Summary Of Evidence

Background

Lower back pain is a common condition in the United States (Deyo, 2002 and Freburger, 2009), which may be isolated to the lower back or radiate to the lower extremities and may involve musculoskeletal structures or nervous structures as well. Recommended initial treatment for chronic lower back pain in most patients is based on an approach starting with self-care (Chou, 2007) and adding more aggressive interventions to include rehabilitation, medications, injections and even surgery in some patients (Chou, 2009).

A number of techniques for surgical fusion of the lumbar spine have been developed, generally classified as interbody fusions or posterolateral fusions (Mummaneni, 2014). There are numerous directional approaches to interbody fusion, including anterior, posterior, lateral and transforaminal approaches, with no clearly demonstrated superiority of one over the rest and as such the best approach is usually a function of patient-specific factors and surgeon comfort (Mobbs, 2015).

Outcome Measurement

Severity of lower back pain and response to treatment can be measured a large number of ways including on a pain scale or using a scale that assesses the ability of a patient to engage in life activities such as the Medical Outcomes Study (MOS) Short Form 36 (SF-36) or Oswestry Disability Index (ODI). Among the many back pain outcome measures available, the ODI is 1 of the most widely used in both research and clinical practice and this particular outcome measure has been recommended for use in clinical practice as an outcome measure for the spine by the American Academy of Orthopedic Surgeons. Moreover, this particular outcome measure has been selected for use in nearly all of the high quality research examining lumbar fusion that was found in the development of this LCD. An important question regarding any outcome measure is: What is the clinical importance of given change in the outcome measure? For the ODI in particular, a 10 point difference is considered the minimum clinically important difference for improvement (Hägg, 2003). However, a 10 point improvement is significantly below the 26 improvement that has been identified to correspond to a global improvement of “much better” in surgical outcome evaluation (Hägg, 2003).

Low back pain without neurologic compromise or radiculopathy

A number of randomized trials have been performed to examine the potential benefits of lumbar spinal fusion in patients with primarily axial low back pain.

One of the earliest large randomized trials of lumbar fusion in patients with low back pain was a study from Sweden in which 294 patients aged 25-65 with at least 2 years of low back pain with radicular pain were randomized to 1 of 3 surgical techniques versus conservative treatment based on physical rehabilitation (Fritzell, 2001). There was not a single primary outcome but rather numerous primary outcome measures including: pain, disability, global self-rating by the patient and back-to-work. Several scales were used, 1 of which was the ODI. The investigators found a statistically and clinically significant difference in all 4 primary outcome measures. There were complications in 17% of the patients randomized in the surgical group and no complications identified in the group managed conservatively.

A study of 349 patients aged 18-55 compared surgical spine stabilization to a rehabilitation program (Fairbank, 2005). The surgical technique was at the discretion of the surgeon. The primary outcome measure was ODI score after 2 years. The surgical group had a change in ODI score from 46.5 to 34.0 and the rehabilitation group had a change in ODI from 44.8 to 36.1. This change reached statistical significance but the investigators noted that this difference was “marginal,” and notably it is well below the minimum clinically important difference identified by Hägg (2003). There were also 19 complications in the surgery group, 11 of whom required re-operation. There were no specific complications identified related to rehabilitation in the conservatively managed arm.

A research report merging data from 2 randomized controlled clinical trials in Norway (Brox, 2010) compared instrumented posterolateral fusion with cognitive intervention and exercises in 124 patients. Patients included had at least 1 year of low back pain and they were excluded if they had evidence of neurologic compromise or radicular symptoms. The primary outcome was the ODI. There was no difference in outcomes between the 2 groups 4 years later in either the intention-to-treat or the as-treated analyses.

A systematic review of comparative and non-comparative studies examining the effect of surgical fusion in axial low back pain (Phillips, 2013) concluded that fusion surgery is supported by the literature as a treatment option in refractory low back pain. A guideline from the Congress of Neurological Surgeons has concluded that fusion or intensive rehabilitation may be appropriate treatment options in patients who have not responded to conventional conservative management (Eck, 2014). The North American Spine Society has published a guideline that suggests spinal fusion may be an appropriate treatment option in patients with refractory axial low back pain due to single level degenerative disc disease who have no psychiatric comorbidities and are not currently smoking (North American Spine Society, 2014).

