, lacks persuasiveness for two reasons. First, the rapid and great extension of the spinal cord involvement observed in our case is unusual for a second delayed injury. In addition, if the local trauma had been the main pathogenetic factor, SPAM appearance and progression would have been occurred shortly after surgery and not with a 48-hour interval.

During the imaging diagnostic work-up, a delayed venous filling was documented by spinal angiography. Relying on this neuroradiological information, we can only hypothesize that venous congestion was the most probable trigger factor of the ascending neurological deterioration. Other authors have already suggested this hypothesis.^13,14,17 How the insertion of the syringosubarachnoid shunt has potentially provoked the venous congestion is still unclear.

There are no actual guidelines for the treatment of SPAM.⁹ Current guidelines for Spinal Cord Injuries (SCI) have been considered to be the best practice in case of SPAM, while the utility of corticosteroids has not been proved yet.¹⁴

In conclusion, this is the first report showing SPAM as a potential severe complication of syringosubarachnoid CSF shunts in the treatment of post-arachnoiditis syringomyelia. The physiopathology of this phenomenon is not clear and probably is related to postoperative rearrangement of venous drainage from the spinal cord. Patients should be warned of this complication when informed consent is obtained.

Disclaimer statements

Contributors AF and VL equally contributed to the conception and design of the article, to the data acquisition, analysis, and interpretation. All the authors were involved in critically drafting/revising the article for important intellectual content. Finally, they all gave final approval of the version to be published.

Disclosure statement The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.

Funding This work received no specific grant from any funding agency in public, commercial, or not-for-profit sectors.