Surgery for chronic musculoskeletal pain: the question of evidence

Key points

1. 01Only 64 of more than 6734 published RCTs on these procedures actually compared the procedure to not performing it (less than 1%)
2. 02Of those 64 comparative trials, only 9 (14%) were favourable to surgery
3. 03None of the 12 blinded trials found surgery to be significantly better than no surgery
4. 04No suitable comparative trials exist for total hip replacement, total shoulder replacement, high tibial osteotomy, or excision of the outer clavicle end
5. 05Total knee replacement was the only procedure with a single supportive trial; all 11 knee arthroscopic meniscectomy trials were unfavourable

How it was conducted

Design

Systematic review counting RCTs; 14 common musculoskeletal surgical procedures identified from Australian hospital data, US board-certification data, and Medicare Australia 2018-2019 billing records

Procedures included

14 procedures: knee arthroscopic meniscectomy, knee arthroscopic debridement for osteoarthritis, total knee arthroplasty, total hip arthroplasty, subacromial decompression, clavicle/AC joint excision, rotator cuff repair, high tibial osteotomy, carpal tunnel decompression, disc replacement, lumbar spine fusion, cervical spine fusion, laminectomy for lumbar stenosis, ankle arthroscopy, total shoulder arthroplasty

Search source

Cochrane Central Register of Controlled Trials (CENTRAL) plus Cochrane Database of Systematic Reviews (CDSR); search completed January 2020

Inclusion criterion

RCTs comparing a listed procedure to a control group that did not receive the procedure (placebo surgery, nonoperative treatment, or no treatment)

Favourable definition

Primary outcome statistically significant (P < 0.05) AND met minimum clinically important difference (MCID) in favour of surgery

What they found

• 9485 total RCTs retrieved across all 14 procedures; after removing duplicates and non-publications, 6734 unique studies remained
• 64 of those 6734 RCTs (less than 1%) compared the procedure to not performing the procedure
• 9 of 64 included RCTs (14%) were favourable to surgery
• 12 of 64 included RCTs (19%) used patient blinding; 9 of those 12 (75%) were for subacromial decompression
• 0 of 12 blinded trials found a statistically significant and clinically important advantage for surgery
• Subacromial decompression: 15 comparative RCTs, 1 favourable, 9 blinded - no blinded trial was favourable
• Knee arthroscopic meniscectomy: 11 comparative RCTs, 0 favourable, 2 blinded
• Knee arthroscopic debridement for osteoarthritis: 8 comparative RCTs, 2 favourable (both the oldest trials), 1 blinded (not favourable)
• Lumbar spine fusion: 9 comparative RCTs, 2 favourable (one for adding fusion to decompression), 0 blinded
• Laminectomy for lumbar stenosis: 4 comparative RCTs, 0 favourable
• Total knee arthroplasty: 1 comparative RCT, 1 favourable (Skou et al., KOOS4 at 1 year)
• Carpal tunnel decompression: 8 comparative RCTs, 1 favourable, 0 blinded
• Total hip arthroplasty, total shoulder arthroplasty, high tibial osteotomy, clavicle excision: 0 comparative RCTs found

Limitations

• Not all procedures are performed solely for chronic pain; some (e.g., knee meniscectomy for mechanical locking, spine fusion for instability) have indications beyond pain, though the authors minimised this by restricting to chronic pain populations
• The list of 14 common procedures may be incomplete, so some common procedures are not represented
• Other risk-of-bias domains (sample size, allocation concealment, selective reporting) were not formally assessed for included trials; no quality weighting was applied
• Trial participants may not represent the full clinical population (e.g., one carpal tunnel trial excluded severe cases and patients with diabetes)

Why it matters

For patients

Most surgeries offered for chronic joint and back pain have not been proven better than non-surgical care in rigorous trials, so patients should ask their surgeon what evidence supports their planned procedure.

For clinicians

Clinicians should be aware that the majority of common musculoskeletal procedures for chronic pain lack comparative RCT support, and the minority that have been tested usually show no statistically or clinically meaningful benefit over non-operative management.

For readers

This review provides a comprehensive snapshot of how little high-quality comparative evidence underpins widespread surgical practice for chronic pain, supporting calls for mandatory pre-adoption RCTs with patient blinding.