Is lumbar fusion necessary for chronic low back pain associated with degenerative disk disease? A meta-analysis

Key points

1. 01Fusion and non-operative care produced similar Oswestry Disability Index (ODI) and VAS pain scores at both short-term (up to 2 years) and long-term (4+ years) follow-up
2. 02At long-term follow-up, the fusion group had a statistically significantly higher complication rate than the non-operative group
3. 03At long-term follow-up, the non-operative group had a statistically significantly higher rate of additional surgeries compared with the fusion group
4. 04No significant between-group differences were found for any outcome at short-term follow-up
5. 05Existing RCTs had lax patient inclusion criteria, so results should be interpreted cautiously pending higher-quality trials

How it was conducted

Design

Systematic review and meta-analysis of prospective studies

Databases searched

PubMed, EMBASE, Cochrane Library, China National Knowledge Infrastructure, up to June 2020

Studies included

6 prospective studies (including RCTs); 3 RCTs for short-term subgroup and up to 4 for long-term subgroup

Population

Adults older than 18 years with chronic low back pain (greater than 1 year) and confirmed degenerative disc disease on plain radiography, CT, or MRI

Comparison

Lumbar fusion (any technique) vs. non-operative treatment (cognitive intervention and exercise programs)

Primary outcomes

ODI, VAS scores for back and leg pain, complication rate, and all-cause additional surgery rate at short-term (discharge to 2 years) and long-term (4+ years) follow-up

What they found

• Short-term ODI: no significant difference between fusion and non-operative groups (random-effects model)
• Short-term VAS back pain: no significant difference between groups
• Short-term VAS leg pain: no significant difference between groups
• Short-term complication rate: no significant difference between groups (fixed-effects model)
• Short-term additional surgery rate: no significant difference between groups
• Long-term ODI: no significant difference between groups (random-effects model)
• Long-term VAS back pain: no significant difference between groups
• Long-term VAS leg pain: no significant difference between groups
• Long-term complication rate: significantly higher in the fusion group compared with the non-operative group (fixed-effects model, P < 0.05)
• Long-term additional surgery rate: significantly higher in the non-operative group compared with the fusion group (random-effects model, P < 0.05)
• Sensitivity analyses did not alter any of the above conclusions

Limitations

• Blinding of patients and providers was not feasible due to the nature of surgical vs. non-surgical interventions, resulting in high performance bias in half the included studies
• All included RCTs had high risk of detection bias because outcomes such as pain are participant-reported and can be influenced by knowledge of treatment received
• There was no consensus diagnostic standard for discogenic low back pain across included studies; lax inclusion criteria may have enrolled heterogeneous patient populations
• Only six prospective studies were available, limiting statistical power and precluding publication bias assessment; sample sizes in individual trials were small

Why it matters

For patients

Patients with chronic low back pain from disc degeneration should know that fusion surgery does not reliably reduce pain or disability more than structured non-operative programs, but does carry a higher risk of complications.

For clinicians

Lumbar fusion should not be offered as a routine first-line treatment for this indication; reserve it for carefully selected patients and weigh the long-term complication burden against the potential reduction in need for additional surgeries.

For readers

This meta-analysis adds long-term follow-up data to an ongoing debate, but the field still needs large, well-designed RCTs with rigorous diagnostic criteria before definitive recommendations can be made.