Anterior versus posterior approach for multilevel cervical spondylotic myelopathy

Abstract

This is the protocol for a review and there is no abstract. The objectives are as follows:

The primary objective of this review is to assess the effectiveness of anterior versus posterior surgical approaches in patients with multilevel cervical spondylotic myelopathy (MCSM), with a focus on pain intensity, quality of life, neurological outcomes, and harms/complications.

Background

Description of the condition

Cervical spondylotic myelopathy (CSM) is a common cause of progressive spinal neurological impairment in elderly people (Young 2000). Congenital spinal canal stenosis, osteophytes, disc herniations, redundant ligamentum flava, or ossification of the posterior longitudinal ligament may lead to spinal neurologic compression in degenerative multilevel cervical spondylotic myelopathy (MCSM) (Sayana 2011). CSM most often involves the C5‐6, C6‐7, and C4‐5 levels in descending order (Montgomery 1992). The clinical symptoms of CSM include neck pain, clumsiness of the hand, myelopathic gait, sensory changes and motor weakness in the hands and feet (Clark 1988; Edwards 2003). However, one study showed that 7.6% of patients over 50 years of age had spinal cord compression without clinical symptoms (Matsumoto 1998).

Myelopathy resulting from multilevel spondylosis in the cervical spine is caused by 1) static factors due to stenotic neurological compression, and 2) dynamic factors due to cervical spinal instability and hypermobility (Hirabayashi 1995). Multilevel instability may contribute to repetitive microtrauma of the spinal cord. A pressure‐induced neuroischaemia is thought to be a final common pathway (Parke 1988). Microtrauma and ischaemia are perceived to play a role in progressive neurological dysfunction. Neck pain is caused by facet osteoarthritis, intervertebral discal tear, instability, nerve root irritation, or muscle fatigue (Rao 2006).

Magnetic resonance imaging (MRI) is now the most valuable radiographic study for the diagnosis of MCSM. The natural history of cervical myelopathy is variable ‐ some patients have quiescent disease for long periods of time with intermittent periods of rapid decline, while others experience a slow, stepwise decline (Clarke 1956; LaRocca 1988; Lees 1963; Martz 2009; McCormick 2003; Nurick 1972). Patients with spinal cord compression are unlikely to improve without treatment (Clarke 1956; Epstein 1988; LaRocca 1988; Nurick 1972), however some authors also suggested that surgery failed to alter the natural history of CSM (Bednarik 1999; Kadanka 2000; King 2005; Rowland 1992). However, the data from these studies were inadequate to provide reliable conclusions due to significant risk of bias, lack of long‐term follow‐up, and small sample size. In a recent systematic review 20% to 60% of patients were shown to have deteriorated neurologically over time without surgery (Karadimas 2013). There is generally consensus that operative treatment is indicated in patients with progressive neurologic symptoms refractory to nonoperative treatment (Boyce 2003; Cusick 1991; Epstein 1988; Law 1993; Lebl 2011; Rao 2006).

A glossary of terms is provided in Appendix 1.

Description of the intervention

Surgical decompression has been accepted for the treatment of CSM (Edwards 2002), but there is controversy regarding the surgical management of CSM involving multiple levels. Anterior cervical decompression with fusion and instrumentation is suitable for cervical myelopathy involving one‐ or two‐levels with low incidence of complications (Emery 1998; Zhang 1983). However, surgical management for MCSM, involving more than two segments, is associated with unpredictable outcomes and a high incidence of complications (Riew 1999; Saunders 1998; Vaccaro 1998). The debate is ongoing as to whether anterior or posterior surgery is the optimal strategy for MCSM.

Anterior approaches consist of anterior cervical discectomy fusion and anterior cervical corpectomy fusion, while posterior approaches consist of laminectomy, laminoplasty, and laminectomy with instrumented fusion. In general, many clinicians and researchers believe that the anterior approach is more effective and that the posterior approach is safer. However, the treatment decision is influenced by the patient's age (e.g. surgeons often use a posterior approach for elderly patients in order to shorten operative time), pre‐operative neurological function, cervical alignment, segments involved, fracture pattern on pre‐operative imaging, and the surgeon's preference for the operative procedure.

Anterior approaches include anterior cervical discectomy fusion and anterior cervical corpectomy fusion. Posterior approaches include laminectomy and laminoplasty.

Anterior approach

Background

In contrast to posterior approaches, the anterior approach directly decompresses the structures most commonly responsible for myelopathy and can relieve cord compression which is caused by cervical kyphosis. In addition, the associated arthrodesis can correct kyphosis and reconstruct cervical alignment, relieve neck pain, restore the height of the decompressed segments, and prevent further degeneration and neurological deterioration over the fused segments.

