Rotator Cuff Pain — Body Handbook

Musculoskeletal · §165

Rotator Cuff Pain

A dull ache deep in the side of your shoulder. A sharp catch when you reach for the seatbelt or the top shelf. Pain that wakes you up the second you roll onto that side. That cluster is what doctors used to call subacromial impingement and now increasingly call rotator cuff related shoulder pain — the most common shoulder problem in adults over forty, and the one where the gap between what the high-quality evidence says and what gets done in clinics is widest. Three things to know up front: an MRI showing a torn rotator cuff is not a diagnosis, the standard "clean it up" surgery does not beat a fake surgery, and a structured twelve-week exercise program clears most of it.

Know · As-needed Evidence Strong Chapter Musculoskeletal

The pain is real and the sleep loss it brings is the part people actually come in for — rolling onto the bad side wakes you up, every night, for months. The fix that works is sustained and unsexy: daily exercises, supervised every other week, for about three months. The fix that doesn't work — but gets sold hard — is the surgery to shave a bit of bone off the top of the joint; two big fake-surgery trials and a Cochrane review have now closed that question. The catch is the effort. Three months of doing the homework through some discomfort is what separates the people who recover from the people who end up booking an operation that the literature says to skip.

The rotator cuff is four small muscles whose tendons wrap the head of the upper arm bone where it meets the socket. They are what let you lift your arm sideways, reach behind your back, and keep the joint stable when you carry a bag of groceries. The tendon that gets sick first and most reliably is supraspinatus — the one running across the top of the shoulder under a small bony shelf called the acromion.

The old story — the one most clinic websites still tell — is that the tendon gets pinched between the bone above and the upper-arm bone below every time you lift your arm. Years of pinching frays the tendon, inflames the small sac of fluid sitting between them, and eventually tears something. Shave the bone off the top, the story goes, and the pinching stops. This is what "impingement" meant.

The current story is more honest. Painful rotator cuff tendons are mostly degenerating, not getting mechanically chewed up — the collagen fibres are disorganised, weaker, and the cells inside the tendon are dying off, the way a frayed rope gives way slowly under everyday loads it used to handle. The mechanical pinching does happen, but it's not the variable that drives the pain. The clearest evidence: when surgeons remove the bone in proper trials where the patient doesn't know whether the surgery actually happened, the pain doesn't go away faster than with the fake operation Beard 2018 Paavola 2018. Specialists who treat shoulders for a living are dropping the "impingement" label and using rotator cuff related shoulder pain or subacromial pain syndrome instead, precisely to stop telling patients a story that turns out to be wrong Diercks 2014 Lewis 2018.

What the surgery trials actually showed

The surgery in question — arthroscopic subacromial decompression, where a surgeon shaves a few millimetres of bone off the underside of the acromion through a keyhole — was for decades one of the most-performed orthopaedic operations in the world. It is no longer a defensible first-line treatment for routine shoulder pain, and the evidence that closed that question is unusually clean.

The BMJ's expert panel responded by issuing a strong recommendation against offering the surgery for ordinary subacromial pain Vandvik 2019. England's NHS down-coded the procedure; Australia restricted reimbursement. A surgery you might still get offered in 2026 is one a major guideline body has told its specialists not to do.

This is for the common picture — the deep ache, the painful arc, the bad nights — without a full-thickness tendon tear from an injury. For traumatic full-thickness tears in younger active people, the calculus changes. A Norwegian trial of 103 patients followed for a full decade showed that early tendon repair beat physiotherapy by a meaningful margin on shoulder function and pain at ten years; fourteen of the physiotherapy group eventually crossed over to surgery and ended up doing worse than the early-repair group Moosmayer 2019. For older patients with small, gradually appearing tears, a Finnish trial found surgery and physiotherapy equivalent at five years Kukkonen 2021. Which means the surgical conversation is real, but narrower than the conversation people are actually having.

The MRI trap

If you bring a sore shoulder to a doctor and they send you for an MRI, the scan will probably show something. A "partial-thickness tear", a "fraying", a "bursitis", a "tendinosis". This is almost certainly not the diagnosis your shoulder needed.

