Fusion or no Fusion After Decompression of the Spinal Cord in Patients With Degenerative Cervical Myelopathy

Background:

Degenerative cervical myelopathy (DCM) is characterized by neck pain and stiffness, weakness and paresthesia of the extremities, sphincter disturbance and bowel and balance disorder. DCM is the most common cause of spinal cord dysfunction in the elderly worldwide (1) and the incidence is 41 per million within North America (2). The mean age for symptoms is 64 years of age, more men than women, 2.7:1, are affected and the most common level is C5-C6 (3).

Mechanism:

DCM is typically the consequence of degenerative disc herniation, osteophyte formation and hypertrophy of the ligamentum flavum that compress the spinal cord. Ossification of the posterior longitudinal ligament (OPLL), which is more prevalent in the Asian population, may also cause compression of the spinal cord.

With non-operative treatments, i.e. medication and physiotherapy, 20-60% of the patients deteriorate neurologically and surgical treatment is indicated (4).

Existing knowledge:

The surgical treatment for DCM is decompression of the spinal cord. Decompression may be achieved with an anterior or posterior approach. Several algorithms have been proposed on whether to choose anterior diskectomy and fusion, anterior corpectomy and fusion, posterior laminectomy with fusion, posterior laminoplasty or, posterior laminectomy alone (5,6). Anterior discectomy/corpectomy with fusion is recommended in patients with a straight or kyphotic spine with compression of less than three levels (6). A posterior approach is recommended in patients with cervical lordosis and compression of more than three levels (7). The WFNS Spine Committee modified these recommendations in 2019 towards a wider use of posterior approaches, e.g. in patients with posterior compression at 1 or 2 levels and patients with a flexible kyphosis (8). It was recommended to address anterior compression with an anterior approach and posterior compression by a posterior approach. Hence, when propensity score matching is performed on the basis of MRI classification and description of the degenerative changes in each patient, anterior and/or posterior compression of the spinal cord, there is no difference between anterior or posterior decompression and fusion approaches (9).

Fusion is recommended in patients with DCM and concurrent signs of instability but there is no definition of instability in the degenerated cervical spine (10). In the traumatic cervical spine, however, instability is defined as >3.5 mm translation or 11° rotation on lateral flexion-extension radiographs (11) but there are no indications that degenerative changes with intact ligaments and unfractured joints would be unstable in the same way.

Kyphosis of cSVA > 40 mm (13) has been correlated to worse postoperative outcome (normal cSVA = 17-11 mm) (14). Consequently the recommendation is to correct kyphosis by an anterior approach (8) but a correction does not seem to affect the outcomes (15).

It remains a matter of debate among spinal surgeons whether posterior fusion after laminectomy for DCM, should be mandatory or not. After reports of post-laminectomy kyphosis in the 1970s and 1980s (16) prophylactic fusion has commonly been combined with the laminectomy procedure (9). In a report from 1999, 34% of the patients developed kyphosis or swan neck deformity after laminectomy compared with 7% of patients surgically treated with laminoplasty, using a muscle-preserving technique (17). However, a muscle-preserving technique that retains the facet integrity as well as the extensor musculature may be used when performing posterior laminectomy as well and is observed to maintain sagittal balance after surgery without progression of kyphosis (18).

Distal junction kyphosis (DJK) is a kyphotic angulation of at least 10° at the distal segment adjacent to a fused level and occurs in 24% of patients within a year after fusion surgery (19). Adjacent segment pathology (ASP) is progression of degeneration at the levels adjacent to a fused level and may also necessitate reoperation with decompression and extended fusion surgery (20).

Considering the existence of muscle-preserving laminectomy techniques that can maintain cervical lordosis (26), there is reason to explore the additional value of instrumented fusion in the cervical spine.

Although both methods are widely used, they are yet to be compared in a randomized controlled study.

Need for a trial:

There exists a controversy among spinal surgeons regarding the need for posterior fusion when laminectomy for DCM is performed.

We hypothesize that laminectomy without fusion results in shorter hospital stay and quicker return to an active life, without reduced patient satisfaction, functional scores, or delayed kyphosis.

It is important to achieve good outcome with a single surgery, to spare this frail group of patients from reoperations. Therefore, reoperation for any reason within five years after the primary surgery will be the primary endpoint of the current study. Long-term follow-up radiographs and magnetic resonance imaging (MRI) is needed to assess differences in the subsequent degenerative changes including spondylolisthesis, kyphosis, and adjacent segment pathology (ASP) to compare the two strategies.

