Story
Shoulder Pain
Last updated
In one pass The top of the upper-arm bone sits like a golf ball on a shallow dish: it can point in any direction, but on its own it is not stable at all.
Educational content, not medical advice — consult a clinician.
Story path
Chapter 1
Why the shoulder is mobile but unstable
The shoulder trades that shallow socket for the largest range of motion in the body, and its stability depends almost entirely on the four small muscles of the rotator cuff holding the ball in place moment by moment. The knee and hip sit in deep sockets and are stable by design; the shoulder takes the opposite route. Most shoulder pain comes back to this trade of stability for movement: either the joint cannot stay in place, or the tendons that keep it steady wear out.
One kind of shoulder pain is not about the shoulder: pain in the left shoulder or arm together with chest tightness, breathlessness, or a cold sweat may be a heart problem. Call emergency services right away.
Mechanism · Why a shallow, muscle-held joint is fragile
This trade has two sides, and most of the kinds of shoulder pain described later follow from it.As Knee Pain explains, joint cartilage has no nerves and no blood vessels, so the cartilage itself cannot hurt. The same is true at the shoulder, but shoulder pain comes more from soft tissue: tendons, the bursa, and the joint capsule, all full of nerves and all painful when squeezed, inflamed, or overloaded.
First, what shallow brings:
The socket (the glenoid, that shallow dish) covers only a small part of the ball, so bone provides almost no restraint. Stability rests almost entirely on soft tissue: at rest on the rim of cartilage (the labrum) and the ligaments, and in motion on the rotator-cuff muscles tightening moment by momentSo the shoulder has two opposite kinds of trouble. One is not staying in place (dislocation, partial dislocation, a torn labrum), common in young people and after injuries. The other is the stabilizing tendons being overused until they degenerate or get pinched (rotator-cuff-related pain), common in middle-aged and older people and in anyone who often works with the arms overhead
Then, why the rotator cuff is a hot spot:
The supraspinatus tendon runs over the top of the ball, through a narrow gap between the ball and the bony roof above it, the acromion and the coracoacromial ligament (the subacromial space)When you raise your arm overhead again and again (swimming, painting walls, lifting weights, carrying a child), this tendon slides back and forth through the gap under tension. Part of it also has a naturally weaker blood supply, so over time it tends to wear and degenerateA common explanation is that a degenerated tendon weakens and cannot hold the ball down, so the ball rides up, the gap narrows further, and a self-reinforcing loop forms
This is why shoulder pain so often has a recognizable pattern: a catch at a certain angle, a struggle to lift the arm overhead, pain at night when lying on that side. It is telling you that the tendon and bursa in that narrow gap are the ones complaining, not that the joint has worn through.
Mechanism · The other parts, and scapular rhythm
Picturing the shoulder as a ball, a shallow dish, and a hand keeps things simple. The real shoulder has a few more parts, and each one maps onto a kind of shoulder pain.Glenohumeral joint: the true shoulder joint, the pairing of that ball and shallow dishLabrum: a ring of fibrous cartilage set around the edge of the dish to deepen it a little, which also gives the ligaments and the long head of the biceps tendon an anchorThe four small rotator-cuff muscles: supraspinatus, infraspinatus, teres minor, and subscapularis. Their tendons wrap the ball from different directions and together form that hand holding the ball inBursa: the subacromial bursa is a fluid-filled cushion that reduces friction between the rotator cuff and the bone above itAC joint: the small joint at the very top of the shoulder where the outer end of the collarbone meets the acromion of the shoulder blade
One more thing is easy to overlook: scapular rhythm. When you raise your arm overhead, it is not only the upper-arm bone that moves. The shoulder blade slides and rotates across the back of the rib cage at the same time, making room for the ball. Arm and shoulder blade move together at roughly a 2:1 ratio. A common view is that when this rhythm goes wrong (a slumped chest, rounded shoulders, weak muscles around the shoulder blade), the space above gets squeezed and the rotator cuff is more easily pinched and worn. That said, slightly irregular shoulder-blade movement is also common in people without pain, and whether it causes pain is still debated.
