Rotator Cuff Tendinopathy: Why It Is Not Tendinitis, and What Actually Works

If you have been told you have rotator cuff tendinitis, the name is probably wrong, and the name matters because it decides what gets prescribed.

When researchers take tissue from tendons that hurt like this and look at it under a microscope, the classic inflammatory cells are usually absent. What they find instead is disorganized collagen: thin, frayed, fragile fibers separated from each other lengthwise and disrupted across.

Alongside that, an apparent increase in tendon cells that have changed shape.

That is degeneration and failed repair, not inflammation. Which is why, as one review puts it plainly, anti-inflammatory drug treatment does not provide significant long term benefit in tendinopathy.

The naming problem, and why it is not pedantry

Three words get used for the same shoulder:

Tendinitis implies inflammation. Mostly inaccurate.
Tendinosis implies degeneration without inflammation. Closer, but describes only one stage.
Tendinopathy means a painful tendon without claiming a mechanism. This is the term most researchers now use.

Many clinicians have gone further and adopted rotator cuff related shoulder pain as an umbrella term, on the grounds that the specific structural label is often unachievable and frequently wrong.

There is a real split here you should know about.

The AAOS still defines this condition as inflammation of the rotator cuff tendons and of the bursa. UK and European physiotherapy and sports medicine sources have largely abandoned that framing. One NHS trust states outright that rotator cuff related shoulder pain replaces the terms bursitis, tendinitis and impingement.

One nuance, because accuracy matters more than a clean story. The leading tendon researchers do not claim inflammation is entirely absent. Inflammatory signalling molecules are elevated. What they argue is that the response is not a traditional inflammatory response, and that it reflects tendon cells reacting to mechanical load.

The safe summary: this is not the inflammation of a sprained ankle, there is no meaningful population of classic inflammatory cells, and anti-inflammatories do not fix it.

What is actually happening in the tendon

The most useful model describes a continuum with three stages, and the important point is that early stages are reversible and late ones are not.

Reactive

A short-term, non-inflammatory response to sudden overload. The tendon cells produce more of a water-binding substance, the tendon swells slightly and thickens to spread the stress. Collagen is still intact. Nerves and blood vessels are unchanged. Fully reversible with reduced load and time.

Dysrepair

An attempted healing response that is not going well. More matrix breakdown, more cells of the wrong type, and enough of the water-binding substance to start separating the collagen fibres from each other. The tendon’s load capacity drops.

Degenerative

Areas of cell death. Regions containing blood vessels, breakdown products and minimal collagen. Little capacity for reversal.

The two ideas worth taking from this

Pathology and pain are not the same thing. A pain-free tendon can contain substantial abnormality. This is why scans mislead.

Reactive-on-degenerative. A tendon can have a degenerative region that is mechanically silent and causes nothing, while the normal tissue next to it takes disproportionate load and flares up. This explains the very common story of a shoulder that was completely fine and then suddenly was not, after something fairly ordinary.

The model’s own authors note its limits: classifying a pain condition by structure is contested, the relationship between structure, pain and function is poorly understood, and imaging cannot detect everything the model needs.

Who gets it

From a systematic review and meta-analysis of 16 studies:

Age over 50 in a working population: 3.31 times the odds
Diabetes: 2.24 times the odds
Overhead work, meaning the shoulder above 90 degrees: 2.41 times the odds, though the studies varied a lot

Weaker, narrative-level evidence from the same review points to heavy manual work, repetitive work, high force exposure, swimming and weight training.

Two things worth flagging.

That review did not find smoking, body weight or genetics to be significant risk factors in the pooled analysis, despite all three being widely quoted. And the diabetes link is thought to work through changes in tendon collagen, which also means blood sugar control is not a side issue if you have diabetes and a painful tendon.

What it feels like

Pain over the outer shoulder and upper arm, related to activity. Worse reaching overhead, or behind your back. Worse at night, often badly enough to affect sleep. Pain reproduced on resisted movement. Often some weakness.

The feature that separates this from frozen shoulder: passive movement is preserved. If someone else moves your arm, it goes. If it does not go even then, the problem is a stiff joint capsule rather than a painful tendon, and the treatment is different.