Additionally, data from this study can be used to identify risk factors for poor outcome to guide surgical decision making.

Objectives:

• To determine the surgical treatment associated with the lowest frequency of reoperations when treating participants with CSM by performing a non-inferiority study comparing laminectomy alone with laminectomy and fusion.
• To evaluate potential differences in outcome including participant satisfaction, functional scores, late degenerative changes including spondylolisthesis, kyphosis, DJK, ASP and, death.

Study setting:

• Uppsala University Hospital
• Karolinska University Hospital, Stockholm
• Ryhov Hospital, Jönköping
• Sahlgrenska University Hospital, Gothenburg

Sample size:

Based on data from the national Swedish spine registry on patients with CSM, reoperation was estimated to 3.4% after standalone laminectomy and 7.9% after laminectomy and fusion. Five year mortality was estimated to 16.3% in the same population.

We further determined that excluding a 5% excess rate of reoperation in the laminectomy group vs laminectomy and fusion was a clinically relevant target for the study, and therefore set the non-inferiority margin at 5 percentage points (pp).

With a sample size of 300 participants and with regards to mortality and an additional 5% loss due to dropout and emigration, we end up with 236 analyzable patients. This results in a power of 87% based on simulation using rerandomization.

Recruitment:

All participants diagnosed with DCM referred for surgical consultation to the orthopedic or neurosurgery departments at the participating centers will be eligible for the study. Oral and written information about the study will be given at the routine physiotherapist appointment before the scheduled doctor's appointment and repeated by the treating surgeon at the following consultation. Participants may be enrolled if they meet the inclusion criteria and sign informed consent.

Allocation:

Participants will be allocated to either standalone laminectomy or laminectomy and fusion through randomization with a 1:1 ratio using the REDcap software (Research Electronic Data Capture), after informed consent and agreement to be included in the study. After inserting the patient´s personal number into REDcap the program reports the random allocation of the patient according to the pre-constructed randomization list. The randomization is stratified for center and participant sex, i.e. using separate lists for each center and sex. The allocation sequence utilizes balanced blocks of three different sizes occurring in random sequence. The principal investigator and study collaborators are blinded to the sequence, the block sizes and block sequence.

Blinding:

Trial participants will not be blinded after assignment to interventions as they have online access to their medical records by a centrally managed system.

The outcome assessors and data analysts will be blinded by using a coding system for the treatment groups.

Dropouts:

Dropouts may be one out of two entities; 1) the participant actively leaves the study or, 2) the participant has died or do not show up on follow-ups for unclear reasons. In case 1, the participant will not be part of the study anymore and data will not be retrieved from other information sources. In case 2, information about living participants will be retrieved from the medical records, radiographs, and the Swedish patient registry.

Statistical methods:

To test for non-inferiority, a two-sided 95% confidence interval (CI) for the difference in failure rates between the two groups will be computed. To account for sparsity of events, the CI will be computed using rerandomization techniques,10 blocked on sex, since the randomization was stratified on sex. Non-inferiority will be claimed if the upper limit of the CI is less than 5 pp. If non-inferiority is demonstrated, superiority will also be tested using the same CI, although the study is likely underpowered to detect this.

All endpoints will be analyzed in the per protocol (PP) population, defined as randomized patients who are still alive without having emigrated or left the study after five years.

The secondary outcomes listed above will be analyzed using ordinal regression models, adjusted for sex. In addition, each secondary endpoint will be dichotomized and analyzed using logistic regression. The dichotomization will be done by comparing baseline and follow-up data, either based on MCID when applicable, or else by defining success as an improvement from baseline. All secondary endpoints will be analyzed at 1, 2 and 5 years of follow-up, but not until the study is closed and the primary results is published.

The study statistician is Lars Lindhagen at Uppsala Clinical Research Center.

Data collection:

Questionnaires including baseline questionnaires and postal follow-up questionnaires as well as validated PROMs will be distributed the participants preoperatively. Postoperatively the participants are routinely followed via the Swespine (swespine.se) and follow-up questionnaires and PROMS will be retrieved from Swespine.41,49,50

Closing statement:

Degenerative cervical myelopathy is the most common cause of spinal cord dysfunction in the elderly worldwide and the incidence is 41 per million within North America. Until now there is no consensus whether to fuse or not when laminectomy is performed and the choice of surgical method is mainly up to the surgeon's preference.

This will be the first randomized controlled trial comparing two of the most common surgical treatments for DCM; the posterior muscle-preserving selective laminectomy and posterior laminectomy with instrumented fusion.