Lumbar Spine Surgery Overview

This article is still missing information.

This article reviews surgical options for chronic low back pain, radicular pain, and lumbar spinal stenosis Other lumbar spine surgery topics, such as scoliosis surgery are not discussed here. As a Musculoskeletal Medicine website, full discussion on areas such as surgical technique are out of scope.

Sham Controlled Trials in Orthopaedic Surgery

There are astoundingly few placebo controlled trials (i.e. sham studies) in all of orthopaedic surgery. All except the Clark study below found no difference compared to placebo surgery, but the Clark study was funded by industry and was only pain reduction in the very short term (44% vs 21% NRS <4/10 at 14 days). Indeed, one systematic review concluded that sham surgery was as effective as orthopaedic surgery for reducing pain and disability.

• Bradley 2002: tidal irrigation v placebo for knee OA - negative study
• Moseley 2002: lavage/debridement v placebo for knee OA - negative study
• Kallmes 2009: vertebroplasty v placebo - negative study
• Buchbinder 2009: vertebroplasty v placebo - negative study
• Clark 2016: vertebroplasty v placebo - positive study
• Firanescu 2018: vertebroplasty v placebo - negative study
• Kroslak 2013: debridement v placebo for lateral epicondylitis - negative study
• Sihvonen 2020: arthroscopic partial meniscectomy vs placebo for degenerative meniscus tears in the knee - negative study (5 year follow up data)
• Paavola 2021: subacromial decompression vs placebo for impingement - negative study (5 year follow up data)

The remaining RCTs in orthopaedic surgery are all non-placebo controlled, i.e. they compare surgery A to surgery B, or surgery A to an active control such as physiotherapy. This includes lumbar spine (and cervical spine) surgery, of which there are no placebo controlled trials. This limits what conclusions can be made. However we can still make some inferences despite these limitations, but the placebo effect from procedures is significant.

Radicular Pain

Sciatica lasting 4-12 months

• Bailey 2020: microdiscectomy superior to nonoperative care (incl. epidural injection) with surgery if needed for pain with 1 year follow up. (pain difference 2.4 scale 0 to 10). At 6 months, the score for leg-pain intensity was 2.8±0.4 in the surgical group and 5.2±0.4 in the nonsurgical group
• At 1 year, the leg-pain intensity score was 2.6±0.4 in the surgical group and 4.7±0.4 in the nonsurgical group; the ODI score was 22.9±2.3 vs 34.7±2.4,

Sciatica lasting 6 to 12 weeks

• Peul 2008: Surgery conferred more rapid relief of pain, but the benefit was not sustained at 6, 12, and 24 months. 20% reported unsatisfactory outcome at 2 years.
• Osterman 2006: Surgery more rapid initial recovery, but no difference at 6 weeks with 2 year follow up. Subgroup analysis discectomy superior if herniation at L4/5 (needs verification)

Sciatica lasting “at least 6 weeks”

• Weinstein 2006: ~80% pain less than 6 months. Major crossover issues. Intent-to-treat both improved. Advantage of surgery in as-treated analysis, worse outcome in post-hoc analysis if >6 months. Significant advantage for non-workers compensation group

Spinal Stenosis

Spinal Stenosis Surgery - Surgery vs non operative management

• Cochrane 2016: 5 RCTs. Studies low quality. “We have very little confidence to conclude whether surgical treatment or a conservative approach is better for lumbar spinal stenosis”.
• Weinstein 2008 (SPORT): n=289. Significant nonadherence and crossover. Surgery vs nonoperative. Intention to treat showed improvement in pain in surgery group but not disability. 2 year follow up. As treated analysis showed improvement in all outcomes in surgery group.
• Malmivaara 2007: n=94. Surgery (segmental decompression and an undercutting facetectomy) vs nonoperative. Surgery superior but benefits diminished overtime with both groups improving. 2 year follow up.
• Brown 2012: n=38, mild decompression vs epidural steroid. At 6 weeks epidural better for function (claudication questionnaire), but worse for VAS. Significant crossovers at 12 weeks preventing further analysis.

