Rotator Cuff Tear: Partial and Full Thickness, and What the Evidence Actually Says

If you are reading this because a scan found a rotator cuff tear, start with one number. In a population screening study of 664 people with an average age of 69, 22 percent had a full thickness rotator cuff tear. Of those tears, 65 percent caused no symptoms whatsoever.

Over the age of 60, an asymptomatic tear is twice as common as a symptomatic one.

That does not mean your pain is imaginary. It means a tear on a scan is a finding to interpret, not a verdict to act on, and the decision about what to do next is more open than most people are told.

What the rotator cuff is, in plain terms

Four muscles wrap the shoulder blade and attach by tendons to the top of the arm bone. Supraspinatus starts lifting the arm. Infraspinatus and teres minor turn it outward. Subscapularis turns it inward.

Three of the four also pull the ball of the shoulder downward into its socket while you lift. That downward pull matters, and it explains a lot of what goes wrong later.

Partial versus full thickness

A partial thickness tear goes part way through the tendon. The tendon is still attached to bone.

A full thickness tear goes all the way through. The AAOS describes it as a hole in the tendon. It can still be small, involving part of the tendon’s footprint, or complete, where the tendon is detached.

Partial tears are also described by which side they are on. Articular sided means the inner surface, facing the joint. Bursal sided means the outer surface. A PASTA lesion is a partial articular sided supraspinatus tendon avulsion, which is the formal name for a specific type of articular sided tear.

This matters more than it sounds, because articular sided tears are more common in ageing shoulders, and the fact that most tears start on the inner surface is one of the strongest arguments against the old theory that bone rubbing on the outside of the tendon causes them.

How tears are measured

Surgeons classify full thickness tears by size: small under 1 cm, medium 1 to 3 cm, large 3 to 5 cm, massive over 5 cm.

Partial tears are graded by depth: grade 1 under 3 mm, grade 2 between 3 and 6 mm, which is around half the tendon thickness, and grade 3 over 6 mm.

Two other things get graded, and they matter more for your outlook than tear size does. Retraction describes how far the torn tendon has pulled back. Fatty infiltration, graded 0 to 4, describes how much of the muscle has turned to fat. Grades 3 and 4 carry a much lower chance of returning to normal function.

How common are they?

Far more common than most people realise, and mostly silent.

From a village screening study using ultrasound on both shoulders, full thickness tear prevalence by decade: none in the 20s to 40s, 10.7 percent in the 50s, 15.2 percent in the 60s, 26.5 percent in the 70s, 36.6 percent in the 80s.

From MRI studies of people with no symptoms at all: 13 percent of 50 to 59 year olds, 20 percent of 60 to 69 year olds, 31 percent of 70 to 79 year olds, and 51 percent of those over 80 have a tear.

A recent meta-analysis of 17 studies covering 5,988 people put overall tear prevalence at 38.6 percent and full thickness at 21 percent. In people who had never sought care for their shoulder, the rate was 23.6 percent. In people attending clinics with shoulder pain, 55.5 percent.

Look at that last pair. Half of people with a painful shoulder have a tear. A quarter of people with no complaint at all also have one. The tear is not a reliable explanation for the pain by itself.

Do silent tears stay silent?

Often, but not always. In prospective follow-up, 51 percent of asymptomatic tears became painful over a mean of 2.8 years.

Traumatic versus degenerative, and why it changes everything

Most rotator cuff tears are degenerative. The tendon wears down slowly over years, the blood supply to the tendon reduces with age, and at some point it fails. The odds of having a tear rise roughly 2.7 times per decade of age.

A minority are traumatic: a fall onto an outstretched arm, a sudden heavy lift, sometimes alongside a dislocation or fracture. Worth knowing: 40 percent of people over 40 who dislocate a shoulder also have a rotator cuff tear.

The distinction changes urgency. A degenerative tear in a 68 year old is managed by symptoms, and there is no clock running. An acute traumatic tear in a 45 year old is treated as time sensitive, for reasons covered below.

What a tear actually feels like

Pain at night and at rest. Pain lifting or rotating the arm. Weakness. Sometimes crackling or grinding. Gradual onset with degenerative tears, sudden with traumatic ones.

On night pain, one honest note. It is the most characteristic symptom of rotator cuff problems and the most commonly asked about. Cleveland Clinic states plainly that the exact cause remains unclear. You will find confident explanations online involving blood flow, pressure and sleeping position. Treat them as theories.

