Shoulder Pain and the Rotator Cuff: Conservative Treatment or Surgery?

For decades the “impingement” theory dominated: the rotator cuff tendons were supposed to be mechanically squeezed under the acromion of the shoulder blade when raising the arm, and the recommended treatment was surgery to remove a piece of bone and soft tissue in the subacromial space (arthroscopic subacromial decompression). This operation was one of the most commonly performed on the shoulder. A series of sham-surgery trials then appeared that tested whether removing the tissue is what helps.

The second layer of explanation is the load model, the same as in other tendinopathies described in this series: a tendon hurts when the load exceeds its current capacity to bear load. Causes may be a sudden rise in the volume of work or training (more overhead exercises, a new sport, home renovation), prolonged holding of the arm raised, weakness of the shoulder blade and cuff muscles, and general factors — sleep, stress, diabetes, smoking. Pain itself is also modulated by the nervous system: it has a tissue dimension but also a psychosocial one, which explains why imaging results correlate poorly with symptoms.

The conclusion from CSAW is eloquent: a procedure that actually removed tissue gave no better result than one in which the key surgical element was omitted. This means that relief after surgery did not stem from mechanically “unblocking” a squeeze. Importantly, all participants had previously been treated conservatively without success, so the trial concerned precisely the group traditionally referred for surgery — and even in them decompression turned out no better than placebo.

The long-term results of the Finnish FIMPACT trial give a similar picture. Ten years after randomisation (193 people randomly assigned to decompression, sham surgery or exercise therapy entered the analysis; 168 completed follow-up) no differences were found in pain at rest or on arm activity between decompression and sham surgery — the difference was −1.5 points (95% CI −8.6 to 5.6) on a 0–100 scale for pain at rest and −3.2 points (−13.0 to 6.5) for pain on activity, with a minimal important difference of 15 points. Compared with exercise, too, no significant differences were observed. The authors stated that decompression offered no benefit over 10 years, either compared with sham surgery or with exercise therapy.

This result is sometimes misread as “exercise does not work”. In fact, in both arms of the trial patients had the care of a physiotherapist and advice on activity; the result says that a more intensive exercise programme was no better than one good session of education and guidance on exercises. The practical conclusion is that what is key is education (pain does not mean damage, movement is safe), moderate maintenance of activity and graded loading, not a specific, elaborate scheme. For most people with a new episode of pain, a sensible, well-explained plan is therefore enough.

Corticosteroid injections into the subacromial space are widely used because they can quickly reduce pain. A systematic review of randomised trials on injections in tendinopathies (Coombes et al., The Lancet 2010) indicated that the short-term efficacy of corticosteroids in rotator cuff tendinopathy is not clear, and at other sites the short-term analgesic effect reversed in the longer term. The GRASP trial, described above, showed no long-term benefit of a subacromial injection within 12 months.

In practice an injection may be a reasonable option when severe pain prevents exercise or sleep, but as a supplement to rehabilitation, not a substitute; it should not be expected to change the course of the disease. The decision on an injection, its frequency and possible risks (e.g. weakening of the tendon with repeated injections) is discussed with a doctor.

The question of surgery becomes truly relevant once a full-thickness tendon tear has been diagnosed. Here the evidence is less clear-cut than in subacromial pain. A 2019 Cochrane review covered nine trials (1007 participants); in the main comparison, surgery (tendon repair with subacromial decompression) versus non-operative treatment (exercises, possibly an injection) at one year probably gave little or no improvement in pain, the improvement in function was small (a mean of 6 points on a 0–100 scale, low certainty of evidence), and the proportion of people rating treatment as successful did not differ significantly.

The authors of the Cochrane review noted, however, important limitations: no trial had a sham-surgery group, all had risk of bias (lack of blinding), and the participants were mainly people with small, degenerative tears of the supraspinatus; the conclusions may not apply to traumatic injuries, large tears involving the subscapularis tendon or young people. The review also did not assess whether surgery prevents degenerative changes in the long term.

This study shows why one cannot answer “always exercise” or “always surgery”. At one year the differences between groups were small, but over the years — especially at 10 and 15 years — they grew in favour of tendon repair, and in unoperated tears enlargement of the damage was observed. At the same time the study has limitations: it concerns small and medium-sized tears, patients could not be blinded (nor was there a sham operation), and in the physiotherapy group some people opted for surgery in the meantime. The results do not translate to shoulder pain without a tear or to all types of damage.

On the basis of the studies discussed and generally accepted principles of management, conservative treatment is usually a sensible first step in shoulder pain without an acute injury. Surgery is considered where conservative treatment is known to have lower chances of success, or where it has failed. The list below is general, and the decision is always made by an orthopaedic surgeon after an examination.

If a doctor proposes subacromial decompression as the only procedure for shoulder pain with intact tendons, it is worth asking about the evidence, including the results of the CSAW and FIMPACT trials, and whether a sufficiently long and well-conducted non-operative treatment has been tried. This is a legitimate, partnership-style question within the process of shared decision-making.

For most people with shoulder pain without an acute injury, the most sensible path is time, education and graded loading — not a quick operation. The results of CSAW and FIMPACT are among the strongest examples in orthopaedics of an intuitively sensible procedure failing to withstand confrontation with a sham placebo. That is good news: you do not need a scalpel to improve subacromial pain, and improvement often comes through changing habits, sleep and consistent exercise.

A full-thickness tear is a separate story: here, especially in younger, active people or after an injury, it is worth consulting an orthopaedic surgeon early and discussing the pros and cons of both paths. The general importance of strength training for shoulders and bones is described in our knowledge-base entries on strength training and on sports recovery.