Chapter 2
Five common causes of shoulder pain
The five most common kinds seen in clinic: rotator-cuff-related pain (the largest group, once called impingement); frozen shoulder (the joint capsule tightens so much that no one else can move the arm either); instability or dislocation (the price of a shallow socket); problems at the AC joint (pain at the very top of the shoulder); and referred pain (felt in the shoulder, but the source may be elsewhere). Full-thickness tendon tears seen on scans become more common with age, yet many people with them feel no pain at all.
Mechanism · Why the label impingement was dropped
Rotator-cuff-related pain is the biggest slice of shoulder pain, and the slice most misled by one old word.The old, overly mechanical model was called subacromial impingement. It assumed that the narrow gap under the acromion knocks the tendon against bone when you lift your arm, and that repeated knocking wears it down. By that logic, shaving off a little bone to open up the space should fix it. That was exactly the rationale for the subacromial decompression surgery that later became popular.
Current understanding leans toward a combined process of load and degeneration, which is why the field now uses a more neutral name: rotator-cuff-related shoulder pain.
Under repeated overhead load, the tendon undergoes degeneration (tendinosis) rather than simple inflammation. It is the same idea as jumper's knee (patellar tendinopathy) in Knee Pain: under the microscope you see disorganized collagen and repair falling behind the load, not a crowd of inflammatory cellsAcromion shape and bone spurs are real, but their role in causing pain has been overstated; many people have the same shapes on scans and no painAn off-balance scapular rhythm and weak rotator-cuff and shoulder-blade muscles may let the ball ride up and shrink the space available during movement; this is the part training can change
Why does the name change matter so much? Because it rewrites the treatment logic directly. If the problem is bone hitting tendon, you should operate to shave bone. If the problem is tendon degeneration plus poorly controlled muscles, then exercise rehabilitation becomes the main treatment, and bone-shaving surgery loses its rationale. Several randomized trials with placebo-surgery controls later confirmed this shift, as the chapter on common misconceptions explains.
One more fact tied to age: degeneration of the rotator-cuff tendons with age is very common, and partial and even full-thickness tears on scans become steadily more common with age in people who have no symptoms at all. That means a tear showed up on the scan and this tear is the cause of your pain are two different things.
Clinical · Details of each kind of shoulder pain
Each of the five kinds of shoulder pain has its own typical picture.Rotator-cuff-related pain: this group covers rotator-cuff tendinopathy, subacromial bursitis, and partial tendon tears. The classic picture is pain as the arm passes through the middle of its lifting arc (roughly 60-120 degrees), difficulty raising the arm overhead, and pain at night when lying on that side. The old name impingement (subacromial impingement) misled people for decades.
Frozen shoulder: the most typical limit is that the arm will not rotate outward even when someone else turns it for you (passive external rotation). The worst part is that it is very painful, but the good news is that it usually runs its own course: a freezing phase, a frozen phase, and a thawing phase, over several months to a year or two. It is more common in people with diabetes and in women aged 40-60.
Instability and dislocation: in young people, an injury (falling onto an outstretched hand, contact sports) pops the ball out of the socket, often tearing the labrum (a Bankart lesion). After one dislocation, the chance of another rises sharply, especially in the young.
AC joint problems: landing on the shoulder in a fall (which drives the outer end of the collarbone down) can sprain or even separate the AC joint. Wear-and-tear arthritis of the AC joint is also common in middle-aged and older adults, causing pain at the top of the shoulder when the arm reaches across the chest.
Referred pain: problems in the neck, heart, gallbladder, or diaphragm can all send pain to the shoulder. One key feature is that moving or pressing on the shoulder does not change this pain. When it needs urgent care is listed in the chapter on when you must see a doctor.
Beyond these five there are long-head biceps tendinopathy, calcific tendinitis, rheumatoid arthritis, and others. But in clinic the large majority of shoulder pain falls into the five groups above, and the first group, rotator-cuff-related pain, accounts for well over half on its own.
