The effects of exercise dosage on neck-related pain and disability: a systematic review with meta-analysis

Key points

1. 01Exercise therapy produced clinically meaningful reductions in pain intensity compared to control at intermediate and long-term follow-up
2. 02Exercise produced small, statistically significant reductions in disability (NDI) overall, but effects were not clinically meaningful
3. 03Neither minutes of exercise per week nor total weeks of exercise predicted pain or disability outcomes in meta-regression
4. 04Wide variation in exercise type, dosage, and reporting across trials makes it impossible to identify an optimal prescription
5. 05Most included trials had high or unclear risk of bias and inconsistent adherence reporting

How it was conducted

Design

Systematic review with meta-analysis of randomised and controlled trials

Databases searched

PubMed, PEDro, SPORTDiscus, CINAHL, Scopus, Cochrane CENTRAL; searches completed through November 26, 2018

Included trials

14 trials included in quantitative synthesis

Participants

Combined sample of 1708 participants across included trials; sample sizes ranged from 15 to 449 per trial

Populations

Adults with acute, subacute, or chronic neck-related disorders receiving supervised exercise therapy versus a non-active control

Primary outcomes

Pain intensity (0-100 VAS) and disability (Neck Disability Index 0-50 or Neck Pain and Disability scale); exercise dosage operationally defined as total supervised exercise minutes over the care duration

What they found

• Overall pain intensity (intermediate to long-term follow-up, 13 trials, n=1080): mean difference -15.32 mm (95% CI: -19.20, -11.44), clinically meaningful, moderate-certainty evidence
• Intermediate-term pain (4-9 weeks, 5 trials, n=242): mean difference -13.36 mm (95% CI: -18.52, -8.24), clinically meaningful, low-certainty evidence
• Long-term pain (10+ weeks, 8 trials, n=838): mean difference -15.51 mm (95% CI: -20.28, -10.74), clinically meaningful, low-certainty evidence
• Overall disability NDI (7 trials, n=577): mean difference -3.64 points (95% CI: -6.19, -1.09), small effect not clinically meaningful, low-certainty evidence
• Intermediate-term disability NDI (4 trials, n=199): mean difference -2.50 points (95% CI: -4.80, -0.20), small effect, low-certainty evidence
• Long-term disability NDI (3 trials, n=378): mean difference -4.23 points (95% CI: -8.55, 0.29), no statistically significant difference, very low-certainty evidence
• Long-term disability NPAD (1 trial, 2 groups, n=101): mean difference -2.58 points (95% CI: -7.24, 2.08), no difference, moderate-certainty evidence
• Meta-regression - minutes of exercise per week: regression coefficient 0.00 (95% CI: 0.00, 0.00), p not significant for pain outcomes
• Meta-regression - total weeks of exercise: regression coefficient 0.00 (95% CI: -0.05, 0.04), p not significant for pain outcomes
• Risk of bias: 7 trials high risk, 4 unclear risk, 3 low risk; interrater reliability kappa=0.84 (95% CI: 0.71, 0.96)
• Total exercise dosage across included trials ranged from 70 to 1800 minutes; exercise frequency ranged from 1 to 7 days per week

Limitations

• Seven of 14 trials had high overall risk of bias, primarily due to unclear allocation concealment and incomplete outcome data
• Heterogeneity in exercise types, dosage parameters, and outcomes reported made it difficult to isolate the effect of dosage from exercise type
• Exercise intensity was rarely reported, preventing a full analysis of dosage effects
• Only 2 studies had both low risk of bias and adherence rates above 75%, making exercise adherence a major gap across trials

Why it matters

For patients

Exercise is effective for reducing neck pain and disability, so engaging in a supervised exercise program is worthwhile regardless of whether it is brief or prolonged.

For clinicians

Clinicians can confidently prescribe exercise for mechanical neck pain, but should tailor the prescription to the individual patient's presentation rather than assuming more exercise minutes will produce proportionally better results.

For readers

This review confirms exercise benefits for neck pain but highlights a critical gap: current trial reporting is too inconsistent to determine an optimal dosage, underscoring the need for standardised reporting using CERT and TIDieR guidelines.