Weakness, and the difference that matters

Most weakness with a cuff tear is pain inhibition. It hurts, so the muscle does not fire properly. Take the pain away and much of the strength returns.

Two terms describe something different and more serious:

Pseudoparesis: active lifting of the arm reduced to under 90 degrees, with some lift preserved.

Pseudoparalysis: essentially no ability to lift the arm actively, often under 45 degrees, while someone else can still move it through full range passively.

The defining feature of both is that passive movement is preserved and there is no nerve injury. The arm can be lifted, just not by you. This is a mechanical failure of the muscles that balance the shoulder, not pain and not a trapped nerve, and it needs specialist assessment. When it becomes established, it can lead on to cuff tear arthropathy.

Diagnosis, and why scans mislead

Clinical tests are indicative rather than definitive. Reported figures: empty can test around 68 percent sensitivity; drop arm test 100 percent specificity, meaning a positive is meaningful but a negative proves little; Neer test around 75 percent overall accuracy. External rotation lag signs are most useful for large posterior tears. Combining tests works better than any one alone.

Ultrasound is very good for full thickness tears, at around 96 percent sensitivity and 93 percent specificity. It is much weaker for partial tears, at around 46 percent sensitivity, meaning it misses more than half of them.

MRI exceeds 90 percent for both sensitivity and specificity in full thickness tears, and runs 67 to 83 percent sensitivity for partial tears.

The AAOS guideline gives strong support for imaging as an adjunct to clinical examination. Note the word adjunct. Given the prevalence figures above, a tear found on a scan in a 70 year old may be a normal age-related finding that has been there for a decade. What matters is whether the imaging matches the clinical picture.

Do tears get bigger?

This is the question people most want answered, and there are real numbers.

For asymptomatic full thickness tears: around 22 percent enlarge by two years, around 50 percent by five years.

For partial thickness tears, progression is slower: around 11 percent at two years, around 35 percent at five years. Conversion from partial to full thickness has been reported anywhere from 10 to 50 percent.

People over 60 were twice as likely to progress to large or massive tears.

Something else progresses alongside the tear, and it is arguably more important. In a group managed without surgery, fatty degeneration of supraspinatus rose from 18 percent at the onset of pain to 41 percent at final follow-up. Infraspinatus went from 13 to 34 percent.

Fatty infiltration does not reverse after repair. It is one of the strongest predictors of a poor result from surgery. That is the real argument for not waiting indefinitely in someone who is likely to need repair eventually.

So do they heal?

Structurally, no. A torn tendon does not knit itself back to bone, and tears more often enlarge than shrink.

Symptomatically, frequently yes. And these two facts coexist. The best available review notes that many patients whose tears become more extensive do not experience worsening symptoms.

When a clinician tells you a tear will not heal and also tells you that you will probably get better without surgery, both statements are true. They are answering different questions.

Physical therapy: what the evidence shows

Better than most people expect.

The AAOS clinical practice guideline gives a strong recommendation that patient reported outcomes improve with physical therapy in symptomatic patients with full thickness tears, while noting that tear progression may occur over 5 to 10 years.

The MOON study is the landmark. It followed 452 patients with atraumatic full thickness tears through a structured home-based physical therapy program. Fewer than 25 percent elected surgery by 12 weeks. Around 25 percent had surgery by two years. The conclusion was that physical therapy is effective in roughly 75 percent of patients over two years, and the group reports that more than 75 percent continued to do well for at least seven years.

The AAOS patient resource puts non-surgical success at relieving pain and improving function in about 80 to 85 percent of patients.

Cortisone injection

The AAOS guideline gives moderate support for a single corticosteroid injection with local anaesthetic for short-term improvement in pain and function. Their patient material reports relief in roughly 67 percent of patients for at least three months.

The contested part is whether injections harm a future repair.

Some observational work suggests multiple injections, or injections within six months of surgery, may raise the re-tear rate. Other work found no association with repair failure. This is genuinely unresolved, and if surgery is on your horizon it is worth raising with your surgeon before accepting a third injection.

Surgery versus physical therapy: the honest picture

Here the evidence conflicts, and you deserve to see the conflict rather than one side of it.