Do I need a scan?

Usually not. NHS guidance states that rotator cuff related shoulder pain often does not require X-rays or scans.

The reason is the asymptomatic findings problem, and the numbers are striking:

54 percent of people over 60 with no shoulder symptoms have a partial or full thickness cuff tear
80 percent of asymptomatic people over 80
79 percent of professional baseball pitchers have abnormal findings in the throwing shoulder
40 percent of elite overhead athletes have partial or full thickness tears, none of whom reported symptoms
96 percent of asymptomatic middle-aged men had some shoulder abnormality
In one study, 55 percent of people with impingement symptoms had cuff pathology, versus 52 percent of people without symptoms

That last line is the one to sit with. The finding rate was almost identical whether the shoulder hurt or not.

The clinical tests are also weaker than their reputation. A systematic review concluded that none of Hawkins-Kennedy, Neer or the painful arc show high diagnostic ability, and that no single physical examination test can be relied on to make a definitive diagnosis. During the full can and empty can tests, eight or nine other muscles are equally active, so the idea that they isolate one tendon does not hold.

The reasonable use of a scan: when something is not adding up, when there is real weakness suggesting a significant tear, or when the question is whether to operate.

What actually works

Exercise

Exercise is the best supported treatment, and it is worth being honest about the size of the effect.

Against surgery, exercise holds up well. A well-constructed exercise approach confers at least equivalent benefit to that derived from surgery.

Against placebo, the margin is modest. In the one high-quality trial in the Cochrane review, manual therapy plus exercise improved function by 22.4 points against 15.6 for placebo, a 7.1 point advantage on a 100 point scale. Pain improved 24.8 against 17.3. Treatment success was 57 percent against 41 percent. Cochrane described the pain result as little or no different to placebo.

Anyone telling you exercise is definitively proven is overstating it. Anyone telling you it does not work is ignoring the comparison that matters, which is against the alternative interventions.

Does the type of exercise matter?

Less than the industry implies. A network meta-analysis compared seven exercise types and ranked concentric strengthening first for both dysfunction and pain, ahead of eccentric, traditional, motor control, scapular-focused, high-load and low-load approaches. But for pain, comparisons between the other types were largely non-significant.

Does the dose matter?

The evidence is low certainty and conflicting. One trial found 20 points better function on a 100 point scale with higher load and volume at three months, though the benefit did not extend to activity or night pain. Another found no benefit of higher load at six weeks.

Does it need to be supervised?

Apparently not. In a randomised trial of 86 patients, a self-managed single exercise was compared against usual physiotherapy. At three months, shoulder pain and disability scores were 32.4 and 30.7, a difference of 3.2 points that was not statistically significant.

Taken together: load the tendon, do it consistently for at least 12 weeks, progress it. No particular flavour has established superiority, and doing it yourself at home appears to work.

Cortisone injection

Genuinely helpful in the short term. Against placebo, pain and function both improve, with effect sizes around 0.68 and 0.62. The AAOS supports a single injection with local anaesthetic for short-term improvement.

The honest caveat: other treatment options are more effective in the intermediate and long term. The benefit washes out rather than lasting.

Two things worth knowing before a second or third injection.

You will read that steroid injections are proven harmful long term. That signal is strongest for tennis elbow, where significant negative effects favouring no treatment were found at intermediate and long-term follow up. For the rotator cuff, the same review found no significant difference from placebo at intermediate follow up. Benefit fading is not the same as harm.

At tissue level, though, the effects are real and dose-dependent. In human rotator cuff cells, a low dose of triamcinolone reduced cell viability to 73 percent of control at day 7 with full recovery by day 21. A high dose dropped viability to 21 percent at day 7 with no recovery by day 21. Re-injecting at 7 days left viability at 40 percent of control, at 14 days 56 percent, at 21 days 87 percent. The authors concluded an interval greater than three weeks was needed to re-administer safely.

There is also a case, made by some clinicians, that an anaesthetic-only injection produces comparable outcomes to steroid combinations up to six months and could reasonably be first line.