In a foundational 1995 study, researchers ran MRI scans on the shoulders of pain-free volunteers and found rotator cuff tears in 34% overall — and 54% of subjects over 60 had a tear with no symptoms at all Sher 1995. Ultrasound studies hit similar numbers; over age 80 it crosses half the population Tempelhof 1999. A Japanese village mass-screening of 683 residents found rotator cuff tears in roughly one in five — and two-thirds of those tears were in people with no pain Yamamoto 2010.

What this means in your clinic visit: if you have a painful shoulder and they find a tear on the scan, the tear may well predate the pain by decades. Treating the tear is not the same as treating the pain. A growing body of guideline and review literature now recommends against routine imaging in the first weeks of shoulder pain, exactly because of the false-signal rate Diercks 2014.

Two other things people get wrong. The first is "rest it" — which fails. Tendons need graded load to remodel; immobilising the shoulder for weeks worsens stiffness and slows recovery. The second is the steroid injection as a quick fix. Subacromial corticosteroid does reduce pain in the first weeks, but the effect is small and gone by three months in meta-analysis; repeat injections accelerate tendon degeneration and can weaken the tissue you're trying to rehabilitate Mohamadi 2017. As a one-off bridge to let you start a loading program, defensible. As an ongoing treatment, not.

What ignoring it actually costs

The first thing people lose is sleep. Lateral shoulder pain has a specific signature at night: the moment you roll onto that side, the bursa gets compressed and the pain wakes you up. You try the other side, your arm flops awkwardly across your body and you wake up anyway. Around 82% of people with shoulder pain lasting more than three months report disturbed sleep — and that's not the version where you fall back asleep easily. It's hours-of-staring, three-times-a-night, by-month-two-you've-forgotten-what-rested-feels-like sleep loss.

What grows out of that, slowly: the meeting where your partner asks why you're so short with the kids. The friend who comments that you look tired in every recent photo. Your morning coffee creeping from one cup to two to three. The week where you skip the gym because you don't have the energy, and then the next week, and then you're someone who used to lift. The chronic-shoulder-pain cohorts measure this as roughly one in five developing clinical depression and one in five clinical anxiety during the months it drags on.

The function loss is the second layer. The bra clasp moves. The seatbelt becomes a daily microflinch. Putting on a coat is a two-step manoeuvre. If your job involves any overhead reach — electrical work, painting, hairdressing, warehouse stocking — productivity drops noticeably and stays dropped. Tennis, swimming, climbing, anything throwing-shaped goes on hold. The reach behind your back to wash or scratch becomes the small daily failure that reminds you something is wrong.

The further-out version is the structural one. A rotator cuff tendinopathy left to its own devices doesn't always progress, but in a meaningful fraction it does. Roughly half of previously pain-free tears developed symptoms over three years in one careful study, and the tears that grew bigger were the ones that started hurting Mall 2010. A small tear that becomes a massive one, that retracts and fills with fat instead of muscle, eventually arrives at a place where surgery to put the tendon back is no longer possible — only joint replacement. The years between are recoverable; the late stage is not.

What actually works

A structured, progressive program targeting the rotator cuff and the muscles that move the shoulder blade, run for about twelve weeks. The tendon-loading core of it is slow, heavy eccentric training — the controlled lowering that a degenerating cuff tendon remodels under. The reference study — Holmgren's Swedish trial — gave one group of long-standing-impingement patients exactly this program; the other group did vague generic shoulder exercises. At three months, the specific-exercise group was substantially better off on every measure, and only one in five had ended up choosing surgery — versus nearly two in three of the generic-exercise group Holmgren 2012. The Dutch Orthopaedic Association's guideline now treats a course like this as first-line management Diercks 2014.

A practical pattern that maps to the trial protocols:

See a physiotherapist who treats shoulders. The first appointment establishes the exercises and the dose; the value of supervision is the load progression, not the hands-on.
Two exercise families, both required. Rotator cuff strengthening — external rotation with an elastic band, scaption (raising the arm at a 30-degree forward angle), eccentric loading. Scapular control — the muscles between and below the shoulder blade (lower trapezius, serratus anterior), which set the platform the cuff works against.
Daily home exercises, supervision every 1–2 weeks for the first three months. Not three sessions and a handout. The supervised contact is what keeps the load progressing.
Train in painful territory, not painful damage. Up to roughly a 4-out-of-10 pain during exercise is fine; pain that lingers more than 24 hours after the session means back off the load, not stop the program.
Stop carrying your arm. Use it for everything you normally would, minus the obvious overhead aggravators for the first few weeks.

