Rheumatoid Arthritis in the Shoulder

Rheumatoid arthritis is famous for the hands. The swollen knuckles, the stiff morning fingers, the classic X-ray of the wrist — that is the picture in every textbook and every pamphlet. So when the shoulder is the joint that starts waking you at 3 a.m., or the joint that stops you reaching a top shelf, it can feel like your disease has gone off-script.

It has not. The shoulder is one of the most commonly involved large joints in rheumatoid arthritis, and it is one of the most functionally expensive. A stiff finger is an annoyance. A shoulder that cannot get your hand above your head, behind your back, or across your body takes away dressing, washing your own hair, driving comfortably, and sleeping on that side. People consistently rate shoulder loss as more disabling than the joint's reputation suggests.

This page is about the shoulder specifically — what is going wrong inside it, how to tell rheumatoid shoulder pain from the three conditions it gets confused with, what imaging is worth doing, what injections genuinely deliver, and what the surgical options really mean when they are eventually offered. The general disease — diagnosis, blood tests, drug classes, systemic complications — lives on the main Rheumatoid Arthritis page. This one assumes you already have the diagnosis and the shoulder is now the problem.

Table of Contents

1. Why the Shoulder Comes Later — and Costs More
2. The Shoulder Is Not One Joint: Four Structures That Matter
3. What Rheumatoid Shoulder Actually Feels Like
4. Telling It Apart: RA vs Cuff, Frozen Shoulder and OA
5. Rotator Cuff Attrition and Cuff Tear Arthropathy
6. Imaging: When Ultrasound and MRI Beat a Plain X-ray
7. The Function You Actually Lose
8. Injections: What They Do and Do Not Do
9. Systemic Control Is Still the Main Lever
10. Movement, Physiotherapy and Joint Protection
11. When Surgery Enters the Picture
12. Hemi, Anatomic Total, and Reverse — the Real Difference
13. Realistic Outcomes and How Long Implants Last
14. Sleeping, Dressing and Daily-Living Adaptations
15. The Neck Complication Nobody Warns You About
16. Key Research Papers
17. Connections
18. Featured Videos

Why the Shoulder Comes Later — and Costs More

Rheumatoid arthritis is a synovial disease. It attacks the lining membrane that lubricates joints and tendon sheaths, and it starts where that lining is most abundant relative to the size of the joint — the small joints of the hands, the wrists, the forefoot. Those joints have a lot of synovium packed into a small space, so a small amount of inflammation produces obvious, early, visible swelling. That is why hands announce the disease.

The shoulder is different in three ways that delay its arrival:

1. It is a big joint with a roomy capsule. A rheumatoid shoulder can accumulate a surprising volume of inflammatory fluid before the capsule is tight enough to hurt. The disease has often been quietly present for a while before you notice it.
2. It hangs. The shoulder is not weight-bearing in the way a knee or hip is, and gravity gently distracts it rather than compressing it. That buys time.
3. It has enormous spare range. The glenohumeral joint has more motion than daily life demands. You can lose a substantial slice of that range before anything you actually do becomes impossible — so the loss is masked until it crosses a threshold, then seems to arrive all at once.

The delay is real, but it is not protection. It is a head start for the disease. By the time a rheumatoid shoulder hurts enough to complain about, the pannus — the invasive, thickened, inflamed synovial tissue that is RA's signature lesion — has often already been chewing on cartilage, on the bone at the joint margins, and, critically, on the rotator cuff tendons. That last part is what makes the rheumatoid shoulder a different animal from an arthritic shoulder in someone without RA.

The classic clinical and radiological survey of the painful rheumatoid shoulder — Ennevaara's study of 200 cases, published in 1967 and still cited — established the pattern that modern imaging keeps confirming: shoulder involvement in RA is common, frequently under-recognised, and involves the soft tissues around the joint as much as the joint surface itself.

The Shoulder Is Not One Joint: Four Structures That Matter

When a doctor says "your shoulder," they are compressing at least four separate structures into one word. Knowing which one hurts changes what helps. Think of the shoulder as a golf ball on a tee (the ball is the humeral head, the tee is the shallow glenoid socket), with the whole tee bolted to a shoulder blade that itself slides freely on the back of the ribcage.

1. The glenohumeral joint — the ball and socket

This is the joint people mean. It is a true synovial joint with a large capsule, and it is a prime RA target. Pain here is deep, felt in the upper arm as much as the shoulder, and it hurts through the whole arc of motion rather than at one point. As cartilage is lost, motion becomes gritty and rotation — especially turning the palm outward with the elbow at your side — goes early.

2. The acromioclavicular (AC) joint — the small joint on top

Where the collarbone meets the bony point of the shoulder blade. It is small, it is synovial, and RA loves it. It is also badly under-diagnosed, because a painful AC joint produces pain right on top of the shoulder that people assume is coming from the main joint. The giveaway is that you can put one finger on the sore spot — glenohumeral pain cannot be pointed to that precisely — and that reaching across your body to the opposite shoulder is exquisitely painful. AC joint pain often responds well to a small, precisely placed injection, which is why identifying it is worth the effort.