Procedure

After exposing the anterior cervical spine, discs are removed completely by blade and curettes. In the anterior cervical corpectomy fusion procedure, the central party of the vertebral body is removed. However, only the osteophytic protrusion around the vertebral body can be addressed in the anterior cervical discectomy fusion procedure. Osteophyte and herniated discs should be removed. The posterior longitudinal ligament will be excised if necessary. Fusion is indicated to overcome the destruction of spinal mechanical integrity caused by decompression. Various implants (e.g. tricortical bone autograft, titanium mesh cages, and polyetheretherketone) can be used in the interspace to restore the height, correct the alignment, and fuse the segments of the cervical spine. Subsequently constrained or semi‐constrained plating systems are applied to stabilise and protect the decompressed and fused spinal levels. Hybrid reconstruction using discectomies and adjacent corpectomies can also be used in the management of MCSM. Research evidence indicates that excellent outcomes can be achieved using anterior management of MCSM (Emery 1998). However, when the anterior approach is used for more than two levels it often has less satisfactory results, with the incidence of hardware failure, nonunion, graft dislodgment, and other complications ranging from 17% to 45% (Bolesta 2000; Emery 1997; Robinson 1962; Simmons 1969; Zdeblick 1997).

Posterior approach

Laminectomy alone

Historically, laminectomy was considered to be the gold standard posterior operation for MCSM, however, it has been associated with many complications, such as segmental instability, postlaminectomy kyphosis, perineural scar formation, loss of cervical motion, and delayed neurologic deterioration (Butler 1992; Crandall 1977; Cybulski 1988; Guigui 1998; Herkowitz 1988; Lonstein 1977; Morimoto 1999). In the procedure of decompression, the posterior column stability is broken down by ablating laminae, central portions of facet joints, ligamentum flava, spinous processes, ligamenta supraspinale, and ligamenta interspinalia (Epstein 2003). Nowadays, due to these complications, laminectomy with fusion or laminoplasty has replaced simple laminectomy for posterior procedures in treating MCSM (Kumar 1999; Heller 2001).

Laminectomy with fusion

The procedure of laminectomy with fusion is similar to laminectomy alone. The aim of implant application is to stabilise the residual structure and prevent postlaminectomy kyphosis, and to allow reliable new bone formation around a stable scaffold. Currently, the most popular instruments are modern lateral mass and pedicle screw fixation systems. Although there are various procedures for placement of lateral mass screws, the most widely accepted methods are the Magerl technique and the Roy‐Camille technique (Hsu 2009).

Laminoplasty

Background

As an alternative to laminectomy, laminoplasty is a simpler procedure and is now more common. In an independent matched cohort analysis of laminoplasty versus laminectomy and fusion, laminoplasty was shown to have a lower incidence of complications and superior clinical outcomes based on objective and subjective evaluation (Heller 2001). Laminoplasty may reduce the rate of postlaminectomy kyphosis in contrast with laminectomy alone (Kamioka 1989), and adjacent segment degeneration without restricting cervical range of movement (Shaffrey 1999). Laminoplasty preserves posterior structures excluding partial ligamentum flavum resection (Mitsunaga 2012).

Procedure

The open and French door methods are most often recommended (Hirabayashi 1981; Kurokawa 1982). In these procedures, a high‐speed burr is used to grind down the lamina to create a hinge or an opening. The hinge is located at the transitional junction between lamina and facet joints. The French door method requires two hinges and a central opening. The open door method involves one hinge and one opening. In both procedures it is possible to enlarge the spinal canal by floating the lamina away from the spinal canal on its hinge. Laminoplasty without fixation has certain advantages, such as lower incidence of complications by avoiding the need for fusion and instrumentation, and by preserving cervical dorsal structures (Rhee 2008), however the procedure is often performed in association with lateral mass instrumentation and arthrodesis.

Advantages and disadvantages of different approaches

The advantages of the anterior approach are as follows: the anterior approach is a muscle‐sparing incision, with low wound complication rates. Because most protrusive pathology is located anteriorly, an anterior procedure may be more direct. It is an effective method for the kyphotic patient because it restores the normal lordotic curvature. It also prevents iatrogenic kyphosis due to hardware failure and reduces late onset neurologic deterioration.

The disadvantages of anterior procedures include the following: a requirement for meticulous surgical technique, longer surgery time, greater intraoperative haemorrhage, potential complications of dural injury and instrumented fusion such as pseudarthrosis, adjacent level spondylotic degeneration, graft dislodgement, and subsidence. The risk of morbidity that accompanies the anterior technique can be avoided (e.g. dysphagia, recurrent laryngeal nerve injury, dysphonia, dyspnoea, oesophageal fistula, impairment of laryngeal, or other pericervical organs).

The posterior approach is most often indicated to decompress the nerve roots and other pathology anatomically positioned in the posterior part of the spinal canal. However, the posterior approach also has its own disadvantages. Procedures without fusion do not stop motion at the pathologic segments, and they do not remove the anterior compressive pathology. Moreover, the incidence of wound infection and other wound complications is higher. Loss of cervical alignment may interfere with the dorsal shift of the cord, and has led to poor clinical outcomes (Kawakami 2000). As a result, this approach is not seen as suitable for patients with cervical kyphosis. Finally, patients often complain of neck pain which might be due to the posterior soft tissue destruction.