The whole intervention is cheap relative to the alternatives: an eight-to-twelve-session physiotherapy course and an elastic band — no MRI scan, no operating theatre on the front end. In most public systems the course is covered or near-free; in private US care it runs a few hundred to about a thousand dollars in total. The expensive paths the evidence tells you to skip are the ones that look like care.

For nights — which is the symptom most people actually want fixed first — sleep on your back with the affected arm supported by a pillow under the elbow, or on the unaffected side with the painful arm hugging a pillow to your chest. Avoid the affected side. Ibuprofen or paracetamol an hour before bed during the first weeks is reasonable. A heat pack on the shoulder before sleep can help. None of this fixes the underlying problem, but you'll start sleeping again within the first week or two of the exercise program working.

Where this goes wrong

The most common reason the exercise program fails is that it never actually happens. One physio visit, an A4 sheet of stretches, two weeks of half-hearted bands at home, and back to the doctor asking about surgery. The Holmgren trial showed the gap between specific progressive loading and generic exercises was huge — but both required showing up Holmgren 2012.

The second failure is the load that never gets harder. An elastic band at week one is the right starting point; an elastic band at week eight is the program failing to progress. The cuff and scapular muscles need to be challenged enough that they're remodelling. If the supervised sessions stopped after the first two, this is the variable that quietly drops.

The third is the rest trap. People assume tendons heal by being left alone, the way a cut on your hand does. Tendons heal by being loaded — at the right dose, with the right progression. Three weeks of slinging the arm and avoiding everything that hurts produces a stiffer, weaker, more sensitive shoulder, not a healed one.

The fourth is steroid injections used as a substitute for the program rather than a bridge into it. Two or three injections over a year is enough to do real harm to the tendon you're trying to keep Mohamadi 2017.

What changes when you do it right

The first thing that comes back is the sleep. Usually within the first two or three weeks of doing the program properly — before the pain on movement has fully cleared — the night-pain element softens. You roll onto that side and you might wake up, but you fall back asleep. By week four or five you stop dreading bedtime. Your partner stops noticing you flinching when you reach for the alarm.

By the second month, the everyday pain dulls. The seatbelt grab stops registering. The dressing-and-undressing sequence collapses back into one fluid motion instead of a sequence of small calculated movements. People at work stop asking what's wrong with your shoulder because they stop seeing you favour it. The thing that started as your condition recedes back into being your shoulder.

By the third month — the end of a real Holmgren-style course — most people are functionally back. The Constant-Murley function scores in the trial improved by roughly 25 points on a 100-point scale, comfortably past the threshold patients report as a meaningful change Holmgren 2012. Overhead reach returns. The bra clasp moves to behind the back where it was. If you swim or throw or lift, you can start adding that load back deliberately.

The further-out payoff is structural and quiet. The tendon adapts to the load it's been given; the cuff is stronger than it was before the episode. The recurrence rate is not zero, but the program you learned how to do is the same program you reach for at the first hint of a flare in five years' time. For the older-patient picture with a small atraumatic tear, that's most likely the entire story — Kuhn's cohort of full-thickness atraumatic tears managed non-operatively saw three out of four patients avoid surgery at two years Kuhn 2013. For appropriately selected younger patients with traumatic full-thickness tears who do go to surgery, the ten-year data on primary repair is good Moosmayer 2019: durable function, durable pain relief, the kind of result that justifies the recovery cost.

Related to look into

If your shoulder pain doesn't match the rotator cuff pattern — pain on top of the shoulder rather than the side, a single point you can press on, pain from carrying a bag rather than reaching overhead — the acromioclavicular joint is the usual alternative. If your shoulder is increasingly stiff rather than weak, with a slow loss of range in all directions, frozen shoulder (adhesive capsulitis) is a different condition with a different course. Sharp shooting pain into the arm with neck movement points at the cervical spine, not the shoulder. Calcific tendinopathy — when a calcium deposit forms inside the cuff — produces a dramatic acute flare that doesn't fit the slow-burn pattern described here. And for sustained overhead work or sport, the conversation about load management and progressive shoulder conditioning is where this entry naturally leads.