Low back pain with spinal stenosis or radicular symptoms

Surgical treatment of patients with low back pain and symptoms of nerve root involvement has been studied with both observational and randomized controlled trials with the highest quality evidence in the treatment of chronically painful conditions without spinal instability or evidence of neurologic compromise.

Prior to the publication of the Maine Lumbar Spine Study, there was a paucity of outcomes data from large studies in the United States comparing surgical to non-surgical management of back pain with nerve root involvement. The Maine Lumbar Spine Study was an observational study that followed patients treated in community orthopedic and neurosurgical practices evaluated and treated for spinal stenosis (Atlas, 2005 A) or radiculopathy (Atlas, 2005 B) with either surgical or non-surgical techniques. Outcomes as far as 8-10 years post-operatively were assessed. This study suggested that surgical treatment may have a role in the management of these 2 conditions but the initial surgical treatment in these studies generally did not involve a spinal fusion.

To date the largest and highest quality research on the effects of lumbar spine surgery has come from the Spine Patients Outcome Research Trial (SPORT) project (Birkmeyer, 2002). This group of studies involved 13 clinical centers and was specifically designed to study the effects of lumbar spine surgery in the groups it was most commonly being used at the time of study design to help address some of the controversies surrounding spine surgery that arose in part based limitations of prior studies. The SPORT project also included an observational study to evaluate outcomes of patients who decline participation in 1 of the randomized trials. Some of the SPORT outcomes have been assessed as far as 8 years out. The SPORT project contained 3 multi-center randomized controlled trials evaluating surgical treatment in patients with intervertebral disc herniation (IDH), degenerative spondylolisthesis (DS) and spinal stenosis (SS). Symptoms of nerve root involvement were required for study enrollment and all patients had tried conservative management prior to study enrollment. Conservative management was tried for at least 6 weeks in the IDH group and for at least 12 weeks in the SS and DS groups.  Primary outcome measures included the SF-36 and ODI.  Subjects who were enrolled and randomized in the SPORT trials under the diagnosis of IDH or SS were surgically treated with discectomy and decompressive laminectomy respectively. Surgical fusion was not part of the planned treatment protocol in either group. Patients who were enrolled in the SPORT trials and randomized for a diagnosis of DS received a decompressive laminectomy and possibly a single level fusion.

Patients whose outcomes were studied in the SPORT project following treatment for IDH had failed 6 weeks of conservative management. There were 501 participants enrolled in the randomized study and 743 were enrolled in the observational study. As noted above, surgical treatment consisted of discectomy with fusion. In the randomized trial of IDH patients, there was no statistically significant difference in the primary outcome measures at 3 months, 6, months or 2 years but a high degree of cross-over limited the intention-to-treat analyses (Weinstein, 2006 A). In the observational cohort of IDH patients, which was an as-treated analysis, patients managed surgically had statistically significantly better primary outcomes at 3 and 6 months as well as at 2 years (Weinstein, 2006 B). In the 4 year follow-up study the as-treated analysis demonstrated persistently better outcomes in those treated with surgery rather than conservatively (Weinstein, 2008 A).

Patients who were enrolled in the SPORT trials for SS following 12 weeks of failed conservative treatment have been followed for at least 8 years and the planned surgical treatment did not involve fusion but rather decompressive laminectomy alone. A total of 289 patients were in enrolled in the randomized trial and 365 were enrolled in the observational cohort. Intention to treat analyses showed that the group who had surgery did significantly better on the bodily pain component of the SF-36, but not on the physical functioning component of the SF-36 or on the ODI at 2 years. As with the study of IDH patients, there was a significant rate of cross-over in the randomized trial, limiting the inferences that can be made from intention to treat analyses. In the as-treated analysis of both randomized patients and patients in the observational arm, patients treated surgically had significantly better outcomes on the SF-36s bodily pain and physical functioning components as well as on the ODI (Weinstein, 2008 B). Of the patients initially enrolled in the trial- 188 of those enrolled in the randomized trial, 252 of those in the observational cohort were available for follow-up at 4 years. The 4 year follow-up data indicated persistence of the findings at 2 years (Weinstein, 2010). At 8 years, 159 of those in the randomized trial and 189 of those in the observational cohort were followed-up. Subjects treated with surgery still appeared to have better outcomes, though there was evidence of a convergence of outcomes over time (Lurie, 2015).

