To fuse or not to fuse: The elderly patient with lumbar stenosis and low-grade spondylolisthesis. Systematic review and meta-analysis of randomised controlled trials

Background: The optimum surgical intervention for elderly patients with lumbar spinal stenosis (LSS) and low-grade degenerative-spondylolisthesis (LGDS) has been extensively debated. We conducted a systematic review and meta-analysis of randomised-controlled-trials (RCTs) comparing the effectiveness of decompression-alone against the gold-standard approach of decompression-with-fusion (D þ F) in elderly patients with LSS and LGDS. Methods: A systematic literature search was performed on published databases from inception to October-2021. English-language RCTs of elderly patients (mean age over-65) with LSS and LGDS, who had undergone DA or D þ F were included. The quality and weight of evidence was assessed, and a meta-analysis performed. Results:


Introduction
Lumbar spinal stenosis (LSS) is the narrowing of the spinal and nerve root canals caused by hypertrophy of osseous and soft tissue structures within the lumbar vertebrae 1 which compresses the spinal nerves and blood vessels exiting the foramen.It clinically manifests with long-term radiculopathy; specifically, back pain (BP) and bilateral radicular leg pain (LP) and paraesthesia, with progression to lower limb weakness. 2,3ymptoms are aggravated by walking upright, standing or hyperextension due to further narrowing of the vertebral canal. 2 It is one of the most prevalent pathologies in the elderly population affecting 200,000 adults in the United States.It is estimated by 2025 that 64 million elderly people will be affected by the condition. 4In patients between the age of 40 to 49 years of age, the prevalence of LSS is estimated to be 3.8% in men and 1.4% in women increasing to 9.8% in men and 5.7% in women between ages 50 to 59 years. 5LSS precipitates BP in both middle-aged and elderly patients resulting in loss of productivity and work hours in the working population and consequently significant economic burden. 6Despite the dramatic decrease in quality of life (QoL) in those suffering and its overwhelming prevalence, an optimal treatment for elderly patients with both LSS and LGDS is yet to be definitively agreed.Surgical rates for LSS have grown significantly over the last decade, and currently, LSS is the most common reason for spinal surgery in patients 65 years and older. 7ecent studies have assessed the clinical effectiveness of DA and D þ F in patients with LSS and LGDS.Two previous systematic reviews demonstrated that DA is not inferior to gold-standard D þ F, irrespective of LGDS. 8,9However, a number of recent studies have been published which may alter these conclusions.This study aims to determine whether DA is as effective as D þ F in elderly patients, over the age of 65 years, with LSS and LGDS.

Methods and materials
A systematic review and meta-analysis of RCTs was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) statement guidance. 10

Eligibility and study selection
The inclusion criteria for both comparisons included RCTs of elderly adult subjects (mean age over 65 years) with LSS and LGDS, comparing outcomes of interest between DA (by open laminectomy, bilateral laminotomy or micro-endoscopic decompression) and D þ F (by posterolateral fusion (PLF), posterior lumbar interbody fusion (PLIF) or anterior lumbar interbody fusion (ALIF)).There were no limitations on geographical location.Exclusion criteria included: samples with mean age less than 65 years, patients with degenerative spondylolisthesis (DS) alone without LSS or with foraminal stenosis and studies with sample sizes of less than twenty subjects.Studies which did not compare both groups or assessed specific techniques of decompression or fusion (such as cage fusions only), studies with patient cohorts who did not undergo any instrumented fusions in their D þ F group and fusions where no autogenous bone graft was used, were excluded.

Outcome measures
Primary outcomes were postoperative BP and LP measured using Visual Analogue scale (VAS) scores ranging from 0 to 10; higher scores indicating greater degree of pain.Secondary outcomes: (a) degree of disability by Oswestry Disability Index (ODI) ranging from 0 to 100; higher score indicating greater degree of disability 11 (b) QoL using 36-item short form (SF-36) survey: physical component summary (PCS) and mental component summary (MCS) scores 12 ; (c) hospital complications: duration of operation, intra-operative blood loss and length of hospital stay and (d) adverse events: total number of surgical complications, incidence of dural tears, postoperative DS and reoperation rate.