How the intervention might work

The minimum goal of surgical intervention is to decompress the spinal cord and nerve roots by increasing the volume of the spinal canal. Appropriate and timely decompression and enlargement of spinal canal space restores spinal cord morphology (Fujiwara 1989), relieves cord oedema (Suri 2003), and likely improves blood supply to the cord, all of which facilitate neurological recovery (Sayana 2011). Decompression may be achieved by one or a combination of excision of pathogenic spur, ligament, and disc element and posterior element removal or release. Further goals may include stabilising the spinal structure, restoring the height and correcting spinal alignment (Koller 2007). Neck pain may be relieved by ablation of torn disc and degenerative articular cartilage. Fixation and fusion can also contribute to relieve pain by stabilisation of spinal segment, distraction of disc space, and indirectly decompression of nerve root entrapment (Rao 2006).

Why it is important to do this review

Posterior approach MCSM is the traditional surgical approach for MCSM. In recent years, with the development of techniques and instruments, anterior approaches have been more widely used. Excellent outcomes have been reported in many studies (Belanger 2005; Boni 1984; Epstein 1992; Hillard 2006; Kirkpatric 1999; Sevki 2004). However, systematic comparisons of these surgical methods is rarely found in the literature. It is commonly believed that direct decompression is more effective than indirect compression, but research evidence indicates that indirect decompression can also achieve good results. In this review we will try to determine which approach is safer and more effective.

Objectives

The primary objective of this review is to assess the effectiveness of anterior versus posterior surgical approaches in patients with multilevel cervical spondylotic myelopathy (MCSM), with a focus on pain intensity, quality of life, neurological outcomes, and harms/complications.

Methods

Criteria for considering studies for this review

Types of studies

As there are few randomised controlled trials (RCTs) comparing anterior to posterior approaches in patients with multilevel cervical spondylotic myelopathy (MCSM), we will include quasi‐RCTs in addition to RCTs. Randomised studies are those in which the patient selection is made truly at random, by using a computer‐generated sequence to assign patients, closed envelopes, block randomisations, and similar. Quasi‐RCTs use a method of allocation such as date of birth or day of the week, i.e not a truly random method.

Types of participants

Inclusion criteria

We will consider studies that include adult patients who underwent the following operations: anterior cervical corpectomy fusion, anterior cervical discectomy fusion, laminectomy, laminoplasty with CSM involving three or more levels. Patients with a clinical diagnosis of MCSM should exhibit clear features of upper motor neuron disorder (e.g. weakness, stiffness, difficulty walking, and loss of dexterity). The diagnosis must be supported by appropriate computed tomography (CT) and MRI findings. Operative treatment should follow the absolute indication mentioned above. Patients will be 20 years of age or older. We will not impose any no restrictions by gender. We will include patients with a neutral to lordotic cervical alignment. We will also include congenital spinal canal stenosis, osteophytes, herniated disc material, redundant ligamentum flavum, and ossification of the posterior longitudinal ligament. The duration of neck pain and neurological dysfunction should be more than six months before operation.

Exclusion criteria

Exclusion criteria are as follows.

1. Patients treated previously with cervical surgery.

2. Patients with a kyphotic alignment.

3. Spondylotic myelopathy involving thoracic segment.

4. Acute cervical disc herniation, intradural pathology, segmental instability (> 3 mm), spinal cord compression due to trauma, tumour, abnormal bone metabolism.

5. Surgery that combined anterior and posterior approaches.

Types of interventions

We will include studies of anterior surgeries compared with posterior surgeries. Anterior surgeries include anterior cervical corpectomy fusion, anterior cervical discectomy fusion, and the hybrid method. Posterior surgeries include laminectomy with instrumented fusion and laminoplasty. All of these surgeries are traditional methods. For example, laminoplasty includes the open and French door methods; we will exclude modified methods such as combined artificial disc replacement.

Types of outcome measures

Outcome measures will include clinical, functional, and radiological parameters. The minimum length of follow‐up will be six months. We will categorise the duration of follow‐up as short‐term (six months to less than one year), medium‐term (one to five years), and long‐term (more than five years). One of the most important roles for reviews of non‐randomised studies is to assess potential unexpected or rare complications of interventions, so we will pay careful attention to all complications. Below, we list the outcome measures we expect to use, however, we will not exclude studies based on outcomes presented. We have not listed some rare complications as secondary outcomes, however, if an additional complication is carefully described and recorded, we will add it to the list.

Primary outcomes

• Pain intensity in the neck, measured by a visual analogue scale or other measure of pain severity (Grant 1999).

• Neurological outcomes compared to pre‐operative status, measured by the Japanese Orthopedic Association scores (Yonenobu 2001), or the Nurick scores (Nurick 1972).

