Functional outcomes of intensive rehabilitation versus fusion surgery among patients with low back pain from lumbar spine degenerative disease: a systematic review and meta-analysis

Tito Guillermo D. Rejante (  jhigzmd23@yahoo.com ) Jose R. Reyes Memorial Medical Center Section Of Neurosurgery Kevin Paul Ferraris Jose R. Reyes Memorial Medical Center Section of Neurosurgery https://orcid.org/0000-0002-0919-6658 Jose Carlos Alcazaren Jose R. Reyes Memorial Medical Center Section of Neurosurgery Joseph Erroll Navarro Jose R. Reyes Memorial Medical Center Section of Neurosurgery Kenny Seng Division of Neurosurgery, Department of Neurosciences, University of the Philippines–Philippine General Hospital, University of the Philippines College of Medicine, Manila, Philippines


Introduction
Back pain is extremely common, with over 600 million individuals a icted, and is the leading cause of years lived with disability worldwide. 1 Low back pain is the hallmark of lumbar degenerative disease; unfortunately, it is a nonspeci c complaint when trying to determine which patients will bene t most from surgery. 2 Initial management is mostly non-operative, but lumbar spinal fusion has been used for nearly almost a century and has shown clinical e cacy in decreasing pain and disability scores as well as giving patients the ability to return to work. 2 The principle of spinal fusion is to provide a biomechanically lasting interbody union, which can be accomplished with the use of different surgical approaches, implants and grafts. 3 However, fusion procedures have not been effective for all patients, and the alternative is physical therapy and rehabilitation.
Physical rehabilitation is the most common method used to apply non-operative treatment of symptoms of patient with chronic low back pain. Therapeutic protocols may include the use of modalities for pain relief, bracing, exercise, ultrasound, electrical stimulation, and activity modi cation. Physical rehabilitation is recommended to reduce pain, to restore range of motion and function, and to strengthen and stabilize the spine, and restore mobility of the neural tissue. 4,5 In low-and middle-income countries where the costs of spinal implants could be prohibitive for the majority of patients with chronic low back pain, this study becomes relevant in determining whether intensive rehabilitation may become a reasonable option.
In randomized controlled trials (RCTs) by Brox and coworkers, 6 Fairbank and coworkers, 7 and Mannion and coworkers, 4 there were no statistically signi cant differences between treatment groups randomized to either lumbar fusion surgery or cognitive intervention and exercises. In contrast, the RCTs by Moller and colleagues 8 and Fritzel and colleagues 9 showed that the patients who were randomized to the surgical group had better outcomes than their non-surgical counterparts. For clinicians who deliberate and ruminate on these two diverging treatment options, it becomes imperative to clarify the weights of evidences in a meta-analysis.

Search Strategy and Selection Criteria
We reviewed articles on the topic that were published between January 2000 and June 2020 in PubMed, the Cochrane Central Registry of Clinical Trials and EMBASE. The keywords used were "chronic low back pain," "lumbar spine," "degenerative disease," "spinal fusion," "lumbar fusion," "surgical stabilization," "physical therapy," and "rehabilitation." The search was limited to human subjects in a prospective, randomized controlled study design. An effort was made to search for grey literatures. RCTs comparing the outcome between groups of patients with chronic low back pain from degenerative disc disease or spondylosis with or without spondylolisthesis who were treated with fusion surgery or rehabilitation, and with random allocation between groups, were considered eligible for inclusion in this meta-analysis. We excluded trials that involved patients with low back pain due to fracture, metastasis, and in ammation. We also excluded trials that compared one form of surgery versus other types of operative intervention.

Study Selection and Reporting
All included articles were independently screened and assessed for validity and eligibility of studies. Appraisal was done to minimize bias using Cochrane Methodological risk assessment tool. The following data were extracted from each of the included trials: author, year of publication, type of population, study design, sample size, duration of study, intervention, comparator, study outcomes, location of population and study outcomes. Our report followed the generally accepted guideline that is the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).