Patients who failed 12 weeks of conservative management for degenerative spondylolisthesis were eligible for this arm of the SPORT project. Surgical treatment involved decompression with possible single level fusion. A total of 304 patients were enrolled in the randomized trial and 303 were enrolled in the observational cohort. The intention-to-treat analysis of the randomized trial showed no benefit at 1 or 2 years of surgery, though cross-over rates were high. The as-treated analyses of both the randomized and the observational groups showed better outcomes for surgery at 1 and 2 years on both the SF-36 and ODI (Weinstein, 2007). Follow-up data collected at 4 years demonstrated persistently better outcomes for the surgically treated patients in as-treated analyses (Weinstein, 2009).

In the spondylolisthesis arm of the SPORT trial, the surgical procedure of choice was laminectomy in addition to possible fusion, without an assessment of laminectomy with vs without fusion in this population. The Spinal Laminectomy versus Instrumented Pedicle Screw (SLIP) trial sought to address the issue of whether fusion plus laminectomy is superior to laminectomy alone in patients with spinal stenosis and spondylolisthesis (Ghogawala, 2016). This was a randomized controlled trial in which 66 patients aged 50-80 years old with stable degenerative spondylolisthesis with symptomatic lumbar SS were randomized to undergo decompressive laminectomy alone versus decompressive laminectomy with posterolateral instrumented fusion. The primary outcome measure was SF-36 core after 2 years and ODI was a secondary outcome measure. The group who underwent decompression and fusion had statistically significantly better SF-36 scores at 2 years, a difference that was maintained at 3 and 4 years following surgery. There was no significant difference in ODI score changes for the 2 groups. The fusion group had greater operative blood loss, time spent in the hospital and operative time. However, the rate of reoperation was significantly less in the fusion group (14%) as compared with the laminectomy alone group (34%).

The Swedish Spinal Stenosis Study was published at the same time as the SLIP trial and also addressed the issue of whether fusion plus decompression is superior to surgical decompression alone in patients within spinal stenosis both with and without spondylolisthesis (Försth, 2016). In this study 247 patients aged 50-80 years old with stenosis at 1 or 2 levels, of which 135 had spondylolisthesis, were randomized to decompression alone (with technique left to the surgeon) vs. decompression plus fusion. The primary outcome measure was the ODI 2 years post-operatively.  There were no differences in ODI change at 2 years in either treatment group, with similar results for both patients with and without pre-operative spondylotlisthesis. A 5 year follow-up of 153 patients also showed no difference between the 2 treatment groups.

Following the SLIP trial and the Swedish Spinal Stenosis Study, a pragmatic observational study of outcomes in clinical practice intended to assess non-inferiority of decompression alone as compared with decompression plus fusion in patients with spinal stenosis and lumbar spondylolisthesis was published (Austevoll, 2017). This study matched 260 patients who had decompression plus fusion with 260 patients who had decompression alone. Primary outcome measures included Numeric Rating Scale scores for leg and for back pain as well as ODI at 3 and 12 months post-operatively.  The authors reported that they were unable to conclude that decompression alone was non-inferior to decompression plus fusion. However, the study results indicate that the only primary outcome for which there was a statistically significant difference in the proportion of patients achieving the pre-specified minimal clinically important difference was Numeric Rating Scale back pain score.

