Real versus sham manual therapy in addition to therapeutic exercise in the treatment of non-specific shoulder pain: a randomized controlled trial

Key points

1. 01All three groups improved over time, but there were no significant differences between the real and sham manual therapy groups for pain, disability or any direction of shoulder motion.
2. 02The benefits appeared to come from the therapeutic exercise program, which everyone received, not from the added manual therapy.
3. 03Pain and disability gains kept growing over follow-up, with the largest improvements at 12 weeks.
4. 04This was a small trial (15 people per group), so a null between-group result is weak evidence that manual therapy has no effect.
5. 05Findings apply only to the two specific manual therapy techniques used here; other manual therapy methods were not tested.

How it was conducted

Design

Evaluator-blinded, allocation-concealed randomized controlled trial with intention-to-treat analysis

Participants

Adults 18 to 60 years with unilateral non-traumatic shoulder pain lasting more than 3 months

Groups

Real manual therapy plus exercise; real shoulder mobilization with sham thoracic technique plus exercise; full sham manual therapy plus exercise

Intervention

Two sessions per week for five weeks; all groups did a progressive isometric exercise program

Primary outcome

Pain intensity on a 0 to 10 visual analog scale (VAS), measured post-treatment and at 4 and 12 weeks

Other outcomes

Shoulder Pain and Disability Index (SPADI) and pain-free active range of motion

What they found

• No significant between-group main effect and no significant time-by-group interaction for pain or disability, meaning real and sham manual therapy did not differ.
• No significant between-group main effect and no significant time-by-group interaction for any range of motion (flexion, extension, abduction, internal and external rotation).
• Significant improvement over time for pain (F = 24.33; p < 0.01; partial eta squared = 0.65) and disability (F = 33.82; p < 0.01; partial eta squared = 0.72) across all groups combined.
• Significant time effects across all groups for range of motion: flexion (F = 25.82; partial eta squared = 0.66), extension (F = 7.25; partial eta squared = 0.15), abduction (F = 14.67; partial eta squared = 0.26), internal rotation (F = 5.12; partial eta squared = 0.12) and external rotation (F = 15.43; partial eta squared = 0.27), all p < 0.01.
• Pooled across all groups, pain fell by 2.56 cm on the VAS at 12 weeks (95% CI -3.38, -1.73).
• Pooled across all groups, SPADI disability fell by 18.73 points at 12 weeks (95% CI -25.01, -12.45).

Limitations

• Only 15 participants per group, so the trial was likely underpowered to detect a real difference between manual therapy and sham.
• No no-treatment control group and blinding success was not tested, so natural recovery or placebo effects cannot be ruled out.
• The two manual therapy techniques used had not been tested in prior randomized trials, making comparison with other research difficult.
• The sample was a convenience group from one university, and the variability of conditions causing non-specific shoulder pain was not accounted for, so results may not generalize.

Why it matters

For patients

If you have long-standing non-specific shoulder pain, a structured exercise program is the likely driver of improvement, and these specific hands-on techniques may not add extra benefit.

For clinicians

Adding these particular glenohumeral and rib-cage mobilizations to exercise did not improve outcomes over sham in this small trial, though the null result is underpowered and other manual therapy methods remain untested.

For readers

This single small RCT suggests exercise, not the added manual therapy techniques studied, explains the improvements, but it is too small to firmly conclude manual therapy has no value.