Search strategy
Two authors [ANONYMOUS] independently performed a literature search using the databases: Ovid Medline, EMBASE, Cochrane Register of Systematic Reviews, Cochrane Register of Controlled Trials (CENTRAL), PubMed and Web of Science, from inception to June 2020.A manual search of reference lists of relevant reviews and their included studies was carried out.

Data extraction
Search strategies are shown in Supplementary File 1. Titles and abstracts were independently screened according to the PICOS criteria (Table 1) by two authors [ANONYMOUS] and full-text articles independently screened and assessed for eligibility.A third author [ANONYMOUS] resolved any discrepancies at title, abstract and full-text screening stages.The extracted data included basic study characteristics including participant age, gender, country of origin, surgical interventions and outcomes.Primary authors for all eligible trials were contacted to request missing data.

Quality assessment
Two authors independently performed data extraction [ANONYMOUS] of included RCTs and three authors assessed risk of bias [ANONYMOUS] in accordance with the Cochrane Handbook for Systematic Reviews of Intervention version 2. 13 The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) methodology was used by the aforementioned authors [ANONYMOUS] to assess the weight of evidence from the findings of the meta-analyses. 14t h e s u r g e o n 2 1 ( 2 0 2 3 ) e 2 3 ee 3 1

Data synthesis
Study heterogeneity was assessed for participant, intervention and study characteristics to determine if there was sufficient homogeneity of data pool.Where study homogeneity was assured by the review team, a meta-analysis was carried out.Meta-analysis results were expressed as weighted mean difference (MD) and 95% confidence intervals (CI), and dichotomous variables were reported as odd ratios (ORs) and 95% CI.Results were regarded as statistically significant if pvalues were less than 0.05.Statistical heterogeneity was measured using the I 2 scores and a fixed-effects model was implemented.Subgroup analysis was performed for primary outcomes in patients undergoing open laminectomy (DA group) and PLF (D þ F group) by excluding studies which used other techniques.Data were analysed using RevMan v.5.4 (Cochrane Collaboration, Oxford, UK). 15

Search results
The literature search (Fig. 1) generated 6690 records.Full-text articles were reviewed for 107 studies and six RCTs were eligible (N ¼ 531).

Study characteristics
The characteristics of included studies are shown in Table 2. Mean follow-up was 27.4 months (range 24 monthse37.2months).There were 256 patients and 275 patients within the DA and D þ F groups respectively (mean age ¼ 66.2 years; 57.8% female).The surgical techniques used were defined as shown in Table 4.All study participants had a diagnosis of LSS and LGDS based on clinical and radiological criteria (Table 5).Clinical criteria in all studies included the presence of typical symptoms of: neurogenic claudication or radiculopathic leg pain with associated neurological symptoms.Three studies used slippage of >3 mm to define LGDS 16e18 .One study used vertebral slippage exceeding 5% to define LGDS 19 and two studies did not specify their criteria 20,21 .In the DA group, the surgical technique consisted of an open laminectomy in four studies 17,18,21 both open laminectomy (82%, n ¼ 98) and bilateral laminotomy (18%, n ¼ 22) in one study, 16 bilateral laminotomy only in one study 20 and micro-endoscopic decompression in one study. 19In the D þ F group, the surgical technique was PLF in all studies, however one study 16 also carried out PLIF in 5% of patients (n ¼ 6) and non-instrumented fusion in 4% of patients (n ¼ 5).One study 19 also carried out PLIF in 47% of patients (n ¼ 8) and ALIF in 6% of patients (n ¼ 1).