Related in the handbook

Helps

— The twelve-week exercise program that fixes most rotator cuff pain leans heavily on slow, loaded tendon work.
— A daily collagen scoop with vitamin C may help the tendon respond to loading. Small, slow, cheap — the work still matters.

Alternatives

— If your shoulder is going stiff in all directions rather than weak and painful on reaching, it's likely frozen shoulder, a different problem.

— A torn cuff on MRI is like a disc bulge on a back scan — common in pain-free people and not, by itself, the diagnosis.
— Hip impingement and gluteal tendinopathy share the shoulder's trap — MRI findings and tempting, oversold surgery.
— Shoulder impingement often traces back to a stiff mid-back; restoring thoracic extension gives the cuff room to work.
— A cranky cuff and a cranky elbow share the same cause and the same loading fix.

Substance + claimed effects

Rotator cuff tendinopathy and subacromial impingement are the two umbrella labels for the most common cause of adult shoulder pain. The rotator cuff is four small muscles — supraspinatus, infraspinatus, teres minor, subscapularis — whose tendons wrap the head of the humerus and produce both stability and elevation. "Tendinopathy" describes degenerative change in those tendons (most often supraspinatus): collagen disorganisation, loss of type-I collagen, neovascularisation, tenocyte apoptosis, with little classical inflammation in chronic disease. "Subacromial impingement" was the dominant 20th-century model — the supraspinatus pinched between the humeral head below and the acromion above — and is now best understood as a clinical syndrome, not a pure anatomic diagnosis; contemporary terminology (Dutch Orthopaedic Association's subacromial pain syndrome, Lewis's rotator cuff related shoulder pain) deliberately drops the mechanistic claim Diercks 2014 Lewis 2018. The condition spans a continuum from reactive tendinopathy through partial-thickness tears to full-thickness rotator cuff tears, with the same pain pattern (lateral deltoid ache, arc of pain on elevation, night pain especially on the affected side) recurring across the spectrum Mitchell 2005. Meaningful consequences scored in this entry: short-term health (this is the dominant axis — pain and function), sleep (night pain is a near-universal complaint and one of the strongest drivers of consultation), mood (chronic shoulder pain carries roughly 20% depression and 80% sleep-disturbance rates), energy and focus (downstream of sleep loss and persistent pain attention costs), effort (the evidence-based first-line treatment is sustained progressive loading), and the framing call on imaging (every reader who pays out-of-pocket for an MRI is the meaningful financial consequence). No effect on longevity, beauty, or appearance.

Evidence by addressing question

Mechanism

The classical Neer impingement model — supraspinatus mechanically pinched as the arm elevates against the undersurface of a hooked or spurred acromion — has driven half a century of surgical acromioplasty. The model is now best treated as one contributor, not a complete account. Histology of symptomatic tendons shows degenerative change rather than mechanical fraying: rounded tenocytes, extracellular-matrix disorganisation, reduced type-I collagen, increased type-III collagen (smaller, mechanically weaker fibrils), neovascularisation, and minimal inflammatory infiltrate in chronic cases. The extrinsic (mechanical) and intrinsic (tendon-degeneration) hypotheses are not exclusive; both feed the same final common pathway of tendon failure. The shift in clinical reasoning is the recognition that "the bone is pinching the tendon" is rarely the operative variable: the FIMPACT and CSAW placebo-surgery trials showed that removing the bone does not improve symptoms over sham Beard 2018 Paavola 2018. Pain in tendinopathy is poorly correlated with structural pathology — peripheral and central sensitisation, scapular kinematics, posture, and load tolerance all matter. The Dutch Orthopaedic guideline accordingly recommends a clinical-syndrome approach over a pathoanatomic one Diercks 2014.