Guidelines from the Congress of Neurological Surgeons indicate that there is a lack of high quality evidence to support the benefit of fusion in addition to laminectomy for patients who have spinal stenosis without spondylolisthesis (Resnick, 2014 A). Additional guidelines from the Congress of Neurological Surgeons developed prior to the publication of the SLP and Swedish Spinal Stenosis studies indicate that fusion may be a reasonable addition to decompression in patients who are getting surgical treatment for spinal stenosis with spondylolisthesis (Resnick, 2014 B).

Persistent Back Pain Following Previous Spine Surgery

One of the largest studies to date examining the outcomes of surgical fusion following an initial spinal surgery reviewed 100 cases of “failed back surgery syndrome” (Arts, 2012). All patients had at least 1 year of persistent pain refractory to conservative treatments after their initial spine surgery and were treated with pedicle screw fixation and interbody fusion in the revision surgery. Etiologies of failed back surgery syndrome specifically identified included previous discectomy, previous laminectomy, adjacent level disease and instability. The primary outcome measure was a dichotomous patient self-report regarding recovery with options of “good recovery” or “bad recovery”. Mean follow-up time after revision surgery with fusion was 14.7 months. On the primary outcome 35% of patients reported good recovery and the remaining 65% reported bad recovery.

Surgical treatment of other spinal conditions

To date there are no large randomized controlled trials specifically regarding the treatment of unstable spine morphologies, significant deformities, neurologic compromise, spinal trauma, major spinal anatomic abnormalities, infection, tumor or re-operation following an initial spine surgery. Spinal fusion is currently often considered a necessary component of surgical treatment for many of these patients based largely on observational studies and consensus guidelines (Floman, 2008; Kotwal, 2011; Roussoly, 2010; Swan, 2006; North American Spine Society, 2014).

An observational study of 327 patients who underwent lumbar surgical decompression and posterolateral lumbar fusion examined the relationship between surgical indication and surgical outcomes (Glassman 2009). Operative indications studied included spondylolisthesis, instability, stenosis, scoliois, disc pathology or need for revision surgery due to nonunion, adjacent level disease or post-discectomy revision. Multiple outcomes at 6 months, 1 year and 2 years were of interest including SF-36 score, ODI and numeric rating scales for back and leg pain. For the full cohort, there was an observed improvement in all outcomes but stratification of outcomes by surgical indication revealed differences among the response to surgery. As measured by ODI score change, improvements from greatest to smallest by indication were as follows: spondylolisthesis, scoliosis, disc pathology, postdiscectomy revision, instability, stenosis, adjacent level disease and non-union from prior fusion. Revision surgery was needed in 10% of the entire cohort with patients who were undergoing the index fusion as a revision surgery having higher rates of revision of the index surgery.

Analysis of Evidence

Studies examining the role of surgical treatment in patients with low back pain without evidence of nerve root involvement have been heterogeneous and have shown a limited benefit of questionable clinical value for lumbar fusion in a population of patients with chronic back pain. Guidelines from 2 major societies recommend consideration of fusion in a select group of patients but the North American Spine Society guideline acknowledged in the guideline text that such a recommendation would be controversial. High quality research does suggest that there is a measurable difference in improvement between patients with degenerative disc disease who are treated with fusion versus those who are treated conservatively. However, this measurable difference is small and not clearly clinically meaningful, while the difference in adverse events is clinically meaningful. Given the limited benefit that high quality research has shown for fusion surgery for the treatment of low back pain, this A/B MAC does not believe that evidence supports coverage for this indication.

A number of high quality studies provide evidence that lumbar decompressive surgery provides a significant benefit in the reducing of symptoms due to lumbar radiculopathy or spinal stenosis from multiple causes in patients who have not improved with conservative management. Fusion in conjunction with decompression has been well studied only in patients with lower extremity symptoms attributable to the spine and spondylolisthesis. While there is high quality evidence that fusion in conjunction with decompression provides clear and meaningful improvements in pain and function as compared with conservative management in patients with stable spondylolisthesis and symptomatic nerve root involvement, it seems that most of the benefit of such a surgery comes from the decompression component. Decompression plus fusion has little benefit over decompression alone in this patient population when considering the benefits and risks of both surgical approaches. While neurosurgical guidelines consider the use of fusion reasonable in this patient population, these guidelines were written prior to the publication of 2 high quality studies which were designed to measure the efficacy of such an intervention in this population and did not clearly find a meaningful benefit. This A/B MAC agrees with voices in the spine surgery community (Greenway, 2016; Joaquim, 2016; Peul, 2016) that the SLIP trial and Swedish Spinal Stenosis Study have provided strong evidence that there is little if any benefit conferred by adding surgical fusion to simple decompression in patients with stable or mild spondylolisthesis and spinal stenosis.