Quality appraisal
Three of the six studies did not perform or did not demonstrate good random sequence generation 17,20,21 (Fig. 2).Only two studies demonstrated adequate allocation concealment. 19,18Nevertheless, all studies demonstrated very low rates of attrition bias.

Primary outcome measures: pain by visual analogue scale (VAS)
Three studies reported BP 19,16,18 and two studies reported LP 16,18 using VAS scores.The data showed no difference in BP or LP by VAS scores between patients who had undergone DA

Degree of disability
Two studies reported data on degree of disability by ODI scores 16,17 (Table 3).The data showed no difference in degree of disability between patients who had undergone DA compared to D þ F (MD 0.50 95% CI -3.31 to 4.31, p ¼ 0.80, N ¼ 294, GRADE: moderate) 18 reported no difference in postoperative disability scores, using Japanese Orthopaedic Association (JOA) scores, between patients who had undergone DA compared to D þ F (MD -1.40; 95% CI -3.85 to 1.05; N ¼ 58).

Quality of life
Only one study 17 reported data on QoL using SF-36 scores.This study showed that those who had undergone a DA had lower PCS score than those who had undergone D þ F at two years (MD -5.70; 95% CI 2.24 to 9.16; N ¼ 66, GRADE: moderate) and four years follow-up (MD -6.70; 95% CI -10.16 to À3.24; N ¼ 66, GRADE: moderate).

Hospital complications
Five studies reported duration of operation and volume of blood lost intra-operatively 16e20 whilst four reported data on length of hospital stay 16e19

Adverse events
Six studies reported the total number of surgical complications 16e21 whilst three studies reported the number of dural tears, 16,18,20 post-operative DS, 18,21 and five studies reported data on reoperation rate 19,16,18,20,21 .The data showed that patients who had undergone DA had fewer total number of surgical complications than those who had undergone D þ F (OR 0.57; 95% CI 1. 0.36 to 0.90, p ¼ 0.02, N ¼ 492, GRADE: moderate).However, those who had undergone DA had experienced more post-operative DS than patients who underwent D þ F (OR 3.49; 95% CI 1.05 to 11.65, p ¼ 0.04, N ¼ 103, GRADE: moderate).There was no difference in incidence of dural tears or reoperation rate between the two groups (OR: 0.94; 95% CI

