Introduction

The development of extensive lesions in the spinal cord after a traumatic spinal cord injury (SCI) is known as post-traumatic myelopathy (PTM).1, 2, 3, 4, 5, 6, 7 Cord tethering, defined as attachment of the spinal cord to the wall of the bony spinal canal,8 is known to be a major factor in the occurrence of such a complication.3, 4, 5, 9, 10 The long-term SCI patient follow-up highlights that PTM is a frequent complication among these patients.2 In addition to urinary, respiratory, vascular or cutaneous complications, PTM should be screened for and treated. PTMs, which can be either cystic (post-traumatic syringomyelia) or non-cystic (post-traumatic progressive myelomalacic myelopathy), are referred to as progressive pathologies.3, 5, 6, 11 The possibility that cord tethering could induce late lesions that would not be extensive but remain strictly limited to the injured area has not been evidenced in the literature. Our experience allowed us to identify patients who presented with such a pathology of the spinal cord.

Patients and methods

This study was conducted within the national reference centre for SCI patients at the university hospital of Nantes, France. Within our cohort of around 1500 patients, we selected all of the traumatic SCI patients who presented with changes in their neurological status (worsening) that were associated with a tethered cord on the magnetic resonance imaging (MRI), and who underwent surgery because of this deterioration. Patients with syringomyelia or extensive myelomalacia,8 as well as those who presented with a worsening within the first month post-SCI, were not included.

Clinical, radiological and therapeutic data were assessed retrospectively at the following four key follow-up periods: immediately after trauma (T0); before the worsening (T1); after the worsening (T2); and 6 months after the surgical treatment (T3). A standardised clinical evaluation was reproduced at T0, T1, T2 and T3. The neurological examination was performed according to the International Standards for Neurological Classification of Spinal Cord Injury.12 The American Spinal Injury Association (ASIA) Impairment Scale (AIS), motor and sensory levels, and motor and sensory scores were established in all of the patients. When pain was present, its typology (nociceptive and/or neuropathic) and location were reported according to the International Spinal Cord Injury Pain Basic Data Set.13 Pain intensity was assessed by the average pain intensity in the last week evaluated with a Numerical Rating Scale (NRS).13 The drugs used were identified and their efficacy was assessed on pain intensity as abovementioned. All patients had their analgesic drugs left unchanged between T2 and T3.

In addition to these clinical data, the following data were evaluated:

At T0, the type of spinal injury was specified using the AOSpine Fracture Classification.14 The type and delay of the surgical stabilisation procedure performed after the trauma was specified if performed.

At T1, spinal canal stenosis was measured using the method previously described by Perrouin–Verbe et al.2 The most recent MRI was examined to identify the post-traumatic spinal cord lesions using the criteria defined by Wang et al.8

At T2, the MRI that was performed because of the neurological worsening was assessed using the abovementioned criteria.

Surgical treatment of the tethered cord was performed in all of the patients. The procedure was initiated with a laminectomy on at least three vertebral levels. The treatment of a residual spinal canal stenosis was necessary in two patients. Then, the dura mater was opened on its midline. A meticulous dissection of the arachnoiditis allowed untethering the cord and roots. The procedure concluded with an expansile duraplasty with a supersized synthetic tissue (Neuro-Patch, B-BRAUN Medical, Melsungen, Germany). The duraplasty was tacked up to the paraspinal muscles laterally, and to the spinous processes at both extremities of the zone of laminectomy.

This study was performed in agreement with the Helsinki Declaration relative to patients’ rights and in accordance with the law on data protection (last version no 2004–801, 6 August 2004). Because the data were collected retrospectively and patient management was not modified and according to French law (last version no 2004–806, 9 August 2004), this study did not require approval by a research ethics committee.

Results

Twelve patients met the inclusion criteria described above and were analysed in this study.

Sociodemographic data, type of initial vertebral lesions and post-traumatic neurological status (T0)

Clinical and radiological data concerning the initial traumatic SCI are detailed in Table 1. Four patients had lesions at the subaxial cervical spine (from C3 to C7) that caused tetraplegia. Eight patients had lesions at the thoracolumbar junction that caused paraplegia. All patients benefited from spinal surgery in the acute phase, which aimed to decompress the spinal cord and to realign and stabilise the spinal lesions. Two patients had a delayed surgical treatment because of their initially unstable haemodynamics (see Table 2).

Table 1 Patient population, type of vertebral injury and initial neurological status (T0)
Full size table
Table 2 Initial spinal lesions and their surgical treatment
Full size table

Clinical and radiological evolution of the patients before worsening (T1)

All of the tetraplegic patients presented with changes in their motor level and motor score, which indicated spontaneous motor recovery. The initial median progression of the motor score in these patients was 7 points (min=6; max=10, see Table 3).

Table 3 Initial motor level (T0) and evolution of the motor score for the patients from T0 to T3
Full size table

A major residual stenosis caused by intracanalar bony fragments was observed on the postoperative computed tomography scan in two patients.