• Quality of Life (for example SF‐36 (36‐Item Short‐Form Survey ‐ quality of life)) (King 2002).

Secondary outcomes

• Related deaths

• Re‐operation related to primary surgery

• Hardware failure with clinical implication

Other outcomes: Patient satisfaction according to Odom's outcome criteria (four‐level assessment of success of surgery in relieving pre‐operative symptoms) (Odom 1958).

Search methods for identification of studies

Electronic searches

We will perform a comprehensive search to identify all relevant studies in the following electronic databases.

• Cochrane Central Register of Controlled Trials (CENTRAL; the Cochrane Library, current issue, which contains the Back and Neck Review Group (CBN) trials register).

• MEDLINE (Ovid, 1946 to present).

• EMBASE (Ovid, 1980 to present).

• Clinicaltrials.gov.

• World Health Organization (WHO) International Clincial Trials Registry Platform (ICTRP).

• Science Citation Index (Web of Science, Core Collection).

• OpenGrey for grey literature.

A draft search strategy for MEDLINE is provided in Appendix 2. We developed the strategy using the methods from Furlan 2015 and Higgins 2011.

Searching other resources

We will comb the references of included studies and analyse citation tracking results. We will also contact other review authors who have carried out systematic reviews on review questions similar to our own. We will not impose any restrictions based on language.

Data collection and analysis

For data collection and analysis we will follow the guidelines recommended in the CBN Methods Guidelines (Furlan 2015), and the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Selection of studies

Two reviewer authors (ZKW, QHZ) will independently assess potentially eligible studies. To resolve disagreement between reviewers, a third review author (JWT) will be consulted.

Data extraction and management

Two independent authors (ZkW, YBQ) will perform data extraction. Authors will not be blinded with regard to the information about the journal name, the authors, the authors' affiliation, or year of publication. We will extract the following data: method of randomisation, blinding (patient/practitioner/analyst), outcomes and follow‐up times, characteristics of the study population (e.g. number of participants/age/gender/baseline scores of the outcome measures), characteristics of the intervention (anterior verus posterior, technique used), loss to follow‐up for each group, sponsorship of the trial, and if there was a conflict of interest for any of the study authors. We will use a standardised form and we will pilot test it to ensure we have included all important data. We will contact study authors if information is lacking. We will report all data in an Excel spreadsheet designed for the purpose.

Assessment of risk of bias in included studies

Two independent review authors will assess the risk of bias (QHZ, YZ). If these two review authors do not reach agreement in this regard, a third review author will be consulted(DEP). We will use the criteria recommended by the CBN Review Group to assess risk of bias of RCTs and quasi‐RCTs; we will describe the items as 'low risk', 'high risk' or 'unclear' (Furlan 2015; Higgins 2011). We will evaluate possible bias according to the generation of the allocation sequence, the concealment of allocation, blinding, incomplete outcome data, selective outcome reporting, and other sources of bias (see Table 1 and Table 2). If an article does not contain information on one or more of the internal validity criteria, we may contact the study authors for additional information. We will pilot test the 'Risk of bias' assessment on similar articles.

Table 1.

Sources of Risk of Bias

Bias Domain	Source of Bias	PossibleAnswers
Selection	(1) Was the method of randomization adequate?	Yes/No/Unsure
Selection	(2) Was the treatment allocation concealed?	Yes/No/Unsure
Performance	(3) Was the patient blinded to the intervention?	Yes/No/Unsure
Performance	(4) Was the care provider blinded to the intervention?	Yes/No/Unsure
Detection	(5) Was the outcome assessor blinded to the intervention?	Yes/No/Unsure
Attrition	(6) Was the drop‐out rate described and acceptable?	Yes/No/Unsure
Attrition	(7) Were all randomized participants analyzed in the group to which they were allocated?	Yes/No/Unsure
Reporting	(8) Are reports of the study free of suggestion of selective outcome reporting?	Yes/No/Unsure
Selection	(9) Were the groups similar at baseline regarding the most important prognostic indicators?	Yes/No/Unsure
Performance	(10) Were cointerventions avoided or similar?	Yes/No/Unsure
Performance	(11) Was the compliance acceptable in all groups?	Yes/No/Unsure
Detection	(12) Was the timing of the outcome assessment similar in all groups?	Yes/No/Unsure
Other	(13) Are other sources of potential bias unlikely?	Yes/No/Unsure

Furlan 2015

Table 2.