Evidence — exercise vs surgery

For subacromial pain syndrome (no full-thickness tear), the question "does the surgery help?" has now been answered to high certainty by three converging lines of evidence. The CSAW trial randomised 313 patients across 32 UK hospitals to arthroscopic subacromial decompression, sham arthroscopy, or no treatment, with the Oxford Shoulder Score primary endpoint at 6 months; decompression was not superior to sham, and both surgical groups exceeded no-treatment only by a small margin attributable to the post-operative care pathway Beard 2018. The Finnish FIMPACT trial randomised 210 patients to decompression, diagnostic arthroscopy (sham), or supervised exercise, with primary endpoints at 24 months on shoulder pain at rest and on arm activity; no clinically meaningful difference between decompression and sham at 24 months, and the result held to 5 years Paavola 2018. The Cochrane review (Karjalainen 2019, 8 trials, 1,062 participants) graded the evidence high-certainty for no benefit of decompression over placebo on pain, function, or quality of life, with a small risk of serious harm Karjalainen 2019. The BMJ Rapid Recommendations panel issued a strong recommendation against subacromial decompression for subacromial pain syndrome on the back of these reviews Vandvik 2019 Lahdeoja 2019. Earlier RCTs pointed in the same direction: Ketola randomised 140 patients to exercise alone vs exercise plus acromioplasty and found no group difference at 2 or 5 years Ketola 2013. On the exercise side, Holmgren randomised 102 patients with persistent impingement to a 12-week progressive rotator-cuff-and-scapular strengthening program vs an unspecific control after a steroid injection; the specific-exercise group showed substantially better Constant-Murley scores at 3 months and far fewer crossing over to surgery Holmgren 2012.

For atraumatic full-thickness rotator cuff tears, the MOON multicenter cohort followed 452 patients through a structured non-operative program; 75% avoided surgery at 2 years with no clinically meaningful difference in function vs surgical cohorts Kuhn 2013. For traumatic or progressive tears in younger active patients, Moosmayer's Norwegian RCT (103 patients, small-to-medium tears) found tendon repair superior to physiotherapy at 10 years on Constant score (+9.6), ASES (+15.7), pain VAS, and range of motion; 14 of the physiotherapy group crossed over to surgery and scored lower than primary repair at follow-up Moosmayer 2019. The Finnish trial by Kukkonen (180 patients >55 with small atraumatic supraspinatus tears) found no clinically meaningful difference between repair and physiotherapy at 1, 2, or 5 years Kukkonen 2021. The two trials disagree on the right call because they sampled different populations — the gap between "younger active patient with a recent or enlarging tear" and "older patient with a small chronic tear" is exactly where the treatment decision belongs.

Protocol

The Dutch Orthopaedic Association guideline recommends 12 weeks of supervised exercise as first-line management, targeting the rotator cuff (external rotation, scaption) and scapular stabilisers (lower-trapezius and serratus anterior), with progressive load and eccentric components Diercks 2014. Holmgren's protocol — daily home exercises + bi-weekly supervised sessions over 12 weeks — is the canonical reference Holmgren 2012. Subacromial corticosteroid injection produces short-term pain relief (≤8 weeks) that does not persist; the Mohamadi meta-analysis (11 RCTs, 726 patients) graded effect sizes as small and transient, and repeated injections accelerate tendon degeneration Mohamadi 2017. Acetaminophen and NSAIDs are reasonable for pain control during the first weeks of a loading program. Activity modification (avoiding repeated overhead loading at the painful arc) supports the program without immobilisation; complete rest worsens stiffness and delays recovery. Sleep position matters: side-lying on the affected shoulder reliably reproduces pain; back-sleeping with an arm pillow or side-lying on the unaffected side with the painful arm supported on a pillow is the standard recommendation, supported by the night-pain symptom mechanism rather than RCT data.

Contraindications

"Don't load through pain that lingers >24 h after a session" is the practical rule. True red flags requiring escalation rather than exercise: acute trauma with sudden loss of active elevation (full-thickness tear should be imaged within weeks in patients who would consider surgery; tear enlargement is associated with symptom development Mall 2010), constitutional symptoms, fever, history of cancer (referred pain), or neurological deficit suggesting cervical radiculopathy or brachial plexus involvement. Diabetes is both a risk factor for cuff disease and a complicating factor for healing; injection regimens should account for glycaemic disruption. Younger active patients with a traumatic full-thickness tear may benefit from earlier surgical consultation Moosmayer 2019.

Misconceptions

Three common errors. First, imaging-driven diagnosis: rotator cuff tears on MRI or ultrasound are common in pain-free shoulders — 34% across all ages in Sher's MRI series (54% over 60) Sher 1995, 23% on ultrasound in Tempelhof (51% over 80) Tempelhof 1999, 20.7% in Yamamoto's village mass-screening (with two-thirds asymptomatic over 60) Yamamoto 2010. Imaging in a painful shoulder will frequently find a "tear" that is not the cause of pain. Second, the impingement-causes-pain story: subacromial decompression doesn't beat sham, which means the mechanical-pinching account is at minimum incomplete Beard 2018 Paavola 2018 Karjalainen 2019. Third, the rest-and-recovery story: prolonged immobilisation and avoidance worsen outcomes; the tendon needs graded load to remodel.