Chapter 3
Five shoulder myths that don't hold up
Five widely repeated claims that do not hold up: decompression surgery will fix it (real surgery did no better than placebo surgery); a tear on a scan must be repaired (many full-thickness tears do not hurt); a frozen shoulder should be forced open early (it usually runs its own course); the more rehab hurts, the better it works; and shoulder pain is always a shoulder problem (the neck, heart, gallbladder, and diaphragm can all send pain to the shoulder).
Myth · Decompression is no better than sham
Shave the bone for shoulder pain was once one of the most common shoulder operations in the world, until a set of overturned it.The CSAW trial (Beard 2018, Lancet): patients with rotator-cuff-related shoulder pain (no full-thickness tear) were randomly split into three groups: real arthroscopic subacromial decompression, a placebo operation (the camera goes in to look but nothing is removed), and no surgery with follow-up only. Result: the decompression and placebo groups showed no clinically meaningful difference in pain and function. Both surgery groups did slightly better than doing nothing, but the gap was too small to reach the threshold of clinical importance, and it very likely came from having an operation at all (the placebo effect, plus the rehab that always follows surgery)The Ketola trial: compared subacromial decompression plus exercise with exercise alone. At two-year follow-up, adding decompression brought no extra benefit; the exercise was the part doing the work
The implication is direct: for the great majority of rotator-cuff-related pain without a full-thickness tear or mechanical catching, the bone-shaving step itself adds almost no extra improvement. The improvement comes mainly from time, rehab, and the placebo effect, not from the sliver of bone the operation removes. This also shows why the name impingement had to go: if the problem is not bone hitting tendon, shaving bone naturally misses the point.
Draw the line clearly, though, and do not overcorrect. This conclusion is about rotator-cuff-related pain (no full-thickness tear, no clear mechanical problem). It does not rule out shoulder surgery that is genuinely indicated, such as repairing a large acute tear in a young person, stabilizing a shoulder that keeps dislocating, or releasing the capsule in a stubborn frozen shoulder. The right order is to do rehab properly first and reserve surgery for people with a clear structural problem whose conservative treatment has truly failed.
Myth · A tear on a scan must be repaired
The link between scans and symptoms deserves its own explanation, because at the shoulder it is even starker than at the knee.Age-related wear in the rotator-cuff tendons is the norm. Ultrasound and MRI studies keep finding the same thing: among ordinary people with no shoulder pain at all, the share with partial or even full-thickness rotator-cuff tears rises steadily with age, and by the sixties and seventies a sizable share of people have a tear but no pain. In other words, a tear seen on a scan is, for many people, just a normal mark of aging, not necessarily the cause of the pain.
Two direct consequences follow:
Scan first and decide later often backfires at the shoulder. A scan is very likely to find something (wear, a tear, a bone spur), and that finding then pulls the attention of both patient and doctor onto it, leading to overtreatment and even unnecessary surgery. So for shoulder pain without red flags and with mild symptoms, guidelines usually advise trying a period of conservative care first, without rushing to imagingI have a tear, so it must be repaired skips a step. What really needs judging is whether this tear is the source of the symptoms and whether it is the kind surgery can fix. Wear-related tears that did not come from an injury usually start with exercise rehab; a large, acute tear from an injury in a young person, with a clear loss of strength, is the group with a clear case for surgical repair
It is the same logic as wear-related meniscus tears in Knee Pain: scans find them everywhere, and many people have no symptoms; the question is not whether a tear exists, but whether it is causing your pain and whether surgery can fix it. Keeping this in mind helps you stay calm in front of a report full of jargon and resist being pushed into an unnecessary operation.
Chapter 4
Gradual loading and patience
The first-line treatment for rotator-cuff-related pain is exercise therapy with gradually increasing load, not an add-on. For wear-related tears that did not come from an injury, exercise and surgery gave roughly similar results at follow-up. For a frozen shoulder, the main path is to control the pain, keep the range of motion, and give it time; it usually runs its own course. With instability, it matters whether this is a first dislocation from an injury or a shoulder that keeps dislocating. A short course of anti-inflammatory pills only buys a window in which you can train; it does not repair the tendon.