Discussion
The results of this systematic review indicate that there is no difference in BP or LP postoperatively between elderly patients with LSS and LGDS, who had undergone DA compared to those who had D þ F. Data for degree of disability showed no  t h e s u r g e o n 2 1 ( 2 0 2 3 ) e 2 3 ee 3 1 difference whether patients had undergone either type of operation.Patients who had DA experienced less hospital complications and lower total number of surgical complications despite higher rates of post-operative DS.
Recent evidence points towards the non-inferiority of DA for treatment of LSS and LGDS compared to D þ F. This hypothesis is supported by findings of a recent Cochrane systematic review of 24 studies with 2352 participants with LSS and LGDS concluding that D þ F is not superior to DA. 23 Similarly, two previous systematic reviews demonstrated no difference in VAS pain or ODI scores between DA and D þ F. 8,9 Similarly, a large observational study 22 of 4259 patients included in the National Swedish Register for Spine Surgery (Swespine), concluded that there was no significant difference in mean VAS LP scores (p ¼ 0.57), ODI scores (p ¼ 0.33) or EQ-5D scores (p ¼ 0.69) between both treatment groups at two-years follow-up; regardless of the presence of pre-operative DS.A recent multicentre study of 306 patients enrolled in the Canadian Spine Outcomes and Research Network (CSORN) database showed clinically significant increased operative time, blood loss, length of hospital-stay and perioperative complications in the D þ F group. 24Nevertheless, both surgical interventions are not without their risks.DA was reported to be associated with post-operative vertebral instability. 25On the contrary, albeit the lower rates of worsening of DS post-operatively, several studies have shown that D þ F is associated with adjacent segment degeneration 26,27 .Another contentious issue is the economic burden of a fusion operation, having greater peri-operative cost implications as well as the economic consequences of higher complication rates (p < 0.001). 28is comprehensive systematic review and metaanalysis has a number of strengths.It presents the most up-to-date evidence from RCTs comparing the effectiveness of DA to D þ F for elderly patients over 65 years with LSS and LGDS.We considered only experimental studies hence providing a higher level of evidence than if we had also included observational studies.The participants of studies represent a large demographic spread from various different healthcare systems which increases the generalisability and global applicability of our findings.All literature searches, data extraction, meta-analyses and quality appraisals were put through a rigorous cross-check by at least two independent researchers at every stage.This study also used the GRADE approach to evaluate the strength of the evidence allowing readers to appreciate the paucity of high-quality evidence on this subject area in the current literature.There are also some limitations worth highlighting.There was a high degree of heterogeneity in defining LGDS between included studies since different clinical and radiological criteria were used.Furthermore, surgical technique varied between studies, and within the patient cohorts of two studies. 19,16We minimised this heterogeneity by clearly defining the acceptable surgical techniques in the inclusion criteria.Due to heterogeneity in follow-up time, there was an inadequate number of studies available to meta-analyse results at individual post-operative follow-up time points.In addition, there was a substantial lack of data with regards to walking ability and patient satisfaction in both comparisons.
Surgeons who operate on elderly patients with LSS and LGDS should be cognisant of the little benefit fusion provides for patients over a DA, as well as the increased risk of Coronal and lateral radiography, CT or CT myelography or MRI were used (criteria not specified).
t h e s u r g e o n 2 1 ( 2 0 2 3 ) e 2 3 ee 3 1 hospital complications and adverse events associated with fusion in this in age group.A thorough assessment of elderly patients with respect to their individual functional requirements as well as their comorbid conditions should be taken into account when justifying the addition of fusion in this age group.The clinical implication of the available evidence is substantially limited by low to moderate quality literature, therefore, further research is necessary to provide high quality evidence-based recommendations.This can be improved by standardising outcome measures and increasing follow-up length in the literature to allow both comparability of studies and address paucity of long-term follow-up data.

Conclusion
D þ F is not a superior intervention to DA for elderly patients with LSS and LGDS.Although DA was found to be associated with lower hospital complications and adverse events, surgeons should balance this with the increased risk of progression of DS post-operatively.Given the low to moderate quality of RCTs comparisons, higher quality RCTs are warranted to ascertain the most appropriate surgical approach in managing LSS with LGDS in elderly patients.

Funding
No funding received.

Declaration of competing interest
None declared.

Table 1 e
PICOS diagram for inclusion and exclusion criteria.

Table 3 e
Summary of results from meta-analysis comparing outcomes in patients with lumbar spinal stenosis (LSS) with low-grade degenerative spondylolisthesis (LGDS) who had decompression alone (DA) versus patients who had decompression with fusion (D þ F); mean difference (MD) for continuous variables, odds ratio (OR) for dichotomous variables and corresponding 95% confidence intervals (CI) reported.
18LF¼ Posterolateral fusion; VAS¼ Visual analogue scale; ODI¼ Oswestry Disability Index; DS ¼ Degenerative Spondylolisthesis. a Subgroup analysis reflects results from only one study.18

Table 5 e
Diagnostic criteria for case definition of Lumbar Spinal Stenosis (LSS) and Low-Grade Degenerative Spondylolisthesis (LGDS).Plain radiographs and imaging studies consisting of a myelogram and contrast-enhanced CT and MRI with LSS at the level of spondylolisthesis.Vertebral slippage exceeding 5% was considered to indicate LGDS.20 Based on history, clinical examination and CT myelography or MRI scan.Mid-saggital diameter of spinal canal of <11 mm was considered stenotic.Instability of <5 mm with rotational instability of <5 mm.
21Spinal claudication symptoms in all patients.