Criteria for a Judgment of ‘‘Yes’’ for the Sources of Risk of Bias

1	A random (unpredictable) assignment sequence. Examples of adequate methods are coin toss (for studies with 2 groups), rolling a dice (for studies with 2 or more groups), drawing of balls of different colours, drawing of ballots with the study group labels from a dark bag, computer‐generated random sequence, preordered sealed envelopes, sequentially‐ordered vials, telephone call to a central office, and preordered list of treatment assignments.Examples of inadequate methods are: alternation, birth date, social insurance/security number, date in which they are invited to participate in the study, and hospital registration number.
2	Assignment generated by an independent person not responsible for determining the eligibility of the patients. This person has no information about the persons included in the trial and has no influence on the assignment sequence or on the decision about eligibility of the patient.
3	Index and control groups are indistinguishable for the patients or if the success of blinding was tested among the patients and it was successful.
4	Index and control groups are indistinguishable for the care providers or if the success of blinding was tested among the care providers and it was successful.
5	Adequacy of blinding should be assessed for each primary outcome separately. This item should be scored ‘‘yes’’ if the success of blinding was tested among the outcome assessors and it was successful or: ‐for patient‐reported outcomes in which the patient is the outcome assessor (e.g., pain, disability): the blinding procedure is adequate for outcome assessors if participant blinding is scored ‘‘yes’’ ‐for outcome criteria assessed during scheduled visit and that supposes a contact between participants and outcome assessors (e.g., clinical examination): the blinding procedure is adequate if patients are blinded, and the treatment or adverse effects of the treatment cannot be noticed during clinical examination ‐for outcome criteria that do not suppose a contact with participants (e.g., radiography, magnetic resonance imaging): the blinding procedure is adequate if the treatment or adverse effects of the treatment cannot be noticed when assessing the main outcome ‐for outcome criteria that are clinical or therapeutic events that will be determined by the interaction between patients and care providers (e.g., cointerventions, hospitalization length, treatment failure), in which the care provider is the outcome assessor: the blinding procedure is adequate for outcome assessors if item ‘‘4’’ (caregivers) is scored ‘‘yes’’ ‐for outcome criteria that are assessed from data of the medical forms: the blinding procedure is adequate if the treatment or adverse effects of the treatment cannot be noticed on the extracted data
6	The number of participants who were included in the study but did not complete the observation period or were not included in the analysis must be described and reasons given. If the percentage of withdrawals and drop‐outs does not exceed 20% for short‐term follow‐up and 30% for long‐term follow‐up and does not lead to substantial bias a ‘‘yes’’ is scored. (N.B. these percentages are arbitrary, not supported by literature).
7	All randomized patients are reported/analyzed in the group they were allocated to by randomization for the most important moments of effect measurement (minus missing values) irrespective of noncompliance and cointerventions.
8	All the results from all prespecified outcomes have been adequately reported in the published report of the trial. This information is either obtained by comparing the protocol and the report, or in the absence of the protocol, assessing that the published report includes enough information to make this judgment.
9	Groups have to be similar at baseline regarding demographic factors, duration and severity of complaints, percentage of patients with neurological symptoms, and value of main outcome measure(s).
10	If there were no cointerventions or they were similar between the index and control groups.
11	The reviewer determines if the compliance with the interventions is acceptable, based on the reported intensity, duration, number and frequency of sessions for both the index intervention and control intervention(s). For example, physiotherapy treatment is usually administered for several sessions; therefore it is necessary to assess how many sessions each patient attended. For single‐session interventions (e.g., surgery), this item is irrelevant.
12	Timing of outcome assessment should be identical for all intervention groups and for all primary outcome measures.
13	Other types of biases. For example: ‐When the outcome measures were not valid. There should be evidence from a previous or present scientific study that the primary outcome can be considered valid in the context of the present. ‐Industry‐sponsored trials. The conflict of interest (COI) statement should explicitly state that the researchers have had full possession of the trial process from planning to reporting without funders with potential COI having any possibility to interfere in the process. If, for example, the statistical analyses have been done by a funder with a potential COI, usually ‘‘unsure’’ is scored.

Furlan 2015

Measures of treatment effect

We will analyse dichotomous outcomes by calculating odds ratios (ORs), and continuous outcomes by calculating the mean difference (MD) with corresponding 95% confidence intervals (CIs). When multiple measures are used for an outcome, we will report a standardised mean difference (SMD). We will analyse heterogeneity of studies carefully, and if appropriate, synthesise the results quantitatively using a random‐effects method. We will define time frames to reflect short‐term, medium‐term, and long‐term follow‐up.

We will use the guideline of a minimal clinically important difference suggested by Ostelo 2008 to assess the clinical relevance of our findings.

Unit of analysis issues

We expect the unit of the trials included in this review to be individual patients. We will only choose patients who underwent the definitive operation for the first time. In our clinical experience, multiple anterior or posterior approach surgeries are unlikely to be performed in one patient due to the limitation of cervical levels. To avoid repeated measurement, we will define several different outcomes, based on different periods of follow‐up, and we will perform separate analyses.

Dealing with missing data

We will perform sensitivity analyses to assess the impact of including or excluding outcomes with high levels of missing data. If the standard deviation of follow‐up data are missing, we will use the study's baseline measure for the related follow‐ups. When the same outcome measures are available from other studies in the review, it may be reasonable to use these in place of the missing standard deviations.