Clinical · What to train, and two things alongside
Exercise therapy is the core treatment, and it has these parts.It follows the same logic as tendinopathy in Knee Pain: a worn tendon does not like rest, it likes to be loaded just right, and gradually increasing tension prompts it to reorganize its collagen and grow stronger and thicker.
Rotator-cuff strength: train outward rotation, inward rotation, and lifting the arm out to the side, within a pain-free or mildly painful but acceptable range (usually with a resistance band or light dumbbells)Shoulder-blade stability: train the serratus anterior and the middle and lower trapezius to restore scapular rhythm and make room above the ballControlled load, increased gradually: several randomized trials show that structured exercise rehab matches decompression surgery over the medium to long term. That is the ingredient that actually did the work when the CSAW and Ketola trials overturned decompression surgery
Two things to do alongside
Relative rest, not complete stillness: cut back for a while on overhead loads that make the pain worse (for example, no painting ceilings or heaving heavy things overhead for a while), but keep moving through the rest of the range. Keeping the arm completely still only makes the shoulder stiffer and the tendon weakerHelping the patient understand what is going on: knowing that this is a problem of tendon wear and load management, not a worn-through joint, and surgery is not a must reduces fear and makes people more willing to stick with rehab. That is part of the treatment effect, not empty talk
Clinical · Training, frozen shoulder, and surgery
Turning you should exercise into something you can actually do.How to train (rotator-cuff-related pain)
Rotator cuff: outward and inward rotation with a band (elbow tucked against your side), raising the arms out to the side up to a pain-free height, and Y, T, and W shapes. Start with no pain or mild acceptable pain, then add resistance step by stepShoulder blade (often neglected): serratus anterior (a wall push-up that pushes the shoulder blades a little farther forward at the top, or the same extra push at the top of a press) and the middle and lower trapezius (Y and T raises lying face down), to restore the upward rotation of the shoulder blade as the arm liftsOverhead work: until the pain settles, avoid high-arc loads that make it worse; once it settles, add overhead movements back graduallyFrequency and timeline: 2-3 times a week, kept up for weeks to months before steady improvement shows. A rotator-cuff tendon is a long project, just like the knee, so do not expect results in days
How to judge pain during rehab (the myth that more pain means better results)
Gradually increasing load allows pain within an acceptable range: a common rule is no more than 3-5 out of 10 during training, and no worse within 24 hoursA bit achy or sore, but not getting worse is fine to train through; more pain with each session and worse the next day means cut back. Pain is a signal, not a ban, and more is not better
How to handle a frozen shoulder
The main path is patience, keeping range of motion within what you can tolerate, and controlling pain. Do not force itA steroid injection early on can shorten the most painful phase; manipulation under anesthesia and capsular release are reserved for the few who stay stubbornly stuck for a long time
Who surgery is for (a last resort)
Rotator-cuff repair: a large, acute tear from an injury in a young person, with a clear loss of strength, is a clear indication. Wear-related tears that did not come from an injury usually start with rehabStabilization surgery: young people whose shoulder keeps dislocating and who have a labral (Bankart) lesionSubacromial decompression: since the CSAW and Ketola trials, no longer routinely recommended for plain rotator-cuff-related painRelease for frozen shoulder (manipulation under anesthesia, MUA, or capsular release): only for the rare person still stubbornly frozen long past the usual course
The logic of the whole treatment: what really changes the course is gradually increasing load and time; injections and surgery deal with symptoms, specific structural problems, and the end stage. It is the same theme as in Knee Pain: the shoulder is almost always a story of move, but move smart, not spare it and keep still.
Chapter 5
Why tendons need load, not rest
So anti-inflammatory pills plus complete rest miss the target. The pills can ease pain for a while, but keeping the arm completely still lets collagen production and strength keep falling. A tendon gets strong again through controlled, gradually increasing tension. Even for acute soft-tissue injuries, the POLICE approach has long since replaced complete rest with optimal loading.
