Shoulder pain and the rotator cuff: what the evidence says about exercise vs. surgery

Shoulder pain is one of the most common reasons people see a clinician. A review of 61 studies from around the world found a median community prevalence of 16%, with estimates generally higher in women than in men [1]. Rotator cuff problems, involving the group of muscles and tendons that keep the shoulder stable, are the most common source, accounting for more than two-thirds of cases [2].

When the pain does not go away, many people start to wonder about surgery. Some of the best research in orthopedics has tested exactly that question.

Surgery for shoulder impingement: tested against placebo

For years, a common operation for shoulder pain was subacromial decompression, which shaves bone and tissue to make more room for the rotator cuff. Two large trials compared it with a “placebo” operation: an arthroscopy in which the surgeon looks inside the shoulder but does not remove anything.

In the UK’s CSAW trial, 313 people with at least three months of shoulder pain and an intact rotator cuff were randomly assigned to decompression, placebo arthroscopy or no treatment [3]:

• No treatment29.4
• Decompression surgery32.7
• Placebo arthroscopy34.2

Source: Beard et al., CSAW trial, The Lancet, 2018 [3]. Bars are scaled to the full 48-point score.

Decompression offered no extra benefit over the placebo operation, and the small advantage of both surgical groups over no treatment was not clinically important [3]. In Finland’s FIMPACT trial, decompression also gave no benefit over placebo arthroscopy at two years, and its advantage over exercise therapy was smaller than the difference patients would notice [4].

A Cochrane review concluded with high certainty that decompression “does not provide clinically important benefits over placebo” [5]. A 2019 international guideline published in the BMJ made a strong recommendation against this surgery for people with long-standing shoulder pain from subacromial pain syndrome [6].

What to do instead

Guidelines point to non-surgical care first. The Dutch Orthopaedic Association’s guideline says subacromial pain syndrome “should preferably be treated non-operatively” and that exercise therapy should be specific, of low intensity and high frequency [7]. It also notes that a corticosteroid injection is an option for persistent or recurring symptoms, and it does not recommend strict immobilization or mobilization techniques [7]. In other words, the active part of the treatment, the exercise, is what carries the weight.

In practice, physical therapy for rotator cuff pain usually includes:

• An explanation of what is going on and how to adjust activities without stopping them
• Progressive strengthening of the rotator cuff and the muscles around the shoulder blade
• Exercises for control and range of motion
• A home program you can keep doing, with the load increasing as the shoulder tolerates it

The BMJ guideline is honest that there is still uncertainty about which non-surgical option is best [6], which is why the plan should be tailored to you and adjusted based on how you respond.

What about a torn rotator cuff?

Tears are a different question, and here the evidence is more nuanced.

In a Finnish trial of 180 shoulders in people over 55 with non-traumatic tears of the supraspinatus tendon, physical therapy alone did as well as surgery after one year [8]:

Group	Constant score at 1 year (0 to 100, higher is better)
Physical therapy alone	74.1
Acromioplasty plus physical therapy	77.2
Tendon repair plus acromioplasty plus physical therapy	77.9

The differences between groups were not statistically significant. Source: Kukkonen et al., 2014 [8].

At two years the results were still similar, and the authors called non-surgical care “a reasonable option for the primary initial treatment” of these tears in older patients [9].

Over the longer term, though, repair can pull ahead. In a Norwegian trial of small and medium tears, patients who had tendon repair had better scores for function and pain at 10 years than those treated with physical therapy, and those who switched to surgery later did worse than those repaired from the start [10]. The American Academy of Orthopaedic Surgeons’ 2025 guideline reflects both sides: physical therapy and surgery both lead to significant improvement for many tears, but without surgery a tear and the muscle around it may worsen over 5 to 10 years [11].

The bottom line: for impingement-type pain, surgery is rarely the answer. For a tear, especially a traumatic tear in a younger or active person, the decision deserves a conversation with an orthopedic surgeon, and physical therapy is part of the plan either way.

When to get checked right away

Most shoulder pain is not an emergency, but some situations need prompt medical attention [12]:

• Sudden or very bad pain, or pain that started after a fall or other injury
• You cannot move your arm, or the shoulder has changed shape or is badly swollen
• Pins and needles that do not go away, or no feeling in the arm
• The shoulder is hot to the touch, or you feel feverish or unwell

Call 911 if you have chest pain that spreads to your arm, neck or jaw. Shoulder or arm pain can be a sign of a heart attack [13].

The takeaway

• Rotator cuff problems are the most common cause of shoulder pain.
• Decompression surgery for impingement did no better than a placebo operation in high-quality trials.
• Guidelines recommend starting with non-surgical care, centered on specific, progressive exercise.
• For rotator cuff tears, physical therapy works well for many people, but repair can be the better long-term choice for some tears.

In Utah you can start physical therapy without a referral. If your shoulder has been bothering you, an evaluation can help sort out what is going on and what to try first.