A MRI of the spinal cord was performed at T1 in 8 out of the 12 patients. A focal cyst was noted in six cases and a focal myelomalacia in two cases (see Table 4).

Table 4 Data of the MRI of the spinal cord performed at T1 and T2
Full size table

Clinical and radiological changes after worsening (T2)

Clinical changes

Five patients complained about the onset of motor weakness (two cases were associated with pain), and seven patients complained about isolated pain. Therefore, these two groups of patients are discussed separately. Detailed data concerning the changes in the neurological status of patients between T1 and T2 are presented in Figure 1.

Figure 1
figure 1

Motor and sensory score changes and pain evaluation, observed between T1 (before worsening) and T2 (after worsening) in the 12 patients of our series. Motor and sensory losses are expressed as points on the motor or sensory ASIA score.12 Pain is evaluated using a 10-point NRS score.

Full size image

Patients with motor loss: Five patients (that is, four tetraplegics and one paraplegic) underwent motor loss in their most caudally preserved muscles. The affected muscles were located at the motor level, just above, or just below (that is, the zone of partial preservation in AIS A patients). In all of these patients, the motor loss was confirmed by two assessments made by at least two examiners. Four of these patients benefited from an electromyographic study, indicating intense acute denervation in the weakened muscles (Table 5). The delay between the initial trauma and the onset of motor loss ranged from 6 to 40 months (median=16.5 months) in four patients. The median loss on the motor score was 4 points (range: 1–8). Two patients also exhibited neuropathic pain with the same dermatomal distribution as motor loss.

Table 5 EMG data in the patients who presented with a motor loss
Full size table

Pain patients: Seven paraplegic patients presented with isolated pain, which had the features of at-level neuropathic pain13 in all cases. The onset of pain ranged from 1 to 4 months (median=2 months). The pain intensity was high with a median evaluation of 8 on a 10-point NRS (see Figure 1).

MRI scan data at T2

All of the patients benefited from a spinal MRI at T2 (see Table 4). A tethered cord was found in all cases. The data were compared with those obtained at T1 in the eight patients who had previously undergone MRIs. Only one patient exhibited changes of his spinal cord lesions (slight extension of a cyst).

Evolution after treatment of the aggravation (T3)

All patients who presented with motor loss benefited from surgery as a first-line treatment, with a median delay time of 2 months (min=1; max=4). In contrast, all of the patients who presented with isolated neuropathic pain were first treated with oral anticonvulsants. Because of failed medical treatment, surgery was later proposed and performed in all of these patients. The median time between the onset of pain and the surgery was 50 months (min=6; max=103) in these patients.

The results of the surgical treatment are presented in Figure 2 for those patients who presented with a motor loss. All of the patients were improved by surgery. The improvement on the motor score ranged from 1 to 7 points (median=3 points). In three patients, this improvement allowed the patient to gain one motor level.

Figure 2
figure 2

Evolution of the ASIA/ISCOS (International Spinal Cord Society) motor score for the five patients who presented with a motor weakness. Scores are given for T0 (time of initial injury), T1 (before worsening), T2 (after worsening) and T3 (after surgical treatment). The lower extremity motor score is given for patient 5.

Full size image

The results of the surgical treatment for the pain patients are presented in Figure 3. The median improvement on the NRS for these patients was 4 points (min=1.5; max=6). Three patients described a partial recurrence of pain within 2 weeks after surgery. In five patients, the drugs could be reduced from T3, as pain was frankly reduced at this time.

Figure 3
figure 3

Effects of cord release surgery on pain in our paraplegic patients (mean NRS score).

Full size image

Discussion

This series of traumatic SCI patients is characterised by the delayed onset of perilesional neurological aggravation (motor weakness and/or at-level neuropathic pain), in the absence of syringomyelia or extensive myelomalacia. The MRI data indicated the presence of a tethered cord in all of the patients. The surgical treatment of this tethered cord allowed symptom improvement, which indicates an association between cord tethering and neurological worsening.

The complication that patients in this series underwent can be considered a PTM; although, remarkable differences from the classical picture of progressive PTMs are present. The nosological framework of PTMs has been established by Edgar and Quail,3 who distinguished between cystic and non-cystic myelopathies. Both entities have been considered progressive. In our patients, the aggravation, in relation to the clinical and radiological data, remained strictly confined to the perilesional zone. With respect to clinical signs, pain patients complained about neuropathic pain, which was strictly at-level. The motor loss only involved the most caudally preserved muscles. In four out of five patients, aggravation led to the loss of one function level (3/5 muscle strength or greater), level which had recovered functional strength between T0 and T1. These patients, therefore, presented with a recovery/worsening sequence. Symptoms, such as spasticity increase, increased autonomic dysreflexia, bowel and bladder dysfunction, or below-level neuropathic pain, which are frequent in progressive PTM, have not been observed in our patients. In the pain patients of this series, the fair hindsight at surgery time (median=50 months) suggests withdrawing the hypothesis of a progressive PTM beginning with at-level neuropathic pain.