3. The subacromial bursa — the cushion under the roof

A flat, fluid-filled sac between the rotator cuff tendons and the bony roof above them. Its job is to let the cuff glide. In RA it can fill with inflammatory fluid and thickened synovium, and a boggy subacromial bursa is one of the most consistent ultrasound findings in a painful rheumatoid shoulder. This is the structure that produces the classic "painful arc" — pain through the middle third of raising the arm out to the side, conventionally about 60 to 120 degrees, so it starts a little below shoulder height, is worst around it, and eases again near the top.

4. The scapulothoracic articulation — the sliding shoulder blade

Not a true joint at all: it is a broad gliding interface where the shoulder blade slides on the back of the ribcage, cushioned by bursae. RA does not attack it directly the way it attacks a synovial joint, but it matters enormously, because it is where compensation happens. When the ball-and-socket joint stiffens, the shoulder blade takes over — it swings and tilts further to get your arm up. That compensation is genuinely useful, and a good physiotherapy programme deliberately trains it. But an overworked scapulothoracic interface can develop its own bursitis, its own snapping and grinding, and its own pain in the muscles between the shoulder blade and spine.

One more to keep in the back of your mind: the sternoclavicular joint, where the collarbone meets the breastbone, is also synovial and also an RA target. Pain there is often mistaken for something cardiac or costochondral. It is usually tender to direct pressure, which most chest-wall pain of concern is not.

What Rheumatoid Shoulder Actually Feels Like

The pattern people describe is fairly consistent, and recognising it is the first diagnostic step:

• Night pain that wakes you, especially lying on that side. This is close to universal and it is the symptom that drives people to finally get help. Lying on the shoulder compresses an already-inflamed bursa and joint. Lying on the other side lets the painful arm drop forward and drag on the capsule.
• Morning stiffness measured in tens of minutes, not seconds. An inflammatory shoulder is stiffest after rest and loosens with gentle use. A mechanical shoulder is the reverse — better after rest, worse with use.
• Deep, poorly located ache in the upper arm. Often felt down to the deltoid insertion, roughly where a shirt sleeve badge sits. Pain below the elbow is not typical and should raise the question of the neck instead.
• Loss of external rotation first. Elbow tucked to your side, turn your forearm outward like opening a gate. This dies early in an inflamed or contracted shoulder and is one of the most useful things you can self-monitor.
• Weakness that is disproportionate to the pain. If lifting a kettle at arm's length feels not just painful but genuinely unsupported, that suggests the rotator cuff is failing rather than simply sore.
• It often becomes bilateral. RA is a symmetrical disease. Even when one shoulder is far worse, the other usually shows something on imaging — which matters, because it removes the option of "just doing everything with the other arm."

Telling It Apart: RA vs Cuff, Frozen Shoulder and OA

Here is the trap: having rheumatoid arthritis does not mean every shoulder pain is rheumatoid. RA patients get ordinary rotator cuff tendinopathy, ordinary frozen shoulder, and ordinary osteoarthritis, and RA actively increases the risk of the first two. So the real question is usually not "which one is it" but "which ones are present, and which is driving the pain right now."

The single most useful bedside test is the active-versus-passive comparison. Lie down, relax the arm completely, and have someone else lift and rotate it for you. If they can move it much further than you can under your own power, the problem is a failing tendon, not a locked capsule. If they cannot move it any further than you can — particularly if external rotation is blocked with a firm, wooden stop — you are dealing with a contracted capsule, and no amount of cuff strengthening will fix that.

And the honest complication: in a long-standing rheumatoid shoulder, all four processes can be present at once. The capsule can be contracted from months of guarding, the cuff can be thinned by pannus, the cartilage can be gone, and the AC joint can be independently inflamed. Which is exactly why a targeted diagnostic injection — put anaesthetic in one structure and see what changes — is often more informative than another scan.

Rotator Cuff Attrition and Cuff Tear Arthropathy

This is the part of rheumatoid shoulder disease that most distinguishes it from garden-variety arthritis, and it deserves plain explanation.

The rotator cuff is four flat tendons that wrap the ball of the shoulder like a hand cupping a grapefruit. Their real job is not power — the deltoid does the lifting — it is centring. They hold the ball down and pressed into its shallow socket so that when the big deltoid muscle fires, the ball rotates in place instead of sliding upward. Without a functioning cuff, the deltoid's pull simply drags the humeral head up toward the bony roof of the shoulder.

In RA, the cuff is attacked from two sides at once. Inflamed synovium inside the joint erodes the tendons from underneath, and inflamed bursal tissue above erodes them from the top. The tendons thin, fray, and eventually perforate. This process is called attrition, and it is deliberately not called a "tear" — nothing snapped. There was no injury. The tendon simply dissolved over years, which is why rheumatoid cuff defects are often broad, ragged, and poorly suited to surgical repair. There is frequently not enough healthy tendon left to stitch to anything.

