Ten years later, subacromial decompression still doesn't outperform placebo surgery
This is one of the more consequential recent studies for how physical therapists explain shoulder pain.
Published in The BMJ in December 2025, the ten-year follow-up of the FIMPACT randomized trial examined arthroscopic subacromial decompression for persistent subacromial pain.
The study was particularly valuable because it included a placebo-surgery comparison.
Participants were assigned to:
Arthroscopic subacromial decompression.
Diagnostic arthroscopy without decompression.
Exercise therapy.
The original trial enrolled 210 adults, with 193 ultimately included in the three treatment groups after the surgical eligibility process.
The investigators obtained ten-year follow-up data from 168 participants.
The results
The researchers measured pain at rest and during arm activity on a 0–100 scale.
For pain during arm activity, the ten-year difference between decompression and placebo surgery was:
−3.2 points
95% confidence interval: −13.0 to +6.5.
The investigators had defined a 15-point difference as clinically important.
For pain at rest, the between-group difference was:
−1.5 points
95% confidence interval: −8.6 to +5.6.
There were no meaningful advantages for decompression in the primary comparisons.
All three groups improved substantially over time. BMJ
Why this challenges a familiar model
The traditional mechanical explanation is straightforward.
The acromion creates insufficient space.
Structures become compressed.
Compression produces pain.
Therefore, removing bone to create more space should improve symptoms.
The model is intuitively appealing.
It is also testable.
And this particular test did not support the predicted clinical benefit.
That does not prove mechanical loading is irrelevant to shoulder pain.
Nor does it prove that subacromial tissues never experience compression.
It demonstrates something narrower and more clinically important:
Removing the proposed mechanical obstruction did not produce better patient outcomes than placebo surgery.
That's a major problem for using that obstruction as the primary explanation for the syndrome.
What about the exercise group?
Decompression also failed to demonstrate a significant long-term advantage over exercise therapy.
For activity-related pain, the estimated difference was −9.4 points, with a 95% confidence interval extending from −19.0 to +0.3.
The exercise comparison was pragmatic and not blinded in the same way as the placebo-surgery comparison.
That makes the placebo comparison the more rigorous test of the procedure's specific effect.
Important limitations
This study does not address every painful shoulder condition.
Participants with full-thickness rotator cuff tears, substantial calcific deposits, significant arthritis, instability, and several other conditions were excluded.
There were also treatment crossovers during the decade of follow-up.
Thirteen participants initially assigned to placebo surgery eventually underwent decompression, which complicates interpretation of the long-term intention-to-treat comparison.
However, the authors' sensitivity analyses continued to support the primary findings.
Monday-morning clinical application
Consider a patient who presents with:
Painful arc.
Pain during resisted abduction.
Night pain.
Difficulty reaching overhead.
Imaging showing acromial morphology that someone has described as impingement.
You can acknowledge that the patient has painful shoulder structures without claiming that a particular bone shape is mechanically trapping tissue and must be corrected.
Instead, examine modifiable factors.
What loads provoke symptoms?
What loads are tolerated?
What is the patient's current strength and endurance?
How does their shoulder respond to progressive exposure?
What activities matter most to them?
You can still use mechanical reasoning.
But the mechanism should remain a hypothesis, not a story protected from contradictory evidence.