Assessment of heterogeneity

We will evaluate clinical heterogeneity for patient characteristics, interventions, controls, outcomes, and other study characteristics. We may perform a test for the statistical homogeneity of studies to evaluate whether the differences among the results of the studies are greater than those that would be found by chance alone. We will assess statistical heterogeneity using the I² test for RCTs and quasi‐RCTs. We will undertake meta‐regression analyses and subgroup analyses to identify important sources of heterogeneity.

Assessment of reporting biases

We will use a funnel plot to assess reporting biases if we include more than 10 trials. For a test for funnel plot asymmetry, we will perform Egger's regression. However, differences in methodological quality or true heterogeneity in intervention effects may lead to funnel plot asymmetry. We will try to use ‘contour‐enhanced’ funnel plots which may help us to differentiate asymmetry due to publication bias from that due to other factors. We will investigate abstracts and conference proceedings to assess if there is a risk of publication bias; if we find publication bias we will not include such studies in the analysis of the review.

Data synthesis

We will only combine the results from studies when we judge them to be sufficiently clinically similar to yield meaningful results. If the data characteristics are heterogenous (i.e. I² > 30% but < 60%), we will try to explain the heterogeneity and use subgroup analysis.

If we are unable to perform a meta‐analysis, we will describe the results from clinically comparable trials narratively in the text.

Regardless of whether there are sufficient data available to use quantitative analyses to summarise the data, we will assess the overall quality of the evidence for each outcome. To accomplish this, we will use the GRADE approach, as recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), and adapted in the Cochrane Back and Neck Group methods guidelines (Furlan 2015). Following GRADE guidelines (Atkins 2004), we will categorise the final grade for quality of evidence for each domain as: high, moderate, low, or very low. We will grade the available evidence on the following domains which are further discussed in Appendix 3 of this protocol: study design, risk of bias, inconsistency (heterogeneity), indirectness (inability to generalise), imprecision (insufficient or imprecise data), publication bias, magnitude of the effect, dose‐response gradient, and influence of all plausible residual confounding.

'Summary of findings' tables

We will present the findings of this review in 'Summary of findings' tables. We will focus on two main comparisons: (1) anterior surgeries compared to laminoplasty, and (2) anterior surgeries compared to laminectomy with instrumented fusion. We will report findings for the following main outcomes: pain intensity, neurological outcomes, and quality of life, as well as adverse effects and major complications (Table 3). Pain intensity in the neck will be measured by a visual analogue scale or other measure of pain severity (Grant 1999); neurological outcomes compared to pre‐operative status will be measured by the Japanese Orthopedic Association scores (Yonenobu 2001) or the Nurick scores (Nurick 1972); quality of Life will be measured by 36‐Item Short‐Form Survey (King 2002); and related deaths, re‐operation related to primary surgery, and hardware failure with clinical implication will be measured by the number of patients with related complications. We will also report on patient satisfaction using Odom's outcome criteria (Odom 1958). Timing of all outcomes will be at short‐term (six months to less than one year), medium‐term (one to five years), and long‐term (more than five years).

Table 3.

Summary of findings

Title Anterior versus posterior approach for multilevel cervical spondylotic myelopathy
Patient or population: patients with multilevel cervical spondylotic myelopathy Settings: inpatient care Intervention: anterior approach Comparison: posterior approach
Outcomes	Outcome type	Outcome measure	Comments
Neck pain	Continuous	VAS or NRS (maximum score of 10)	at short‐term (six months to less than one year), medium‐term (one to five years), and long‐term (more than five years)
Neurological outcomes	Continuous	Japanese Orthopedic Association scores (maximum score of 17) or Nurick scores (maximum score of 5)	at short‐term (six months to less than one year), medium‐term (one to five years), and long‐term (more than five years)
Quality of Life	Continuous	36‐Item Short‐Form Survey ‐ quality of life (maximum score of 100)	at short‐term (six months to less than one year), medium‐term (one to five years), and long‐term (more than five years)
Related deaths	Dichotomous	number of patients with operation‐related deaths since initial surgery	at short‐term (six months to less than one year), medium‐term (one to five years), and long‐term (more than five years)
Re‐operation related to primary surgery	Dichotomous	number of patients with secondary surgery at the operated level; 'yes' is defined as any secondary surgery at the operated level, regardless of the time since initial surgery	at short‐term (six months to less than one year), medium‐term (one to five years), and long‐term (more than five years)
Hardware failure with clinical implication	Dichotomous	number of patients with secondary surgery at the operated level; 'yes' is defined as hardware dislodgement since initial surgery	at short‐term (six months to less than one year), medium‐term (one to five years), and long‐term (more than five years)
Patient satisfaction	Continuous	Odom's outcome criteria (4 grade)	at short‐term (six months to less than one year), medium‐term (one to five years), and long‐term (more than five years)
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Subgroup analysis and investigation of heterogeneity

If there are adequate studies to justify subgroup analyses (more than 10 studies), we will analyse subgroups according to diagnosis of MCSM (according to symptoms, MRI and/or CT), pre‐operative neurological function (according to the Nurick scale (Nurick 1972) or Japanese Orthopedic Association scores (Yonenobu 2001)), pre‐operative patient age, duration of symptoms, ambulation status, analgesic use, sagittal alignment, number of cervical levels involved, surgical procedures, concomitant interventions, and length of follow‐up time by meta‐regression.