Once the cuff is gone, the shoulder enters the state Charles Neer named cuff tear arthropathy: a massive irreparable cuff defect plus destruction of the joint surface. Its radiographic signature is unmistakable once you know it — the humeral head rides up and articulates against the underside of the acromion, which over years remodels into a smooth concave surface, and the humeral head itself rounds off. Radiologists describe this as the acromion becoming a second socket and the humeral head taking on a hip-like shape. Functionally it means the arm can no longer be raised at all by its own muscles, even though the joint may move freely when someone else lifts it — a state often called a pseudoparalytic shoulder.

Two consequences follow, and they are the reasons the rest of this page is written the way it is:

1. Cuff status, not cartilage status, is what determines which operation you can have. A shoulder replacement that relies on the cuff to work will fail in a shoulder that has no cuff. This is the entire rationale for the reverse prosthesis described below.
2. Systemic disease control protects the tendon, not just the cartilage. Every month of uncontrolled synovitis is a month of tendon being digested. This is a strong argument against tolerating a "mildly grumbling" shoulder for years.

Imaging: When Ultrasound and MRI Beat a Plain X-ray

An X-ray of a painful rheumatoid shoulder is worth having, but it answers a narrow question. It shows bone and the gap between bones. It shows almost nothing about the tendons and bursa, which in RA are often where the trouble is.

What the plain X-ray does tell you

• Joint space. RA narrows it uniformly — the whole gap shrinks at once, because the pannus attacks the whole surface. Osteoarthritis narrows it unevenly, usually worst where the load is.
• Erosions. Small bites taken out of the bone at the joint margins, where the pannus first reaches bone.
• Thin bone around the joint (periarticular osteopenia) — an early and characteristic inflammatory finding.
• Glenoid erosion. Crucially for surgical planning, RA erodes the socket medially, boring inward toward the shoulder blade rather than wearing it back-to-front the way OA does. Medial erosion consumes the bone a future glenoid implant would need to anchor into.
• Superior migration of the humeral head. Measured as the acromiohumeral distance — the gap between the top of the ball and the underside of the bony roof. The commonly quoted normal range is roughly 7 to 14 millimetres; a narrowed gap is a strong indirect sign that the cuff is no longer holding the head down, and a very narrow one suggests a large or massive cuff defect. Radiographs are a crude tool for this, but it is a genuinely useful measurement that costs nothing extra, and it changes the surgical conversation entirely. A narrative review of superior migration in chronic cuff disease sets out the mechanics in detail, and a study of acromiohumeral distance after cuff repair confirms the link between the measurement and whether the cuff is actually intact.

When ultrasound earns its place

Ultrasound is the workhorse for the rheumatoid shoulder, and it is under-used. It is cheap, it needs no radiation and no contrast, it can be done in the clinic in minutes, and it is dynamic — the sonographer can move your arm and watch the tendons glide (or fail to). It shows:

• fluid and thickened synovium in the subacromial bursa and the joint itself;
• active inflammation on power Doppler — blood flow within synovium, which distinguishes an actively inflamed joint from an old, burnt-out, mechanically damaged one. This is the single most useful thing ultrasound adds, because it separates "this will respond to anti-inflammatory treatment" from "this is structural damage and needs a mechanical solution";
• full-thickness rotator cuff defects, which it detects with accuracy broadly comparable to MRI in trained hands;
• a real-time target for a guided injection, in the same appointment.

A study comparing ultrasound findings of the shoulder in RA directly against physical examination illustrates the core point: examining a shoulder with your hands is not a reliable way to know what is inside it. Things are routinely found that clinical examination did not predict.

When MRI is the right call

MRI is the better test when the surgical decision is close, when the question is bone rather than tendon, or when ultrasound cannot see far enough. Specifically:

• Fatty infiltration of the cuff muscles. This is MRI's decisive contribution. When a tendon has been non-functional for years, its muscle belly is gradually replaced by fat — and fatty muscle does not recover even if the tendon is repaired. This finding, more than the size of the tear, tells a surgeon whether repair is pointless and a reverse replacement is the honest option.
• Bone marrow oedema and hidden erosions not yet visible on X-ray.
• Glenoid bone stock and version before surgery, where CT is often added as well.

What MRI is not good for is monitoring. If the question is "is this shoulder inflamed today," ultrasound with Doppler answers it faster and cheaper.

The Function You Actually Lose

Shoulder disability is measured in a strangely specific way, because the shoulder's failures are specific. Three arcs of motion carry most of daily life. Measured in degrees, an inflamed or contracted shoulder follows what is classically called the capsular pattern: external rotation is the most restricted, elevation next, and internal rotation the least restricted of the three. What people notice, though, tracks function rather than degrees — and because so many daily jobs hang off the hand-behind-the-back arc, a modest loss there is felt long before a larger loss of overhead reach is. Both orderings are worth knowing, because they answer different questions: the first tells a physiotherapist what the capsule is doing, the second tells you what you are about to lose.

