Rotator Cuff Tear? When Surgery Helps — and When Physio Works Just as Well

"You've got a rotator cuff tear." For a lot of people that sentence lands like a verdict: something is torn, so someone will have to sew it back together.

The research tells a more interesting — and more hopeful — story. Whether surgery genuinely helps depends on which shoulder problem you actually have. Honest care means splitting them apart, because the evidence points in different directions for each.

First: "Tears" Are Common in Shoulders That Don't Hurt

Just like spinal imaging, shoulder MRI findings are common in people with no symptoms at all. In a classic MRI study of people with no shoulder pain (Sher et al., 1995), around a third had rotator cuff abnormalities — and among those over 60, more than a quarter had a full-thickness tear they didn't know existed.

Certainty: moderate — an older but repeatedly replicated finding; every subsequent imaging study of asymptomatic shoulders points the same way.

So a tear on your scan is a finding, not automatically the cause of your pain — and not automatically a surgical problem.

Situation 1: Shoulder Pain Without a Full-Thickness Tear

This is the most common presentation — often called subacromial or rotator-cuff-related shoulder pain. For decades the standard operation was subacromial decompression ("shaving the bone to make room").

Then the CSAW trial (Beard et al., 2018, The Lancet) compared that operation against placebo surgery — keyhole inspection with nothing actually removed. The result: decompression was no better than the placebo operation, and neither was meaningfully better than no treatment.

Certainty: high — a multicentre randomised placebo-controlled trial, the strongest design there is for a surgical question.

For this group, guidelines now point firmly at progressive, loaded exercise as first-line care. That's exactly the kind of structured, strength-led programme we build.

Situation 2: Gradual-Onset (Degenerative) Full-Thickness Tears

What about a genuine full-thickness tear that came on gradually with age rather than an injury? A Finnish randomised trial (Kukkonen et al., 2014) compared physiotherapy alone against surgical repair for non-traumatic supraspinatus tears in older adults — and found no meaningful difference in outcomes at one year.

Certainty: moderate — a well-run randomised trial, though in a specific group (older adults, smaller degenerative tears); results shouldn't be stretched to every tear.

Situation 3: Acute Tears and Younger, Active Shoulders — Where Surgery Earns Its Place

Here's the honest other side. A Norwegian randomised trial (Moosmayer et al.) followed people with small-to-medium full-thickness tears for a decade. At 10 years, the group who had primary tendon repair had better shoulder scores, less pain, and higher satisfaction than the physiotherapy-first group — moderate differences, but real ones. Notably, about a third of the physio-first group later chose surgery along the way.

Certainty: moderate-to-high — one high-quality long-term randomised trial, with the direction maintained at its 15-year follow-up.

And for acute, traumatic tears — a fall or wrench followed by sudden weakness, especially in a younger or highly active person — early surgical opinion is widely recommended, because those tears behave differently from degenerative ones.

Putting It Together

The Bottom Line

The question isn't "is a tear there?" — it's "which problem is this, and what does the evidence say helps that problem?" For most shoulder pain, a properly progressed loading programme is the right starting point, with surgery kept for the situations where it genuinely outperforms rehab. A good clinician will be straight with you about which group you're in — and will change the plan if your shoulder isn't tracking.