Failure modes

Exercise therapy fails in roughly 20–30% of patients in the high-quality trials — Holmgren saw 20% of the specific-exercise group cross to surgery vs 63% in the unspecific-exercise group Holmgren 2012. Failure modes: inadequate dose (1–2 supervised sessions then dropped at home), wrong target (general shoulder mobility instead of progressive cuff + scapular loading), under-progression (the load never moves beyond elastic-band level), ignoring scapular kinematics, and high baseline psychological distress (which independently predicts worse outcomes across musculoskeletal pain). Surgical failure modes after rotator cuff repair: retear (25–60% on imaging at 1 year for larger tears; clinical outcomes often acceptable despite retear), stiffness, and infection. Surgical failure after decompression for impingement: the procedure didn't fix what was causing the pain.

Practicalities

Annual prevalence of rotator-cuff-attributable shoulder pain is roughly 1–4% in the general adult population, rising sharply with age and overhead occupational exposure; in primary-care shoulder-pain populations rotator cuff disease accounts for 70–85% of presentations Littlewood 2013 Mitchell 2005. Risk factors: age over 50, manual or overhead occupations, sustained overhead sports (swimming, throwing), diabetes, obesity, smoking. Conservative care costs in most public systems are low — 8–12 supervised physiotherapy sessions are usual; private rates run roughly $50–$150 per session. The crossing point in cost is surgery: arthroscopic rotator cuff repair runs $5,000–$15,000+ in private US care, with weeks-to-months of recovery. Time to functional recovery: 6–12 weeks for tendinopathy responding to exercise, 3–6 months for surgical recovery (sling and graded protocol), 9–12+ months for full athletic return after large repairs.

Stakes

Untreated rotator cuff disease typically does not resolve spontaneously and tends to progress in a meaningful subset: Mall's prospective ultrasound cohort showed 51% of previously asymptomatic tears became symptomatic over a mean 2.8 years, and tear enlargement (mean +10.6 mm in the newly symptomatic vs +3.3 mm in the still-asymptomatic) predicted pain development Mall 2010. The lived stakes are sleep (82% of patients with chronic shoulder pain >3 months report disturbed sleep), mood (22% rate of depression, 19% anxiety in the same cohort), and function (loss of overhead reach, dressing, driving comfort, work capacity for any overhead task). Progression to massive (>5 cm), retracted, fatty-infiltrated cuff tears risks cuff-tear arthropathy and reverse-shoulder arthroplasty as the only late option.

Payoff

Treatment response is real and reproducible. Holmgren's specific-exercise group at 12 weeks showed Constant-Murley score improvements of ~25 points (clinically meaningful threshold ~10–15) and the surgical-crossover rate dropped to 20% from 63% in controls Holmgren 2012. The MOON cohort showed 75% of full-thickness atraumatic tears managed non-operatively avoided surgery at 2 years Kuhn 2013. Sleep returns within weeks of pain control; the night-pain element is usually the first to remit. For appropriately selected surgical candidates (younger, traumatic, full-thickness), Moosmayer's 10-year data shows durable advantage of primary repair over physiotherapy on function and pain Moosmayer 2019.

Out-of-scope (adjacent)

Frozen shoulder (adhesive capsulitis) and shoulder osteoarthritis present with overlapping pain but distinct clinical patterns and management. Acromioclavicular joint pathology, calcific tendinopathy, biceps tendinopathy, and cervical-radiculopathy referred pain are the differential diagnoses every primary-care assessment should at least rule out. Posture and load-management for occupational and sport overhead exposure interact with this condition over years.

The credibility range

Optimist case (for the impingement-decompression model)

A defender of the classical model would argue: the placebo trials don't disprove decompression in all patients, only in the heterogeneous subacromial-pain populations recruited. Surgeons identify a subgroup (younger, type-3 hooked acromion, painful arc unresponsive to rehab) for whom acromioplasty does work — the trial averages mask this. Rotator cuff repair for traumatic full-thickness tears has the durable Moosmayer 10-year data behind it, and waiting risks irreversible muscle atrophy and fatty infiltration. The community of orthopaedic shoulder specialists has not collectively abandoned arthroscopic decompression because they observe responders.