1. External rotation goes first, on both counts. Elbow at your side, forearm swinging outward. Losing it means you cannot reach the far side of your own head to wash or brush hair, cannot easily reach across the car to the seatbelt, cannot get a hand to your opposite ear on the phone, and cannot hold a steering wheel comfortably at the top.
2. Reaching behind the back is what people miss next. This is the "hand behind the back" arc — a blend of internal rotation, extension and adduction: fastening a bra, tucking in a shirt, reaching a back pocket, wiping after the toilet, scratching your own back, getting a coat sleeve on. In pure range terms internal rotation is the last of the three to be severely limited, but it carries the most daily tasks, so even a partial loss bites early. It is the arc people find most humiliating to lose, and the one they are least likely to volunteer in a consultation. It is worth naming explicitly, because it changes the treatment target.
3. Overhead reach is the most visible loss. Top shelves, hanging laundry, changing a bulb, putting a bag in an overhead locker, washing hair. Elevation is usually restricted earlier than the hand-behind-back arc when someone measures it, but it degrades gradually and the shoulder blade quietly compensates, so it is often reported last.

Then there are the losses that do not fit an arc at all. Carrying becomes painful long before range is lost, because holding a weight at arm's length loads the cuff hard. Sleep is often the first casualty and the last to return — and lost sleep amplifies pain perception and worsens fatigue, so a painful shoulder becomes a whole-body problem faster than a painful knee does. And pushing up out of a chair loads the shoulder in a way most people never think about until they cannot do it, which matters especially if the hips or knees are also affected.

An older evidence summary of suprascapular nerve block described it reducing chronic shoulder pain and disability in people with degenerative disease or rheumatoid arthritis, which is a reminder that when a shoulder proves stubborn, targeted pain interventions beyond the standard steroid injection do exist and are worth asking about.

Injections: What They Do and Do Not Do

A corticosteroid injection is the most common intervention offered for a painful rheumatoid shoulder, and it is genuinely useful — provided everyone is honest about what it is for.

What it does

It puts a potent anti-inflammatory directly into the inflamed tissue at a concentration you could never achieve with tablets. It reduces synovial inflammation and swelling, and it therefore reduces pain and often improves range within days. A common preparation is around 40 mg of triamcinolone or methylprednisolone mixed with a local anaesthetic, though the dose is adjusted for the target — a small AC joint takes far less than a glenohumeral joint.

The anaesthetic is not just for comfort. It is a diagnostic tool. If numbing the subacromial bursa abolishes your pain within minutes, the bursa was the pain source. If it does nothing, the pain is coming from the joint itself, or the AC joint, or the neck — and you have learned something more valuable than the steroid will deliver.

What it does not do

• It does not modify the disease. It suppresses one joint's inflammation temporarily. The underlying rheumatoid process continues, and so does the cartilage and tendon damage.
• It does not last indefinitely. Relief typically runs weeks to a few months. If it lasts only days, that is diagnostic information — usually meaning the pain is predominantly structural rather than inflammatory, and that a different plan is needed.
• It does not rebuild a tendon. Repeated steroid injections around a tendon are generally believed to weaken it further, and in a rheumatoid shoulder where the cuff is already being eroded, that is not a trivial concern. The widely used rule of thumb — no more than three or four injections into the same joint in a year, spaced out — is a sensible caution rather than a hard law, but the principle behind it is sound.
• It does not fix stiffness on its own. An injection into a contracted capsule buys a window of reduced pain. If nobody uses that window to do range-of-motion work, the shoulder stiffens straight back up. The injection and the physiotherapy are a package; either alone underperforms.

Two safety points specific to a rheumatoid patient

• Joint infection is rare, but you are not an average patient. Septic arthritis after an injection is uncommon, and it is the complication that matters. Rheumatoid patients are usually on drugs — biologics, JAK inhibitors, corticosteroids, sometimes methotrexate — that both raise the risk of infection and blunt the fever and redness that would normally announce it. The rule is simple: a shoulder that becomes hotter, more swollen and more painful in the days after an injection rather than better should be assessed the same day, not waited out. A short-lived post-injection flare in the first day or two, easing after that, is common and is a different thing.
• Injected steroid is absorbed into the body. It is not a purely local drug. Blood glucose commonly rises for several days afterwards, which matters if you have diabetes and is worth planning for; flushing, a few nights of poor sleep, and a temporary mood lift or dip are also common and settle.

Guided or blind?

This is a live question and the answer differs by target. Landmark-guided (blind) injections into the subacromial bursa are usually reasonably accurate, because the bursa is a large flat space directly under an easily felt bony landmark. Randomised comparison of ultrasound-guided against landmark-guided subacromial injection has been published in sports medicine journals, and the overall message from that literature is nuanced: image guidance reliably improves the accuracy of needle placement, but the difference it makes to symptom outcomes is smaller and less consistent than that accuracy advantage suggests.