Sensitivity analysis

A sensitivity analysis is a repeat of the primary analysis or meta‐analysis, substituting alternative decisions or ranges of values for decisions that were arbitrary or unclear (Higgins 2011). We will use a sensitivity analysis to assess analysis methods, eligibility criteria, publication bias, and evaluation of missing data.

Appendices

Appendix 1. Glossary of terms

Anterior surgery: involving an anterior incision and approach to the vertebral body

Arthrodesis: surgical fusion of two bones

Cervical kyphosis: refers to the abnormally excessive convex kyphotic curvature of the spine as it occurs in the cervical region

Decompression: removal of bone and soft tissue surrounding nerve structures

Dysphagia: difficulty in swallowing caused by oesophageal or neurological injury

Dysphonia: disorders of the voice caused by tracheal or neurological injury

Dysponea: difficulty in respiration caused by tracheal injury

Facet joints: small stabilising joints located between and behind adjacent vertebrae

Horner syndrome: a syndrome caused by sympathic nerve damage

Iatrogenic kyphosis: kyphosis caused by spinal surgery

Kyphosis: refers to the abnormally excessive convex curvature of the spine as it occurs in the cervical, thoracic and sacral regions

Lordotic cervical alignment: physiological cervical curve

Myelopathic gait: difficulty walking, lower extremity stiffness and hyperreflexic jerking

Myelopathy: any neurologic deficit related to the spinal cord

Neuroischaemia: diminished neurological nutrition and oxygenation

Perineural scar formation: scar formation between the dura and muscle

Postlaminectomy kyphosis: a common occurrence which describes a loss of the cervical physiological curve

Posterior surgery: involving a posterior incision and approach to the posterior structures

Pseudoarthrosis: pseudo joint formation without fusion

Spinal canal stenosis: narrowing of the spinal canal

Spondylotic: degeneration of the spinal column from any cause

Appendix 2. MEDLINE search strategy

1. Cervical Vertebrae/

2. Neck/

3. Cervical Cord/

4. Intervertebral Disc/

5. cervical.mp.

6. spine.mp.

7. spinal canal.mp.

8. or/1‐7

9. ((single or double) adj2 (level or segment)).mp.

10. 8 not 9

11. ((cervical or spine or spinal) adj2 (compress$ OR stenos$)).mp.

12. neck pain.mp.

13. (ossifi$ adj3 (posterior longitudinal ligament)).mp.

14. ligament calcinosis.mp.

15. Spinal Stenosis/

16. Spinal osteophytosis.mp.

17. Spondylosis.mp.

18. Ligamentum Flavum.mp.

19. Intervertebral Disc Displacement/

20. Intervertebral Disc Degeneration/

21. ((spine OR spinal) adj3 disease$)

22. cervical adj3 (myelopath$ or radiculomyelopath$).mp.

23. cervical spondylosis.mp.

24. Spinal Cord Diseases/

25. Spinal Cord Compression/

26. Spinal Osteophytosis/

27. or/11‐26

28. exp Surgery/

29. exp Surgical Procedures, Operative/

30. osteotom$.mp.

31. decompress$.mp.

32. Laminectomy/

33. laminectom$.mp.

34. laminotom$.mp.

35. osteotom$.mp.

36. enlargement.mp.

37. Laminoplasty/

38. laminoplast$.mp.

39. Spinal Fusion/

40. Diskectomy/

41. dorsal$.mp.

42. ventral$.mp.

43. ((open OR france OR double OR two) adj2 door).mp.

44. hybrid reconstruction.mp.

45. ACCF.mp.

46. ACDF.mp.

47. anterior$.mp.

48. posterior$.mp.

49. or/28‐48

50. animals.sh. not (humans.sh. and animals.sh.)

51. randomized controlled trial.pt.

52. controlled clinical trial.pt.

53. Random Allocation/

54. Double‐blind Method/

55. Single‐blind Method/

56. clinical trial.pt.

57. exp Clinical Trials/

58. (clin$ adj25 trial$).mp.

59. groups.ab.

60. controlled trial.mp.

61. ((singl$ or doubl$ or trebl$ or tripl$) adj25 (blind$ or mask$)).mp.

62. Placebos/

63. placebo$.mp.

64. random$.mp.

65. Research design/

66. volunteer$.mp.

67. or/51‐66

68. 67 not 50

69. 10 and 27 and 49 and 68

Appendix 3. The GRADE approach to evidence synthesis

We will categorise the quality of evidence as follows.