Large randomized trials examining the effects of surgical fusion as compared with treatment without fusion have not been done in patient populations with spinal instability related to a fracture or unstable spondylolisthesis, major spinal deformities, tumors, infection, neurologic compromise, trauma or revision after an initial spine surgery. The nature of these conditions is such that high quality research may be unfeasible or unethical. While there is a paucity of high quality evidence for these conditions, there appears to be little disagreement in the spine surgical community regarding the general approach to appropriate management recommended in treatment guidelines. In spite of the paucity of high quality evidence, these conditions likely warrant treatment. As such this A/B MAC is basing coverage determination on low quality research that is available.

Definitions and Scope

This local coverage determination (LCD) discusses fusions of the lumbar spine.

Background

This LCD will discuss indications for surgical fusion of the lumbar spine. Total disc arthroplasty is not covered by this LCD.

Requirements

Lumbar fusion must be performed by a qualified surgeon in an operating room with the necessary equipment. For elective surgeries, providers should verify that the necessary equipment and hardware to perform the procedure on a patient is available prior to the induction of anesthesia.

Documentation must demonstrate that the patient met at least 1 of the indications for the procedure as described below. Where possible, there must be documented shared decision making with the patient or the individual who is serving as the proxy decision maker for the patient with the appropriate discussion of anticipated risks and benefits of the procedure.

Indications for lumbar fusion include any of the following.

A patient must have at least 1 of the following:

1. Radiographic or clinical evidence of instability due to any 1 of the following:

     a. Congenital deformities,

     b. Trauma,

     c. Fractures,

     d. Chronic degenerative conditions,

     e. Tumor,

     f. Infection,

     g. Erosive conditions,

     h. Space-occupying lesions, or

     i. Iatrogenic causes, including expected instability as a consequence of another medically necessary spine procedure.

      If more than 1 level is fused under this indication, the record must reflect that each fused level is affected by 1 of the above conditions.

 2. Symptomatic spinal deformity (in the absence of instability or neural compression) which meets the following criteria:

     Both of the first 2 criteria must be met:

     a. Functional limitation in daily activities due to back pain or discomfort and

     b. Nonresponse to at least 1 year of non-operative treatment

     Any 1 of these 3 criteria must be met:

     c. Sagittal or coronal imbalance by at least 5 cm as measured on radiographic imaging of the entire spine, or

     d. Progression of deformity by at least 10 degrees, or

     e. Scoliotic curvature of greater than 30 degrees

 3. Revision surgery for pseudarthrosis following an initial spine surgery if the following conditions are all met.

     a. The patient had a period of reduced pain initially following surgery and

     b. The time since the prior surgery has been at least 1 year and

     c. There is clear radiographic evidence of pseudarthrosis  and

     d. The patient has exhausted available conservative treatment measures

The medical record must clearly reflect which conservative treatments the patient has tried or is not a candidate for and why, including medical therapies, physical and exercise therapies and injections. Outcomes for fusion in revision surgery, usually do not lead to pain relief and as such fusion is considered a last resort treatment option only when all other treatment options have failed. This information must be communicated to the patient prior to surgery to allow for appropriate shared decision making with a well-informed patient. The medical record must reflect that this counseling was done and that the patient wished to undergo surgery with the appropriately informed consent.

 4. Symptomatic compression of neural elements for which disc excision is necessary for decompression. The record must reflect disc compression of neural elements for each level fused based on this indication.