Active and passive physical therapy in patients with chronic low back pain: a level I Bayesian network meta-analysis

Key points

1. 01Active physiotherapy achieved the lowest pain scores (SMD 1.00) compared with passive and combined approaches.
2. 02Active physiotherapy also produced the lowest Roland-Morris Disability Questionnaire scores (SMD 0.94) and the lowest Oswestry Disability Index scores (SMD -1.23).
3. 03All three treatment groups were well matched at baseline for age, sex, BMI, symptom duration, and outcome scores.
4. 04Passive physiotherapy still improved pain and disability compared with no treatment, but to a lesser degree than active exercise.
5. 05The network was statistically consistent across all three outcomes (global Wald test p values 0.2-0.8), supporting the reliability of indirect comparisons.

How it was conducted

Design

Bayesian hierarchical random-effects network meta-analysis of RCTs (Level I evidence)

Databases searched

PubMed, Web of Science, Google Scholar, Embase (searched June 2025, no date restriction)

Included studies

29 RCTs

Participants

2768 patients with chronic (>=3 months) mechanical or non-specific LBP; mean age 46.9 years, mean BMI 25.8 kg/m2

Groups compared

Active physiotherapy (n=1271) vs passive physiotherapy (n=912) vs combined active and passive (n=521)

Primary outcomes

Pain (VAS/NRS), Roland-Morris Disability Questionnaire (RMQ), Oswestry Disability Index (ODI); mean follow-up 6.2 months

What they found

• Active physiotherapy had the lowest pain scores: SMD 1.00 (95% CI -3.28 to 5.28); global Wald test PWald = 0.8, no significant inconsistency.
• Active physiotherapy had the lowest RMQ score: SMD 0.94 (95% CI -4.96 to 3.09); PWald = 0.2, no significant inconsistency.
• Active physiotherapy had the lowest ODI score: SMD -1.23 (95% CI -9.83 to 7.36); PWald = 0.6, no significant inconsistency.
• Baseline pain scores were comparable across groups: active 5.3+/-1.1, passive 5.8+/-0.9, combined 5.3+/-1.3 (p=0.3).
• Baseline RMQ scores were comparable: active 10.4+/-2.9, passive 10.0+/-2.8, combined 10.2+/-2.9 (p=0.9).
• Baseline ODI scores were comparable: active 35.4+/-10.7, passive 30.8+/-6.3, combined 37.8+/-16.9 (p=0.5).

Limitations

• Included trials differed substantially in the specific type of active and passive physiotherapy used and in therapeutic regimens, limiting sub-group analysis.
• Most studies had short follow-up (mean 6.2 months); long-term effects and recurrence rates are unknown.
• Blinding of participants and personnel was not possible in most trials, resulting in moderate performance and detection bias.
• Concomitant pharmacological therapy varied across studies and could not be controlled for in the analysis.

Why it matters

For patients

People with chronic low back pain are likely to get better and more lasting relief from engaging in active exercise-based physiotherapy than from passive treatments alone.

For clinicians

When prescribing physiotherapy for mechanical or non-specific chronic LBP, prioritising active exercise programmes over passive modalities is supported by Level I network meta-analytic evidence.

For readers

This Bayesian network meta-analysis provides the most direct head-to-head comparison of active, passive, and combined physiotherapy for chronic LBP published to date, though the wide confidence intervals suggest further high-quality RCTs are still needed.