Skeptic case

The defender's appeal to a responder subgroup is unfalsifiable until prespecified subgroups in adequately powered trials show benefit, and they have not Karjalainen 2019. The strongest evidence available — two large multicentre placebo-controlled RCTs and a high-certainty Cochrane review — converges on no benefit over sham. The "impingement" diagnosis itself misframes the mechanism: pain in tendinopathy correlates poorly with the mechanical-pinching account, and removing the bone does not remove the pain. The cost is non-trivial: surgery carries small but real risks (frozen shoulder, infection, anaesthetic complications) for no expected benefit. The clinical implication is that exercise is the first and often only treatment that works for the majority of subacromial pain.

Author's call

For subacromial pain syndrome without full-thickness tear, the BMJ Rapid Recommendation panel's strong-against position for subacromial decompression surgery is the correct read of the literature Vandvik 2019 — high-certainty, consistent across multiple high-quality trials. Exercise is first-line, full stop. For atraumatic full-thickness tears in older patients with low functional demand, conservative care is reasonable and often successful Kuhn 2013 Kukkonen 2021. For traumatic full-thickness tears in younger active patients, primary repair carries durable functional advantage Moosmayer 2019 and the surgical conversation should happen early — within weeks, not after a year of failed rehab — before fatty infiltration becomes irreversible. The entry's tone is conservative-skeptic on imaging (a tear on MRI is not a diagnosis), pro-exercise as the default action, pro-surgery in a specifically defined subgroup. Controversy score reflects active ongoing reframing in the field (impingement → rotator cuff related shoulder pain) and substantial gap between trial evidence and surgical practice rates.

Stakeholder + incentive map

Orthopaedic surgeons — financial and professional incentive to perform decompression / repair; subacromial decompression was historically among the highest-volume orthopaedic procedures. The CSAW and FIMPACT placebo trials directly threaten that volume.
Physiotherapists / physical therapists — incentive to position exercise as first-line; the evidence is on their side for subacromial pain syndrome. Risk: under-progressing loading programs.
Patients — pulled toward imaging and surgery by the narrative that "something is torn and needs to be fixed"; pulled toward conservative care by cost, recovery time, and the placebo-trial evidence.
Insurance / public payers — incentive to restrict decompression coverage (UK NHS England, Australian commissioning bodies have explicitly down-coded the procedure post-Beard).
Imaging providers — incentive to image more shoulders; the asymptomatic-tear prevalence data is a counterweight.
Steroid-injection clinics — short-term pain relief is real but transient and tendon-degenerative on repeat; commercial structure encourages repeat use.

Population variability

Age — prevalence rises steeply from age 50. Sher: 4% under 40, 54% over 60 with cuff tears on MRI in asymptomatic shoulders Sher 1995. Tempelhof: 23% baseline, >50% over 70 on ultrasound Tempelhof 1999.
Occupation — overhead workers (painters, electricians, mechanics, carpenters) carry several-fold higher prevalence; the Littlewood review documents working-population estimates 2–40% vs general 1.2% Littlewood 2013.
Sport — overhead-throwing athletes, swimmers, climbers, weightlifters. Pattern of presentation differs (often internal impingement at end-range external rotation in throwers).
Diabetes — clearly associated, both with risk and with impaired tendon healing.
Obesity — BMI ≥35 associated with 3-fold higher likelihood of repair surgery.
Sex — slight female predominance in primary-care presentations but not consistent across all studies.
Traumatic vs atraumatic onset — traumatic full-thickness tears in patients <65 have a distinct treatment calculus (early surgical consultation); atraumatic chronic tendinopathy is the typical exercise-responsive presentation.