Evidence · Three ways to load a tendon
The idea that gradually increasing tension repairs tendons comes mainly from the three trials below. They studied the Achilles and patellar tendons; applying them to the rotator cuff is reasoning from the same principle, and direct evidence in the rotator cuff is thinner.Eccentric training (the muscle works while being lengthened): the classic Achilles eccentric program from Alfredson 1998 established the idea that gradually increasing load repairs tendonsHeavy slow resistance (HSR): in a 12-week controlled trial in people with patellar tendinopathy, Kongsgaard 2009 found heavy slow resistance worked as well as eccentric training, and both kept their gains at the half-year follow-upIsometric contractions (the muscle works but the joint does not move): in a small number of athletes with patellar tendinopathy, Rio 2015 found that isometric holds eased pain on the spot, letting people start loading safely even while pain is high. It was a small trial, and later studies have not agreed
Myth · Why pills and rest often fail the shoulder
This one deserves its own page, because it harms the most people and is especially typical at the shoulder.The common approach (often useless): at the first twinge, put the arm in a sling, strap it up, take anti-inflammatories, and keep it completely still. In the short term it may not hurt (because there is no load), but as soon as activity resumes the pain returns, on and off for years. Worse, long stillness can make the joint capsule contract, dragging a rotator-cuff problem toward the stiffness of a frozen shoulder.
Why it is wrong:
Chronic rotator-cuff pain is fundamentally degeneration with few inflammatory cells, so an anti-inflammatory drug has almost nothing to act onThe shoulder is especially sensitive to stillness. It relies on muscle for stability at every moment, so once it stops moving, rotator-cuff and shoulder-blade strength drop quickly and the capsule tightens, a vicious circleIt also clashes with modern care for acute injuries: even for acute soft-tissue injuries, the POLICE approach (Bleakley 2012) has replaced the old Rest with Optimal Loading, because complete stillness only slows healing
The right approach:
Do not stop loading; set the load in a suitable range (no more than 3-5 out of 10 during training, and no worse within 24 hours) and keep increasing it gradually within that rangeFor a while, avoid the high-arc overhead movements that make it worse (relative rest), but keep moving through the rest of the range. Do not put the whole arm in storageGive the tendon gradually increasing tension with eccentric, heavy slow resistance, and isometric work, while training shoulder-blade stability to restore the rhythm
So whether it is rotator-cuff pain or the thawing phase of a frozen shoulder, the shoulder is almost always a story of move, but move smart, not sling it and leave it alone. The only cases that truly need the arm held still and a doctor seen first are the ones listed in the chapter on when you must see a doctor.
Chapter 6
When to see a doctor
Pain in the left shoulder or arm with chest tightness, breathlessness, a cold sweat, or nausea, especially if it gets worse with effort and does not change when you move or press the shoulder, may be pain referred from the heart. Call emergency services immediately; do not rub it like a stiff neck. A shoulder that is red, swollen, and hot, with fever or chills, may be a joint infection. Go to the emergency department right away. After an injury, if the shoulder looks deformed or you cannot lift the arm, it may be dislocated or broken. Do not try to push it back yourself. Weakness or numbness that keeps getting worse; constant pain at night with weight loss, a low fever, night sweats, or a lump you can feel; or pain that still gets worse after several weeks of proper loading: see a doctor soon. This site does not replace a doctor's diagnosis.