The picture changes for harder targets. Getting a needle reliably inside the glenohumeral joint without imaging is considerably less certain, and the AC joint is a small, obliquely angled slot that is easy to miss. For those two targets — and in any shoulder where the anatomy is distorted by years of rheumatoid destruction — asking for ultrasound guidance is entirely reasonable. Guidance also brings a diagnostic bonus: the operator sees the bursa and the cuff on the way in, and you leave knowing more than you arrived with.

Systemic Control Is Still the Main Lever

It is easy, when one joint dominates, to start thinking locally — injections, physiotherapy, braces, eventually surgery. But the shoulder is a window onto the disease, not a separate problem. A shoulder that is actively synovitic is telling you the disease is not adequately suppressed, and every month of that is more cartilage and more tendon lost.

The specific relevance for the shoulder is the cuff. Cartilage loss can eventually be answered with a joint replacement. Cuff loss narrows which replacement is possible, and it cannot be undone. Protecting the tendon is therefore an argument for treating a shoulder flare as a reason to review systemic therapy rather than simply injecting the joint and moving on. The drug options — methotrexate and the conventional DMARDs, the biologics, the JAK inhibitors, and where steroids fit — are covered on the DMARDs and Biologics page.

The adjuncts matter too, though they are adjuncts and not substitutes: the evidence around omega-3 fats is among the better-supported dietary interventions in RA — it can genuinely reduce tender joints and morning stiffness and lets some people take fewer NSAIDs, but it has never been shown to slow the erosion an X-ray measures, which is exactly the damage a shoulder cannot recover from — and dietary pattern more broadly is covered on the Anti-Inflammatory Diet page. This is the one place to say it plainly: none of that replaces disease-modifying drugs, and a shoulder is a poor joint to experiment on, because what it loses it does not get back.

Movement, Physiotherapy and Joint Protection

The instinct with a painful shoulder is to stop using it. That instinct is wrong, and it is wrong in a specific and costly way: a shoulder held still stiffens faster than almost any other joint in the body. The capsule contracts, and you convert a painful shoulder into a painful and frozen one.

A shoulder programme for rheumatoid arthritis has to do three different jobs, and they run in this order:

1. Preserve range, every single day. This is non-negotiable and it is the highest-value thing you can do. Gentle, unloaded range work — pendulum swings with the arm hanging, assisted movement using the good arm or a stick, sliding the hand up a wall, lying on your back and letting gravity help the arm overhead. It should not be forced into significant pain. It should be frequent and short: several small sessions beat one heroic one.
2. Strengthen what still works, without hammering what does not. The scapular stabilisers — the muscles that control the shoulder blade on the ribcage — can almost always be trained safely even when the ball-and-socket joint is a wreck, and they are the compensation mechanism the shoulder falls back on. Rotator cuff strengthening is valuable where the tendons are intact but should be light, controlled, and kept below shoulder height, typically with a band and the elbow at the side.
3. Change how you load it. Keep the elbow below shoulder height for anything heavy. Carry things close to the body rather than at arm's length — the load on the shoulder rises steeply with distance from the body, which is why a light kettle held out straight can hurt more than a heavy bag held against your hip. Push rather than pull. Use two hands. Slide objects along a counter instead of lifting them.

General principles and a fuller programme are on the Exercise and Joint Protection page. The shoulder-specific addition is that range work should continue even during a flare, at reduced intensity. Strengthening can pause; range work should not.

When Surgery Enters the Picture

Shoulder surgery in rheumatoid arthritis is offered for one reason above all: pain that no longer responds to medication, injection, or activity change — particularly pain that destroys sleep. Function usually improves too, but pain relief is the reliable outcome and the honest headline.

The triggers that usually start the conversation:

• night pain that has stopped responding to injections;
• X-ray showing the joint space effectively gone;
• an inability to raise the arm that comes from cuff failure rather than pain;
• a shoulder that has become the limiting factor on independence — dressing, washing, driving.

The timing argument that rarely gets made clearly enough: in rheumatoid arthritis there is a cost to waiting that does not exist in osteoarthritis. RA erodes the glenoid medially — it bores inward, thinning and hollowing the socket. A glenoid implant needs bone to anchor into. Wait long enough and there may not be enough left, which removes the better reconstructive options and leaves only the ones with less reliable pain relief. Similarly, waiting while the cuff dissolves removes the anatomic replacement from the menu. This is a genuine reason to have the surgical conversation earlier than instinct suggests — not to have the operation earlier, necessarily, but to know where you stand and what is still on the table.

There are also lesser operations for narrower problems. Synovectomy — clearing out inflamed synovium, usually arthroscopically — can help a shoulder with persistent synovitis but preserved cartilage. Excision of the outer end of the collarbone can work very well for an isolated AC joint that is the dominant pain source — but the qualifier is load-bearing: it is the diagnostic injection that makes the case, because operating on an AC joint that was not actually generating the pain changes nothing. Neither is a substitute for replacement in a destroyed joint, but both address real problems that a replacement would be an over-treatment for.