Knowledge gaps

The unresolved questions: which exercise-program features matter most (specific cuff + scapular loading vs general shoulder exercise — Holmgren says specific is better, but the meta-analyses are mixed) Holmgren 2012; whether a responder phenotype exists for surgery in subacromial pain syndrome (no adequately powered subgroup analysis has identified one); the right timing of surgical consultation in atraumatic full-thickness tears in active middle-aged patients (Moosmayer says earlier in the Norwegian sample, Kukkonen says later doesn't matter in the older Finnish sample); the role of pain neuroscience education and centrally-acting interventions in non-responders; the long-term tendon-degenerative effect of repeated corticosteroid injection Mohamadi 2017. What would change the call: a placebo-controlled RCT of rotator cuff repair in young active patients with traumatic tears (logistically and ethically difficult); a validated phenotyping system that prospectively identifies the surgical responders the optimist case posits.

Scope decisions. The brief named four consequences — pain, function, sleep, recovery — plus a comparison of exercise-based and surgical management and interpretation of imaging findings. All six are covered in the body. Pain and function are the dominant axis (scored under health_short_term); sleep gets its own dimension and its own paragraphs in stakes, protocol, and payoff; recovery is the through-line of the protocol, failure-modes, and payoff sections. The exercise-vs-surgery comparison is the spine of evidence; imaging interpretation is the lead of misconceptions.

Title. Used "Subacromial Pain" rather than "Impingement" in the title because the contemporary literature is actively retiring "impingement" — Diercks 2014, Lewis 2018, and the BMJ Rapid Recommendation Vandvik 2019 all use the newer terminology. Kept "tendinopathy" in front for searchability — that's still what most patients and primary-care doctors will type.

Excluded from the body.

Frozen shoulder, calcific tendinopathy, AC joint pathology. Each is a distinct condition with a distinct course; lumping them in would dilute the actionable specifics. Signposted in out-of-scope as separate-entry candidates.
Detailed surgical technique comparisons (open vs arthroscopic, single vs double-row repair). Reader doesn't choose between these — the surgeon does. Out of altitude for a reader-facing entry.
Platelet-rich plasma, stem-cell, prolotherapy injections. Evidence is too thin to recommend or to confidently dismiss; including them would lead readers down expensive paths the literature doesn't support. Flagged for a future entry once trial data matures.
Massive cuff tear / reverse shoulder arthroplasty. Late-stage surgical territory; out of scope for a literacy entry on the common adult presentation. Mentioned briefly in stakes as the bad-progression endpoint.

Rating difficulties.

health_short_term scored 4 not 5. The pain and function effect is dominant by topic but treatment response is a 12-week project, not a one-week transformation; 5 implies a new baseline that comes faster than the rehab realistically delivers.
energy and focus scored 2 each. Both are downstream of sleep and pain rather than primary effects; scoring them 0 would under-represent how chronic shoulder pain actually shows up in daily life, but scoring them higher would overstate the direct effect.
cost_burden scored 1. The first-line action (physiotherapy course) is genuinely cheap in most contexts; the expensive paths (MRI, surgery) are the ones the entry argues against. Could be argued at 2 in private-pay US contexts, but the entry's recommendation keeps cost low.
effort_burden scored 3. Three months of daily home exercises plus bi-weekly supervised sessions is substantial sustained effort; the high crossover-to-surgery rate in Holmgren's control group is the practical evidence that adherence is the binding constraint.
controversy scored 3. The pathoanatomic "impingement" model is being walked back across the field, and surgical practice rates substantially exceed what the BMJ Rapid Recommendation justifies. Active reasonable disagreement.

Hard call: surgery framing. The literature on subacromial decompression is unusually clean — two large placebo-surgery RCTs and a high-certainty Cochrane review pointing the same direction. Wrote the article confidently on that. The traumatic-full-thickness-tear subgroup gets a single warning callout in protocol rather than a separate addressing section, because the entry's audience is the general adult-with-shoulder-pain reader and the surgical-candidate subset is small. Reviewer should check whether this balance reads as appropriately conservative-against-surgery without burying the legitimate surgical conversation.

Future links. "Frozen Shoulder (Adhesive Capsulitis)", "Acromioclavicular Joint Pain", "Calcific Tendinopathy", "Cervical Radiculopathy", and a separate "Progressive Loading for Tendinopathy" entry (the protocol logic generalises to Achilles, patellar, lateral elbow). Wire cross-links in once those entries land.

Separate-entry candidates. Progressive loading as a general tendinopathy principle warrants its own entry; the exercise-protocol detail here only scratches the surface and is reused for Achilles / patellar / lateral epicondylopathy.