References · 12
- Sophia Fox, A. J., Bedi, A., & Rodeo, S. A. (2009). The basic science of articular cartilage: structure, composition, and function. Sports Health, 1(6), 461–468. 10.1177/1941738109350438
- Yamamoto, A., Takagishi, K., Osawa, T., Yanagawa, T., Nakajima, D., Shitara, H., & Kobayashi, T. (2010). Prevalence and risk factors of a rotator cuff tear in the general population. Journal of Shoulder and Elbow Surgery, 19(1), 116-120. 10.1016/j.jse.2009.04.006
- Rangan, A., Brealey, S. D., Keding, A., Corbacho, B., Northgraves, M., Kottam, L., et al. (2020). Management of adults with primary frozen shoulder in secondary care (UK FROST): a multicentre, pragmatic, three-arm, randomised controlled trial. The Lancet, 396(10256), 977-989. 10.1016/S0140-6736(20)31965-6
- Kukkonen, J., Joukainen, A., Lehtinen, J., Mattila, K. T., Tuominen, E. K. J., Kauko, T., & Aarimaa, V. (2015). Treatment of nontraumatic rotator cuff tears: a randomized controlled trial with two years of clinical and imaging follow-up. Journal of Bone and Joint Surgery (Am), 97(21), 1729-1737. 10.2106/JBJS.N.01051
- Beard, D. J., Rees, J. L., Cook, J. A., Rombach, I., Cooper, C., Merritt, N., et al. (2018). Arthroscopic subacromial decompression for subacromial shoulder pain (CSAW): a multicentre, pragmatic, parallel group, placebo-controlled, three-group, randomised surgical trial. The Lancet, 391(10118), 329-338. 10.1016/S0140-6736(17)32457-1
- Ketola, S., Lehtinen, J., Arnala, I., Nissinen, M., Westenius, H., Sintonen, H., et al. (2009). Does arthroscopic acromioplasty provide any additional value in the treatment of shoulder impingement syndrome? A two-year randomised controlled trial. Journal of Bone and Joint Surgery (Br), 91(10), 1326-1334. 10.1302/0301-620X.91B10.22094
- Alfredson, H., Pietilä, T., Jonsson, P., & Lorentzon, R. (1998). Heavy-load eccentric calf muscle training for the treatment of chronic Achilles tendinosis. The American Journal of Sports Medicine, 26(3), 360–366. 10.1177/03635465980260030301
- Kongsgaard, M., Kovanen, V., Aagaard, P., Doessing, S., Hansen, P., Laursen, A. H., Kaldau, N. C., Kjaer, M., & Magnusson, S. P. (2009). Corticosteroid injections, eccentric decline squat training and heavy slow resistance training in patellar tendinopathy. Scandinavian Journal of Medicine & Science in Sports, 19(6), 790–802. Single-blind RCT, 39 men with patellar tendinopathy: corticosteroid injection, eccentric decline squats or heavy slow resistance for 12 weeks. All groups improved at 12 weeks; at the half-year follow-up the gains held with eccentric and HSR training but deteriorated in the corticosteroid group. HSR had the highest satisfaction and increased collagen turnover (abstract, PMID 19793213). 10.1111/j.1600-0838.2009.00949.x
- Bleakley, C. M., Glasgow, P., & MacAuley, D. C. (2012). PRICE needs updating, should we call the POLICE? British Journal of Sports Medicine, 46(4), 220-221. Argues complete rest (PRICE's 'R') impairs healing; proposes Protection + Optimal Loading + Ice + Compression + Elevation (POLICE) as the modern acute-injury framework. 10.1136/bjsports-2011-090297
- Rio, E., Kidgell, D., Purdam, C., Gaida, J., Moseley, G. L., Pearce, A. J., & Cook, J. (2015). Isometric exercise induces analgesia and reduces inhibition in patellar tendinopathy. British Journal of Sports Medicine, 49(19), 1277–1283. 10.1136/bjsports-2014-094386
- Kelley, M. J., Shaffer, M. A., Kuhn, J. E., Michener, L. A., Seitz, A. L., Uhl, T. L., Godges, J. J., & McClure, P. (2013). Shoulder pain and mobility deficits: Adhesive capsulitis. Clinical practice guidelines linked to the International Classification of Functioning, Disability, and Health from the Orthopaedic Section of the American Physical Therapy Association. Journal of Orthopaedic & Sports Physical Therapy, 43(5), A1-A31. APTA clinical practice guideline for the stiff, painful shoulder; its differential-diagnosis section is what separates a shoulder that needs loading from one that needs a physician. 10.2519/jospt.2013.0302
- Gray, J. C. (2012). Visceral referred pain to the shoulder. In Physical therapy of the shoulder (pp. 267-304). Elsevier. Why the heart, diaphragm, gallbladder and lung apex can present as shoulder pain, and the features that separate referred visceral pain from a musculoskeletal source - notably that it does not change when the shoulder is moved or pressed. 10.1016/b978-1-4377-0740-3.00012-x