Ten years on, shoulder decompression still had not beaten placebo surgery

A decade after surgery, patients who had their shoulder decompressed arthroscopically were no better off than patients who were put under anaesthetic, had a camera inserted and nothing done [s1]. The difference between the two groups on a 100-point pain scale was 1.5 points at rest and 3.2 points on arm activity — against a threshold of 15 points that the trial had set in advance as the smallest difference that would matter [s1].

This is the strongest form of evidence a surgical procedure can be subjected to, and the reason it is rare: a placebo-controlled operation requires anaesthetising healthy-enough people, cutting them, and doing nothing. Where it has been done, the results have been consistent enough to change what several common orthopaedic operations are for.

FIMPACT

The trial ran in the orthopaedic departments of three public hospitals in Finland. It enrolled 210 adults aged 35 to 65 with symptoms consistent with subacromial pain syndrome for more than three months, beginning on 1 February 2005, and followed them to 20 September 2023 [s1]. Participants were randomised 1:1:1 to arthroscopic subacromial decompression (59 participants), placebo surgery in the form of diagnostic arthroscopy (63), or exercise therapy as a pragmatic comparator (71) [s1]. Participants and outcome assessors were blinded in the surgery-versus-placebo comparison, which is the part of the design that carries the weight [s1].

The primary outcomes were shoulder pain at rest and on arm activity at ten years, each on a 0–100 visual analogue scale, with the minimally important difference set at 15 points [s1]. A total of 168 participants (87%) completed the ten-year follow-up [s1].

In the intention-to-treat comparison of decompression against placebo surgery, the mean difference was −1.5 points (95% CI −8.6 to 5.6) for pain at rest and −3.2 (−13.0 to 6.5) for pain on arm activity [s1]. Against exercise therapy, the differences were −4.0 (−11.0 to 3.0) and −9.4 (−19.0 to 0.3) [s1]. No significant differences appeared in any secondary outcome or in adverse events [s1]. The conclusion the authors draw is that the procedure offered no benefit over placebo surgery or exercise therapy over ten years [s1].

Note what the confidence intervals allow. The activity-pain interval against exercise reaches −19.0, so a moderate advantage for surgery over exercise cannot be excluded — but the placebo-controlled comparison, the one that isolates the operation itself, brackets zero comfortably in both directions.

The knee version

The same question was put to arthroscopic partial meniscectomy, the operation performed for degenerative meniscal tears. OMEX randomised 140 participants with a degenerative meniscal tear and no or minimal radiographic osteoarthritis, 1:1, to surgery or twelve weeks of exercise therapy [s2]. Its ten-year primary outcome was structural rather than symptomatic: progression of radiographic knee osteoarthritis on the OARSI atlas sum score [s2].

The adjusted mean difference in change was 0.39 (95% CI −0.19 to 0.97), with more progression in the surgery group [s2]. Radiographic knee osteoarthritis developed in 23% of the surgery group and 20% of the exercise group, an adjusted risk difference of 3% (−13% to 19%) [s2]. No clinically relevant differences were found in patient-reported pain and knee function or in isokinetic muscle strength, and both treatments were associated with improvement in pain and function over the decade [s2].

OMEX used exercise rather than sham surgery as its comparator, which means it cannot separate the operation from the ritual around it. What it can say is that after ten years the two paths ended in the same place.

A discrepancy in the pooled analysis

A meta-analysis of six randomised trials covering 1,078 participants reports that exercise therapy carried a lower risk of knee osteoarthritis progression than arthroscopic partial meniscectomy, with a risk ratio of 1.27 (95% CI 1.05 to 1.53), and no statistically significant differences on any of the five KOOS subscales — pain, symptoms, activities of daily living, sport and recreation, or quality of life [s3].

A risk ratio above 1 accompanied by a claim of lower risk in the group being described means the ratio must be oriented surgery-versus-exercise while the sentence is written exercise-first. The abstract does not state the reference group, so the direction has to be inferred from the conclusion rather than read off the number. We report it as printed and note that the orientation is not specified in the source.

Why placebo-controlled surgery matters more here than elsewhere

Surgery is a large intervention with a large ritual attached: a consultation, a decision, a theatre, an anaesthetic, a recovery, an expectation of improvement. Pain is the outcome most responsive to all of that. An uncontrolled surgical series in which patients report less pain afterwards is close to uninformative, because patients would be expected to report less pain afterwards even if nothing had been done to the joint — which is precisely what the FIMPACT placebo arm demonstrates [s1].

That does not make the improvement fake. FIMPACT's placebo group genuinely felt better; so did OMEX's exercise group [s1] [s2]. It makes the attribution wrong. Deciding whether to operate depends on whether the operation is doing the work, and only a sham-controlled trial can answer that.

What to watch

Sham-controlled orthopaedic trials remain scarce, and the two conditions with the strongest such evidence — subacromial pain syndrome and degenerative meniscal tear — are among the most commonly operated. Whether decade-scale follow-ups like these translate into changes in operating volumes is the measurable question, and it is a question about health systems rather than about the evidence, which is now unusually clear.

This article is informational and is not medical advice. Decisions about surgery belong with a reader and their clinician.

Sources

• Arthroscopic subacromial decompression versus placebo surgery for subacromial pain syndrome: 10 year follow-up of the FIMPACT randomised, placebo surgery controlled trial — BMJ, 2025-12-02
• Arthroscopic partial meniscectomy versus exercise therapy for degenerative meniscal tears: 10-year follow-up of the OMEX randomised controlled trial — British Journal of Sports Medicine, 2024-09-26
• Long-term effects of exercise therapy versus arthroscopic partial meniscectomy for degenerative meniscal tear: A meta-analysis of randomized controlled trials — Asian Journal of Surgery, 2024-03-25