Hemi, Anatomic Total, and Reverse — the Real Difference

Three operations get called "shoulder replacement," and they are not variations on a theme. They are mechanically different solutions to different problems. Understanding which one you are being offered, and why, is the single most useful piece of preparation you can do.

Hemiarthroplasty — replace the ball only

A metal ball on a stem replaces the humeral head. The socket is left alone. It is a smaller operation with nothing to loosen on the socket side, and it is chosen mainly when the glenoid cannot take an implant — because the bone has been eroded away, which in RA is common. Its weakness is straightforward: a metal ball still grinding against a damaged, arthritic socket does not relieve pain as completely or as durably as replacing both surfaces. In rheumatoid arthritis specifically, where the socket is usually diseased as well, hemiarthroplasty tends to be a compromise driven by bone stock rather than a first choice. A large series from the Mayo Clinic compared total shoulder arthroplasty against hemiarthroplasty in 303 consecutive rheumatoid shoulders and remains a key reference for that comparison.

Anatomic total shoulder arthroplasty — replace both, keeping the natural geometry

A metal ball on the humerus and a plastic socket on the glenoid, in the same arrangement nature used: ball on the arm, socket on the shoulder blade. When it is the right operation it is a very good one — pain relief is excellent and motion is usually close to natural.

It has one absolute prerequisite: a functioning rotator cuff. The prosthesis reproduces normal anatomy, and normal anatomy depends on the cuff centring the ball in the socket. Implant an anatomic total shoulder into a cuff-deficient shoulder and the deltoid drags the new ball upward, the head grinds against the acromion, and the plastic socket component levers loose — a failure mode surgeons describe as "rocking horse" loosening. It is a predictable failure, which is why cuff status is assessed so carefully before this operation is chosen. Long-term follow-up of the classic Neer total shoulder prosthesis established the durability that this design is judged against.

Reverse total shoulder arthroplasty — swap the ball and socket over

This is the operation that changed what is possible in the rheumatoid shoulder, and it is worth understanding properly.

The reverse prosthesis puts the ball on the shoulder blade (a metal hemisphere called a glenosphere, bolted into the glenoid) and the socket on the arm. It sounds like an eccentric idea until you see what it does mechanically. Reversing the parts moves the joint's centre of rotation inward and downward, which lengthens the deltoid's leverage — the muscle now pulls on a longer arm and can lift the arm on its own. The design also converts the deltoid's upward pull into useful rotation instead of destructive migration, because a fixed ball on the socket side cannot ride upward the way a loose humeral head can.

The result: it works without a rotator cuff. That is the whole point, and it is precisely why it matters in rheumatoid arthritis, where the cuff is so often gone. A shoulder that could not be raised at all before surgery can often be raised afterwards, powered entirely by deltoid. The early experience with the Grammont reverse prosthesis in rheumatoid patients with non-reconstructible cuff lesions is one of the papers that established this indication.

The trade-offs are real and should be stated:

• Rotation is less reliably restored than elevation. Raising the arm forward and out usually improves well; turning it outward and reaching behind the back improve less predictably, because those movements depended on the cuff you no longer have. Reaching a back pocket may remain difficult.
• The complication rate is higher than for anatomic replacement. Reverse-specific problems include scapular notching (the arm-side component contacting and eroding the edge of the shoulder blade), instability, and acromial stress fractures — the last being of particular concern in RA, where bone is often thin from disease and steroids.
• It is a bigger commitment in a younger patient. RA patients often reach this point earlier in life than typical cuff-arthropathy patients, which means more years for an implant to survive and a higher likelihood of eventually needing revision. A systematic review specifically pooled complications and implant survivorship for reverse replacements in people under 65 — an age band a great many rheumatoid patients fall into.

Realistic Outcomes and How Long Implants Last

What to expect, stated as honestly as the evidence allows:

• Pain relief is the strong, dependable result. Across designs and diagnoses, shoulder replacement is very effective at relieving arthritic pain, and it is the outcome most likely to meet expectations. If the goal is to sleep through the night again, the odds are good.
• Motion improves, but less dramatically than pain. Expect meaningful, useful gains — and do not expect a normal shoulder. Reverse replacements in particular buy back elevation more reliably than rotation.
• Implants last a long time but not forever. Long-term series of shoulder replacements report survivorship well past a decade for many patients, which is the relevant time frame for most people receiving one. The honest caveat is that "survivorship" counts implants that have not been revised, not shoulders that feel perfect.
• Rheumatoid patients carry extra risk on both ends. Thin bone, thin skin, and immunosuppressive medication all raise the risk of fracture during surgery, of wound-healing problems, and of infection. Managing DMARDs and biologics around the operation — when to pause, when to restart — is a genuine planning task, not a formality, and it needs the rheumatologist and surgeon talking to each other rather than each assuming the other has handled it. This is the one point on this page where you should insist on an explicit answer rather than accepting a general reassurance. Declare your supplements alongside your prescriptions while you are at it: fish oil and turmeric both have a mild antiplatelet effect, and while the trial evidence has not shown fish oil increasing surgical bleeding — so this is usually a disclosure rather than a reason to stop — it carries more weight if you also take warfarin, a direct oral anticoagulant, or an antiplatelet drug. Let the surgical team make that call rather than a supplement label.
• The other joints shape the plan. If the elbow or hand on the same side also need surgery, or if a hip or knee replacement is coming that will require crutches, the order of operations matters — a replaced shoulder should not be loaded through a crutch handle. Sequencing is worth raising early.