• High (⊙⊙⊙⊙): further research is very unlikely to change the confidence in the estimate of effect.

• Moderate (⊙⊙⊙○): further research is likely to have an important impact in the confidence in the estimate of effect.

• Low (⊙⊙○○): further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.

• Very Low (⊙○○○): any estimate of effect is very uncertain.

We will grade the evidence available to answer each subquestion on the domains in the following manner.

1. Risk of bias

Limitations in the study design and implementation may bias the estimates of the treatment effect. Our confidence in the estimate of the effect and in the following recommendation decreases if studies suffer from major limitations. We will examine all studies on five types of biases as follows.

• Selection (random sequence generation, allocation concealment, group similarities at baseline).

• Performance (blinding of participants, blinding of healthcare providers).

• Attrition (dropouts and intention‐to‐treat analysis).

• Measurement (blinding of outcome assessors and timing of outcome assessment).

• Reporting bias (selective reporting)

When it comes to grading the quality of the evidence, we will not downgrade evidence from studies judged as being at"low" or "unclear" risk of bias for all categories. We will downgrade evidence as follows:

• Byone level when three or fewer categories for each study are judged to have a "high" risk of bias.

• By two levels when four or more categories for each study are judged to have a "high" risk of bias.

2. Inconsistency

Inconsistency refers to an unexplained heterogeneity of results. Widely differing estimates of the treatment effect (i.e. heterogeneity or variability in results) across studies suggest true differences in underlying treatment effect. Inconsistency may arise from differences in: populations (e.g. drugs may have larger relative effects in sicker populations), interventions (e.g. larger effects with higher drug doses), or outcomes (e.g. diminishing treatment effect with time). We will downgrade the quality of evidence as follows.

• By one level: when the heterogeneity or variability in results is large (for example: I² above 80%).

• By two levels: when the heterogeneity or variability in results is large and there was inconsistency arising from populations, interventions, or outcomes.

3. Indirectness

Indirect population, intervention, comparator, or outcome – the question being addressed in this systematic review is different from the available evidence regarding the population, intervention, comparator, or an outcome in the included randomised trial.

We will downgrade the quality of evidence as follows.

• By one level: when there is indirectness in only one area.

• By two levels: when there is indirectness in two or more areas.

4. Imprecision

Results are imprecise when studies include relatively few participants and few events, and thus have wide confidence intervals around the estimate of the effect. In this case, we judge the quality of the evidence lower than we otherwise would because of resulting uncertainty in the results. We will consider each outcome separately.

For dichotomous outcomes we will consider imprecision for either of the following two reasons.

1. There is only one study. When there is more than one study, the total number of events is less than 300 (a threshold rule‐of‐thumb value) (Mueller 2007).

2. 95% confidence interval around the pooled or best estimate of effect includes both: 1) no effect; and 2) appreciable benefit or appreciable harm. The threshold for 'appreciable benefit' or 'appreciable harm' is a relative risk reduction (RRR) or relative risk increase (RRI) greater than 25%.

We will downgrade the quality of evidence as follows.

• By one level: when there is imprecision due to (1) or (2).

• By two levels: when there is imprecision due to (1) and (2).

For continuous outcomes we will consider imprecision for either of the following two reasons.

1. There is only one study. When there is more than one study, total population size is less than 400 (a threshold rule‐of‐thumb value; using the usual α and β, and an effect size of 0.2 standard deviations, representing a small effect).

2. 95% confidence interval includes no effect and the upper or lower confidence limit crosses an effect size (standardised mean difference) of 0.5 in either direction.

We will downgrade the quality of evidence as follows.

• By one level: when there is imprecision due to (1) or (2).

• By two levels: when there is imprecision due to (1) and (2).

5. Publication bias

Publication bias is a systematic underestimate or an overestimate of the underlying beneficial or harmful effect due to the selective publication of studies. We will downgrade the quality of evidence as follows.

• By one level: when the funnel plot suggests publication bias.

Contributions of authors

Drafting the protocol: Zhen‐kai Wu, Li Zhao.

Selection of studies: Zhen‐kai Wu, Qing‐hua Zhao, Ji‐wei Tian.

Data extracting and management: Zhen‐kai Wu, Yong‐bing Qian, Yi Zhou, Daniel E Porter.

Methodological evaluation: Fan Yang, Yi Zhou.

Zhen‐kai Wu and Qing‐hua Zhao contributed equally to this work.

Sources of support

Internal sources

• Shanghai Jiaotong University School of Medicine, China.

External sources

• No sources of support supplied

Declarations of interest

Zhen‐kai Wu has nothing to disclose.

Li Zhao has nothing to disclose.

Qing‐hua Zhao has nothing to disclose.

Ji‐wei Tian has nothing to disclose.

Yong‐bing Qian has nothing to disclose.

Yi Zhou has nothing to disclose.

Daniel E Porter has nothing to disclose.

New