Spinal Stenosis - Decompression plus fusion vs decompression-alone

• Forsth 2016: n=247, decompression surgery equal to decompression plus fusion with 5 years of follow up. (pain and disability)
• Ghogawala 2016: n=66 with grade I spondylolisthesis, fusion better than decompression (“physical quality of life”) by year 2 (but not by year one). Nut no improvement in disability score.
• Cochrane 2016: Newer techniques are not superior to decompression. Didn’t include above two studies.
• Chang 2017: meta-analysis (15 studies) including the above newer studies, no benefit for decompression plus fusion over decompression alone at 2 years. Fusion has higher risk of complications. “We believe decompression alone to be a sound choice for LSS”

Fusion for Non-Specific CLBP

• Harris 2018:
  • Analysis of 33 meta-analyses mostly of 4 non-placebo controlled RCTs (!). Fritzell 2001, Brox 2003, Fairbank 2005, Brox 2006.
  • The most recent moderate quality review (Chou 2009): not more effective than intensive rehabilitation.
  • None of the post 2009 trials was of sufficiently high quality to change this conclusion
  • Complications are common at 16%
• Mannion 2013
  • Long term follow up (8-11 years) of the three classic RCTs Fritzel, Brox, Fairbank
  • Surgery no better than CBT and exercise both intention-to-treat and as-treated.
  • Findings mirror the two years results (Swedish study outlier but reporting bias and less effective non-operative treatment)
  • Just 50-60% of patients in each group reported their back was better/much better.
  • Just 40% had disability scores in the normal range
  • 25% crossover
  • Swedish trial the outlier, they reported positive outcomes at 2 years, but no difference at 10 years. But significant reporting bias using “global assessment” as the reported measure Non operative treatment was unstructured, nonspecific physiotherapy
• Good Editorial 2016 Spine Mannion
  • ITT analyses may be supported by analyses in which cases are grouped as “per protocol” (those who underwent the treatment to which they were assigned and completed all the follow-ups) or “as treated” (the treatment actually received) or “worst case” (group crossovers considered as “failing” the treatment to which they were randomized). These subgroup analyses present their own problems. Patients moving from non-operative to surgical treatment can be tracked with relative ease, but not those having surgery, failing it, and then going on to have successful non-operative treatment. Any subgrouping based on group changes in one direction only and on the assertion that a group change from non-operative to surgery indicates failure of non-operative treatment but revision surgery does not count as failure of surgery will inevitably be biased
• Mino 2017: Cohort study, 95% of those on opioids preoperatively continued on them postoperatively with 2 year follow up.
• Surgeries: posterolateral fusion in Norway, up to surgeon in UK studies
• There is no evidence that surgery undertaken without a valid diagnosis is effective for chronic low back pain
  • Brox 2003: Eligibility CLBP >1 year, and degeneration L4/5 and/or L5/S1 on XR.
  • Fritzell 2001: Eligibility severe CLBP >2 years, back>leg pain, clinician should interpret pain arising from L4/5 and/or L5/S1 using history/exam/radiology, major disability >1 year, degenerative change at above levels
  • Fairbank 2005: eligible if patient/surgeon uncertain which treatment best, >1 year CLBP
  • It is unclear whether discography is useful for patient selection Harris 2018

Lumbar Disc Replacement

• Hellum 2011
  • Lumbar disc prosthesis v rehabilitation, 2 year follow up
  • Statistically significant benefit for surgery
  • Disability: less than clinically significant benefit at 8 points on 0 to 100 scale.
  • Pain: 12 point difference favouring surgery.
  • only study comparing with rehab
• Cochrane 2013
  • Systematic review comparing disc replacement v fusion
  • No blinding and bias with sponsorships.
  • Statistically significant but not clinically significant benefit for replacement
  • Prevention of adjacent level disease and/or facet disease was not properly assessed.
• Not really done in NZ

Bottom Line

• There are no placebo controlled trials for any spinal surgery, below all low quality evidence
• Microdiscectomy for Sciatica
  • Possibly effective if pain 4 – 12 months [LOE 1c]
  • If done early then possibly more rapid recovery but no difference by 6 weeks.
• Spinal Stenosis Surgery
  • Possibly effective but benefits diminish over time
  • The addition of fusion to decompression is probably not necessary, and adds significant extra complications.
• Fusion for CLBP
  • No more effective than intensive rehabilitation
  • Possibly more effective than non-intensive rehabilitation, but benefits not sustained.
  • Disc Replacement for CLBP
  • Possibly statistically but no clinically significant benefit for surgery over non-intensive rehabilitation
  • No more effective than fusion

See Also

Lumbar Spine Arthrodesis