Sleeping, Dressing and Daily-Living Adaptations

These are the changes that improve life next week rather than next year.

Sleep — usually the biggest single win

• Do not lie on the painful shoulder. Obvious, but worth stating because people try to tough it out.
• Lying on the good side needs support. Left unsupported, the painful arm falls forward across the body and hangs, dragging on the capsule all night. Hug a pillow — a full body pillow or a folded duvet — so the sore arm rests on top of something at roughly chest height, elbow supported, rather than dangling.
• Lying on your back often works best. Put a small folded towel or thin cushion under the upper arm, just behind the elbow, so the arm is not pulled backward into extension by gravity. This one adjustment is frequently the difference between waking and not waking.
• A wedge pillow or recliner helps a lot of people. Sleeping semi-upright reduces shoulder pain for many, and it is a cheap thing to test before buying anything.
• Do not sleep with the arm overhead. It compresses the structures under the bony roof for hours.

Dressing

• Painful arm in first, out last. The universal rule for a stiff limb. Dress the sore side first while the garment is loose; undress it last, after the other arm is free and the fabric has slack.
• Front-opening beats overhead. Button shirts, zip tops and cardigans avoid the overhead arc entirely. This is the single most effective clothing change.
• Front-fastening bras, or fasten at the front and rotate. Sports-style front-closing designs remove the behind-the-back arc, which is often the most painful.
• A dressing stick and a long shoehorn cost very little and remove a surprising number of daily provocations.

Washing and grooming

• A long-handled sponge or brush removes the reach-behind-the-back arc from bathing.
• A handheld shower head on a slider means you bring the water to you instead of raising the arm.
• Wall-mounted or stand-held hair dryers remove a sustained overhead hold, which is one of the most reliably painful everyday tasks.
• Pump bottles rather than squeeze bottles; shelves at chest height rather than above the head.

Kitchen and carrying

• Move everything you use daily to between hip and chest height. The most effective single afternoon's work you can do for a bad shoulder.
• Lightweight pans; a kettle you can fill from a jug rather than lifting to the tap; slide heavy pots along the counter rather than lifting.
• Carry loads close to the body, and split shopping into more, lighter bags.
• Use a crossbody bag on the unaffected side, or better, a wheeled trolley. A shoulder bag strap on a diseased shoulder compresses exactly the wrong structure.
• Push doors and trolleys rather than pulling them.

Driving

• A seatbelt reacher, or simply reaching across with the good hand, removes a daily jolt of external rotation.
• Both hands lower on the wheel; adjust the seat closer so the arms are not extended.
• Reversing by mirror and camera rather than twisting and reaching over the seat back.

The Neck Complication Nobody Warns You About

Two things about the cervical spine belong on a rheumatoid shoulder page, and they are routinely omitted.

First, the neck can masquerade as the shoulder. Rheumatoid arthritis affects the cervical spine, and an irritated nerve root in the neck refers pain into the shoulder and upper arm in a pattern that can be very hard to distinguish from a shoulder problem. Two clues point to the neck rather than the shoulder: pain travelling below the elbow, especially with pins and needles or numbness in the fingers, and a shoulder that hurts even when it is completely still and someone else moves it through a full painless range. If your shoulder pain came with hand tingling, mention it.

Second, and more important: if shoulder surgery is being planned, the cervical spine must be assessed first. Rheumatoid arthritis can loosen the ligaments that stabilise the top two vertebrae, allowing the uppermost vertebra to slip relative to the one beneath it. Someone in that state having their neck extended for airway management under general anaesthesia is at risk of spinal-cord injury. The risk is well known to rheumatologists and anaesthetists, but checking for it is not automatic: the usual approach is to assess the neck clinically and image it in long-standing, erosive or severe rheumatoid disease, rather than to screen every patient — so it depends on somebody knowing you have RA and thinking to ask. If you are being booked for a shoulder operation, ask directly whether your cervical spine has been assessed. It is a reasonable question and any competent team will welcome it.