Radiological data confirmed that the myelopathy in our patients was localised. The signal intensity changes in the cord, whether cysts or malacia, were confined to the immediate surroundings of the initial lesion. When previous MRI data were available (8 patients), the MRIs performed at T2 were unchanged compared with the former MRI data in all of the patients but one (see Table 4 and Figure 4).

Figure 4
figure 4

MRI scans of patient 4 before and after the neurological deterioration. (a) MRI scan performed at T1, before worsening. A focal cyst is visible at level C6-C7, and the cord is tethered ventrally and dorsally. (b) MRI scan performed at T2 (after worsening), 3 months later, showing the cyst, possibly slightly bigger and the tethering.

Full size image

Therefore, it appears that among the PTMs, the pathological condition that is described in this series emerges. This focal entity consists in myelopathy and/or radiculopathy, which is strictly confined to the lesional and perilesional zones. A radiculopathy can be suspected because of an associated tethering of the roots in 9 out of 12 patients. We hypothesise that the term perilesional myeloradiculopathy (PMR) could be used to describe this condition. Although the natural evolution of a PMR to a progressive myelopathy is possible, this does not appear to be systematic, given our data. We, thus, consider PMR a distinct pathological entity rather than an incipient form of progressive PTM.

Progressive PTMs and PMR share the same pathophysiological mechanism, which is the cord tethering. Williams considered the presence of scars of the dura and arachnoid at the site of injury to be a critical factor in the genesis of a syrinx.4 Edgar and Quail emphasised the role of the tethered cord in both forms of PTMs.3 Many works have confirmed the causal link between arachnoiditis and syrinx15, 16 or myelomalacic myelopathies.17 In our study, all of the patients had spinal lesions that were located at the subaxial cervical spine or at the thoracolumbar junction. The anterior cervical spinal fusion allows significant motion of the spinal cord. In thoracolumbar lesions, a short posterior fusion allows residual motion of the lumbar spine in relation to the above fused segments. The tethering in these mobile zones could have induced secondary lesions of the cord and/or roots due to traction forces during spinal movements.18 These slight lesions were not visible on MRI scans but responsible for clinical and electromyographic changes.

We hypothesise that PMR treatment must follow the same principles and techniques as treatment of progressive PTMs. The aim is to release the tethered cord and prevent relapse by creating an enlarged subarachnoid space (expansile duraplasty). If there is an associated compression of the cord due to spinal canal stenosis, for example, by bony fragments, this issue must also be treated. Surgical treatment led to partial motor recovery in two patients and complete recovery in three patients. In five out of the seven pain patients, surgery was followed by an improvement of at least three points on the 10-point NRS. These encouraging results and the lack of efficacy of the previous medical treatment in these pain patients provide arguments for performing surgery for pain patients with a PMR.

Cord release surgery is delicate. The arachnoidolysis must be performed carefully to prevent additional lesions of the cord or roots. No patient of our series exhibited postoperative worsening. However, three patients experienced partial recurrence of pain within the first two weeks after surgery, possibly caused by re-tethering of the cord and roots, which has been previously emphasised.5 Because arachnoiditis is a form of healing, the possibility to prevent postoperative arachnoiditis is poor. The prevention of postoperative re-tethering could be based upon the quality of the expansile duraplasty, which must be tacked up to the paraspinal muscles or spinous processes, as well as the absence of a subarachnoid haemorrhage. Lee et al.5 also advocated the frequent turning of the patient from side-to-side into the prone position in the postoperative period, which could prevent the deposit of sediment in the recumbent position and the creation of a new tethering scar. Nevertheless, preventing re-tethering after surgery of a tethered cord remains poorly controlled. Because the risk of pain recurrence, possibly due to re-tethering, is well known, the pertinence of performing a DREZotomy in association with the cord release should be discussed. DREZotomy has been shown to be effective in relieving pain in SCI patients with chronic at-level neuropathic pain. Some studies have demonstrated pain improvement in 68%19 or 74%20 of patients with this technique. Sindou et al. recently described a two-step technique for SCI patients with at-level neuropathic pain.19 The first step was to release the cord and roots, and the second step was the DREZotomy. Our study supports the efficacy of cord release surgery alone. A prospective study that compares the results of cord release surgery alone versus cord release associated with a DREZotomy could help to identify the best treatment for at-level neuropathic pain in traumatic SCI patients.

Conclusions

PMR is a potential complication after a traumatic SCI, consisting in motor loss and/or at-level neuropathic pain. PMR can occur as soon as in the first 6 months post injury. It is caused by damage to the spinal cord and roots if these structures are tethered and submitted to traction, lengthening and compression constraints. This study stresses the necessity of performing an MRI for every SCI patient who presents with chronic neuropathic pain. An MRI can detect a curable cause, such as a tethered cord, for which surgical treatment is effective. This work suggests that to relieve symptoms and prevent further deterioration, early surgery should be proposed for patients who present with PMR.

Data Archiving

There were no data to deposit.