The case that repair adds little

The Cochrane review of nine trials covering 1,007 participants found surgery improved pain by 0.87 points on a 0 to 10 scale, around a 9 percent improvement. Function was around 6 points better on a 0 to 100 scale, described as little or no improvement. Treatment success was 87 percent with exercise and 94 percent with surgery. Quality of life was essentially identical. The conclusion: surgery probably provides little or no clinically meaningful benefit.

The Finnish trial randomized 150 shoulders in patients with a mean age of 71 and small non-traumatic supraspinatus tears of around 10 mm, into physiotherapy, acromioplasty plus physiotherapy, or repair plus physiotherapy. At over five years there was no significant difference between the three groups. Satisfaction ran 88 to 92 percent across all arms.

The case that repair matters

The Moosmayer trial randomized 103 patients with tears up to 3 cm and followed them for ten years. Repair came out ahead: Constant score 80.5 versus 71.8, ASES score 94.0 versus 80.0, both statistically significant. Twenty-seven percent of the physiotherapy group crossed over to surgery later, and those late crossovers did worse than patients repaired first time. The advantage persisted at 15 years.

In the physiotherapy arm, unrepaired tears widened by around 10 mm on average, and the 41 percent whose tears grew by 10 mm or more had markedly worse function.

How to hold both

The trials enrolled different people. The Finnish study took older patients with very small tears, where repair adds little. Moosmayer took a younger cohort with larger tears and followed them long enough for tear progression to show up.

The reasonable reading: for a small degenerative tear in someone in their seventies, physical therapy first is well supported and surgery may add nothing. For a larger tear in someone in their fifties with a long life of shoulder use ahead, the case for repair is stronger, and delaying has a measurable cost.

The AAOS position sits between the two: for small to medium tears, both physical therapy and surgery produce significant improvement, and a repair that heals produces better outcomes than physical therapy alone or a repair that fails.

When surgery is urgent rather than elective

An acute traumatic tear, particularly in a younger person with real weakness, is treated differently. Specialist guidance recommends repair as early as possible, preferably within three months.

The mechanical reasoning is sound: less retraction, better tissue to work with, less fatty degeneration, and a better chance of restoring anatomy. In the available comparison, final functional scores were similar between early and late repair, but time to full recovery was around 14 months for early repair versus 34 months for late. For tears of 3 cm or more, 44 percent of late repairs needed a graft, against none of the early repairs.

Be aware that the certainty of this evidence is rated very low, based on 134 patients across three cohort studies with no standard definition of early. The rationale is better established than the trial data.

For degenerative tears, the AAOS suggests considering surgery after 6 to 12 months of failed non-surgical treatment, or for large tears with good tissue quality.

If you do have a repair, will it hold?

The UKUFF trial imaged 217 repaired shoulders at 12 months. Overall healing was 56 percent. By tear size: small 66 percent, medium 68 percent, large 47 percent, massive 27 percent.

Mean age of patients whose repairs healed was 61. Mean age of those whose repairs failed was 64. The authors concluded that age is the dominant risk factor for repair failure, independent of tear size.

Other reported figures: 87 percent healing for tears under 10 mm versus 62 percent for those over; 67 percent for single tendon tears versus 36 percent for multi-tendon tears. In patients over 65, a single tendon tear had around a 43 percent chance of healing.

Even a healed tendon reaches only about a fifth to a tenth of normal tissue material properties. High blood sugar levels correlate with higher re-tear rates.

If you are heading for surgery, our guide to rotator cuff repair recovery covers the sling, the phases and the precautions in detail.

When to seek urgent attention

Same day: a hot, swollen shoulder with fever, suggesting infection. An unreduced dislocation.

Within a week: unexplained lump or deformity, night sweats, unremitting night pain in someone with a history of cancer.

Urgent specialist referral: an acute rotator cuff tear after an injury, especially with weakness. Weakness turning the arm outward particularly suggests a large acute tear, and this is the situation where the three month window matters.

Related conditions

Rotator cuff tears sit on a spectrum with several other shoulder problems, and the labels overlap more than the textbooks suggest: rotator cuff tendinopathy, shoulder impingement or subacromial pain syndrome, subacromial bursitis, and at the severe end cuff tear arthropathy.

This article is for general education and is not medical advice. Decisions about rotator cuff surgery depend on your age, your tear, your tissue quality and what you need your shoulder to do. Discuss them with your surgeon or physical therapist.