Some conservative interventions are more effective than others for people with chronic non-specific neck pain: a systematic review and network meta-analysis

Key points

1. 01A network meta-analysis combines direct and indirect comparisons so many treatments can be ranked against each other even when they were never tested head to head.
2. 02Multimodal non-pharmacological combinations (for example exercise plus manual therapy, or exercise plus kinesio taping) were the most effective options for pain and disability at short and intermediate follow-up.
3. 03Anti-inflammatory drugs (NSAIDs), alone or combined with another treatment, did not beat inert treatment for pain at 1 month or 3 to 6 months.
4. 04At 12 months no intervention was clearly better than inert treatment, and on disability several conservative treatments were actually worse than inert care.
5. 05Certainty of evidence was low to very low throughout, driven by small trials, indirect evidence, imprecision, and reporting bias, so the rankings carry real uncertainty.

How it was conducted

Design

Systematic review and network meta-analysis of RCTs (PROSPERO CRD42019124501), GRADE via CINeMA

Search

Six databases from inception to May 2021; 119 RCTs included (12,496 patients, 277 arms, 32 interventions), median 24 participants per arm

Intervention

All conservative pharmacological and non-pharmacological treatments (alone or paired) versus inert treatment in chronic neck pain of 3 months or more

Outcomes

Pain intensity and disability at 1 month, 3 to 6 months, and 12 months, plus adverse events; SMDs back-translated to 0-to-10 (pain) and 0-to-100 (disability) scales

Analysis

Random-effects frequentist network meta-analysis (Stata) with SUCRA rankings, transitivity and inconsistency checks, sensitivity analyses

What they found

• Pain at 1 month: multimodal non-pharmacological treatments effective versus inert (MD range 0.84 to 3.74 on 0-to-10 scale); manual therapy SMD -0.42 (95% CI -0.82 to -0.01); top ranked was exercise with kinesio taping (SUCRA 93.2%).
• Pain at 3 to 6 months: education with exercise, mind-body practices, and physical agents effective versus inert (SMD range 0.53 to 0.75).
• Disability at 3 to 6 months: five treatments significant versus inert (SMD range 0.50 to 1.47); top ranked acupuncture or dry needling with exercise (SUCRA 88.2%).
• NSAIDs with or without another treatment did not reduce pain versus inert at 1 month or 3 to 6 months; at 12 months no treatment was superior to inert, and inert was significantly better than four conservative treatments on disability.
• Risk of bias low in 50.4%, unclear in 22.7%, high in 26.9% of trials; most mild adverse events occurred with acupuncture or dry needling (1,288 events in 2,353 participants); no serious events; overall certainty low to very low.

Limitations

• Overall certainty of evidence was low to very low, mainly from indirect comparisons, imprecision (small nodes, wide CIs), and reporting bias.
• Disability at 1 month and pain at 3 to 6 months showed inconsistency, so only pairwise (not network) estimates could be reported there.
• Trials were small (median 24 per arm) and mostly single-centre (90.8%), and many did not report effect modifiers needed to test transitivity.
• No-treatment, waitlist, sham, and placebo were combined into one inert node, which may blur differences between control types.

Why it matters

For patients

For long-standing neck pain, combining active and hands-on treatments like exercise and manual therapy may help more than pills, though benefits are modest and not certain.

For clinicians

Favor multimodal non-pharmacological care over NSAIDs for chronic neck pain, but interpret the specific rankings cautiously given low certainty and weigh acupuncture's higher rate of mild side effects.

For readers

This is the largest comparison of conservative neck pain treatments to date, yet weak underlying trials mean it points to a direction rather than a definitive winner.