Shockwave therapy

A meta-analysis of 16 randomised trials covering 1,093 patients found significantly better pain and function scores, with a total effective rate 3.64 times control, and only mild transient side effects. Benefit appeared regardless of energy intensity or device type.

Important caveat: most of the strongly positive shockwave evidence historically sits in calcific tendinitis, which is a different problem. If you do not have a calcium deposit, the evidence is weaker than the headline suggests.

PRP

Contested. A meta-analysis found no significant pain difference at one and three months, a significant difference at 12 months, and better function scores on one measure but not another. The studies were highly inconsistent and showed signs of publication bias, and PRP preparation varied across trials.

Against that, the AAOS states that limited evidence does not support the routine use of platelet rich plasma for cuff tendinopathy or partial tears. Both positions are on the record.

Manual therapy

Supported as an addition to exercise rather than a treatment on its own. The Cochrane comparison above, which tested the combination against placebo, is the ceiling on what it delivers.

Therapeutic ultrasound: this one does not work

Stated plainly by the review evidence. Therapeutic ultrasound did not provide greater benefit than placebo or advice for pain or function, added nothing when combined with exercise, and was inferior to laser for pain. The conclusion was that clinicians should not use ultrasound therapy until higher quality evidence exists.

If it is on your treatment plan, it is worth asking why.

How much pain is acceptable during exercise?

This is the question patients ask most and get answered least clearly.

For a shoulder that is not highly irritable, pain in the order of 2 to 5 out of 10 during exercise is acceptable provided it settles soon afterward. NHS guidance agrees that it is normal for these exercises to provoke shoulder pain and safe to continue.

The widely used monitoring approach adds two more rules: pain should return to your baseline by the next morning, and it should not climb week on week.

For an irritable shoulder, the approach changes. Avoid band or equipment work done quickly under high load and repetition. Use isometric exercises with the arm supported, performed in the direction of pain, and gentle short-lever loading.

One principle worth adopting: change one variable at a time and judge it over 24 hours. People who change load, reps and frequency all at once have no idea which change caused the flare.

How long does it take?

NHS guidance says most people get better over a few months with activity modification and exercise. Published recommendations are for a comprehensive program of at least 12 weeks with progressive loading to fatigue.

And around a quarter to a third do not settle. In a five-year follow-up, 82 of 109 patients, or 75 percent, were pain-free at five years.

The sobering line from the same work: patients who do not get better with non-operative treatment do not get better with operative treatment either.

Predictors of a worse outcome were symptom duration over a year, living alone, no professional education, sick leave over two weeks before treatment, low work satisfaction, and AC joint degeneration. Notice how many of those are social rather than structural.

Should I have surgery?

For tendinopathy without a significant tear, the evidence says no.

The BMJ Rapid Recommendation makes a strong recommendation against subacromial decompression surgery for shoulder pain lasting more than three months, based on high quality evidence, stating that almost all informed patients would choose to avoid surgery because there is no benefit but there are harms.

A ten-year randomised trial comparing exercise plus acromioplasty against exercise alone in rotator cuff tendinopathy found both groups improved significantly and there were no significant differences between them. The authors do not recommend arthroscopic decompression for this condition. The detail on the placebo-controlled trials behind this is in our guide to shoulder impingement and subacromial pain syndrome.

One important boundary. All of that applies to tendinopathy and subacromial pain without a significant tear. It does not transfer to an acute traumatic full thickness tear in a younger person, where the calculation is different.

When to get checked urgently

Same day: fever with a hot swollen shoulder. Pain unrelated to activity and not relieved by rest with feeling systemically unwell. Chest pain, breathlessness or sweating with shoulder pain, which needs emergency assessment. Obvious deformity after trauma.

Within a week: a history of cancer with unremitting night pain, night sweats, or an unexplained mass.

Night pain is both a routine feature of tendinopathy and, in a different context, a red flag. The distinction is whether it comes with systemic symptoms or a relevant history.

Related conditions

Rotator cuff tears · Shoulder impingement · Subacromial bursitis · Cuff tear arthropathy

This article is for general education and is not medical advice. If your shoulder is not improving with loading over 12 weeks, or if you have significant weakness, see a physical therapist or doctor.