Statistical Analysis
Data were extracted from tables and from the associated text describing the outcomes before and after treatment. The functional outcomes were pretreatment and posttreatment differences based on Oswestry Disability Index (ODI), low back pain or leg pain determined through the Visual Analog Scale (VAS) score, and overall patient outcome. The homogeneity of ORs was tested using Cochran's Q statistics. If homogeneity was rejected at the 0.1 level, then the ORs were treated using the random effects model, with the presumption that there were multiple potential sources of heterogeneity being present in the studies that were included. The overall odds ratio (OR) was computed using the Cochran-Mantel-Haenszel method. Meta-analysis results were presented as mean difference (MD) and standard deviations (SD), with 95% con dence intervals (CI), and graphically presented as forest plots. Inverse variance method was used to calculate estimates for continuous variables. Review Manager version 5.4 was the software used for the analyses.

Search Results
The search revealed 220 articles from database search and 15 articles from bibliography search. After duplicates and non-RCTs have been been excluded, 61 articles were screened and full-text articles were reviewed. Of the 11 screened-in articles, 6 were further excluded because they different outcome interest and similarity in patients involved. Five studies ful lled eligibility criteria and were analyzed after full-text systematic review. Figure 1 shows the ow of selection process while Table 1 Table 2 gives a summary of the risk of bias assessment for included studies. The subgroup analysis (Figure 4) shows the mean difference in improvement in leg pain between the Lumbar Spine Fusion Surgery and Rehabilitation groups. There was a 7.2 improvement in VAS score for leg pain (95% CI -8.58-22.97; p=0.37; I 2 =98%); however, the effect was not signi cant.

Overall Patient Outcome
The subgroup analysis ( Figure 5) shows the odds ratio in overall patient outcome between the Lumbar Spine Fusion Surgery and Rehabilitation groups. There was no signi cant difference in the overall patient outcome (95% CI 0.23-1.08; p=0.08; I 2 =78%).

Discussion
This meta-analysis of ve RCTs showed signi cant improvement in the change in disability index scores and low back pain among patients who underwent surgery compared to those who underwent intensive rehabilitation. However, there were no signi cant differences in the improvement of leg pain and overall patient outcome.
The Forest plot on change in the disability score ( Figure 2) shows the rather high heterogeneity of I 2 =98%. An important factor that may be considered is the duration of evaluation. Duration of follow-up may confound the assessment for disability and pain. A sensitivity analysis was done to include only the studies that followed up the patients after 2 years. A high heterogeneity is noted after sensitivity analysis wherein other factors may be considered such as the differences on the plan for rehabilitation and surgical technique. Studies of this kind of treatment wherein it is di cult to blind the patients and the caregiver were prone to performance bias.
Pain is subjective hence a validated tool to quantify pain was used in the included studies. There was signi cant improvement in pain scores for low back pain following surgery. For leg pain however, there was no signi cant improvement in the pain scores. The overall patient outcome measured in the study is equivalent to the patient's subjective perception of improvement. There was no statistical difference between lumbar spine fusion surgery and rehabilitation in terms of overall patient outcome.
The results of our meta-analysis corroborate that of Chou and colleagues 10 wherein surgery was found to be equivalent to an intense rehabilitation program for patients with for low-back pain and lumbar spondylolisthesis. The argument for fusion surgery is stronger in patients with lumbar degenerative disease presenting mostly as stenosis with evidence of instability. The landmark Spine Patient Outcomes Research Trial (SPORT) with randomized (n = 304) and observational (n = 303) patients who have lumbar spondylolisthesis, it was concluded after as-treated analysis that surgery was better in each outcome measure at 2-, 4-, and 8-year follow-up. [11][12][13] Given the overall ndings, lumbar spine fusion surgery can somehow provide more bene t in lowering disability and low back pain among patients with chronic low back pain from degenerative causes like spondylolisthesis. Intensive rehabilitation presents as a respectable alternative for patients who does not want to undergo surgery because it appears equivalent to lumbar spinal fusion surgery when it comes to improving leg pain and overall patient outcome. Considering costs and accessibility, intensive rehabilitation can also be offered to patients who refuse to undergo surgery and who have comorbid illnesses for which general anesthesia and surgery are contraindicated.

Conclusion
Among patients with chronic low back pain, lumbar spine fusion surgery showed improvement in functional outcomes of change in disability and low back pain when compared with intensive rehabilitation. However, the two treatment options showed no differences in terms of improvement of leg pain and overall patient outcome.

Declarations Con ict of Interest
The authors declare that the research was conducted in the absence of any commercial or nancial relationships that could be construed as a potential con ict of interest.

Study
Randomised patients (surgery/rehabilitation or physical therapy)  Table 2 Flow Diagram for Search Strategy