Key Research Papers

1. Ennevaara K. Painful shoulder joint in rheumatoid arthritis: a clinical and radiological study of 200 cases, with special reference to arthrography of the glenohumeral joint. Acta Rheumatologica Scandinavica, 1967;12(suppl 11):3–106. The foundational clinical and radiological survey of the painful rheumatoid shoulder.
2. Kim HA, Kim SH, Seo YI. Ultrasonographic findings of the shoulder in patients with rheumatoid arthritis and comparison with physical examination. Journal of Korean Medical Science, 2007;22(4):660. Directly compares what ultrasound shows in a rheumatoid shoulder against what clinical examination shows.
3. Naqvi GA, Jadaan M, Harrington P. Accuracy of ultrasonography and magnetic resonance imaging for detection of full thickness rotator cuff tears. International Journal of Shoulder Surgery, 2009;3(4):94. Head-to-head accuracy of the two imaging tests for full-thickness cuff defects.
4. Zakryś M, Zakryś K, Stupnicki S, et al. Superior migration of the humeral head in chronic rotator cuff disease: a narrative review of pathomechanisms, diagnosis, and treatment. Cureus, 2026. Detailed account of why the humeral head rides upward once the cuff fails, and how it is measured.
5. Kholinne E, Kwak JM, Sun Y, et al. The relationship between rotator cuff integrity and acromiohumeral distance following open and arthroscopic rotator cuff repair. SICOT-J, 2021;7:23. Links the acromiohumeral distance measured on plain film to whether the cuff is actually intact.
6. Cole BF, Peters KS, Hackett L, Murrell GAC. Ultrasound-guided versus blind subacromial corticosteroid injections for subacromial impingement syndrome. The American Journal of Sports Medicine, 2016;44(3):702–707. Randomised comparison of image-guided against landmark-guided subacromial injection.
7. Hackett L, Peters K, Murrell GAC. Ultrasound-guided versus blind subacromial corticosteroid injection for subacromial impingement. Journal of Science and Medicine in Sport, 2014;18:e29. Companion report from the same research group on injection guidance.
8. Symmons D. Suprascapular nerve block reduced chronic shoulder pain and disability in degenerative disease or rheumatoid arthritis. Evidence-Based Medicine, 2004;9(2):50. Evidence summary of a targeted nerve-block option for stubborn shoulder pain including rheumatoid disease.
9. Sperling JW, Cofield RH, Schleck CD, Harmsen WS. Total shoulder arthroplasty versus hemiarthroplasty for rheumatoid arthritis of the shoulder: results of 303 consecutive cases. Journal of Shoulder and Elbow Surgery, 2007;16(6):683–690. The large rheumatoid-specific series comparing replacing both surfaces against replacing the ball alone.
10. Torchia ME, Cofield RH, Settergren CR. Total shoulder arthroplasty with the Neer prosthesis: long-term results. Journal of Shoulder and Elbow Surgery, 1997;6(6):495–505. The long-term follow-up against which anatomic total shoulder durability is benchmarked.
11. Rittmeister M, Kerschbaumer F. Grammont reverse total shoulder arthroplasty in patients with rheumatoid arthritis and nonreconstructible rotator cuff lesions. Journal of Shoulder and Elbow Surgery, 2001;10(1):17–22. Early experience establishing the reverse prosthesis as an option for the cuff-deficient rheumatoid shoulder.
12. Goldenberg BT, Samuelsen BT, Spratt JD, Dornan GJ, Millett PJ. Complications and implant survivorship following primary reverse total shoulder arthroplasty in patients younger than 65 years: a systematic review. Journal of Shoulder and Elbow Surgery, 2020;29(8):1703–1711. Pooled complications and survivorship in the younger age band many rheumatoid patients fall into.

Live PubMed Searches

These links run a fresh search on PubMed each time, so they stay current as new work is published.

1. PubMed: rheumatoid arthritis and glenohumeral joint involvement
2. PubMed: rotator cuff attrition and tears in rheumatoid arthritis
3. PubMed: cuff tear arthropathy and superior migration of the humeral head
4. PubMed: shoulder ultrasound and power Doppler synovitis in RA
5. PubMed: shoulder MRI and fatty infiltration of the rotator cuff
6. PubMed: reverse total shoulder arthroplasty in rheumatoid arthritis
7. PubMed: anatomic total shoulder arthroplasty and glenoid loosening in RA
8. PubMed: corticosteroid injection of the rheumatoid shoulder
9. PubMed: adhesive capsulitis versus inflammatory arthritis
10. PubMed: shoulder physiotherapy and exercise trials in rheumatoid arthritis
11. PubMed: cervical spine instability in RA and anaesthetic risk
12. PubMed: acromioclavicular joint involvement in rheumatoid arthritis

Connections

• Rheumatoid Arthritis — the canonical overview: diagnosis, blood tests, systemic features and prognosis.
• DMARDs and Biologics — the drugs that protect the shoulder's cartilage and tendons.
• Exercise and Joint Protection — range-of-motion work and load management principles.
• Omega-3 and Fish Oil — the best-supported dietary adjunct in RA: modest gains in tender joints and morning stiffness, and NSAID-sparing in some people, but no demonstrated effect on radiographic erosion, so it sits beside a DMARD and never in place of one.
• Anti-Inflammatory Diet — dietary pattern as a supporting measure.
• Herbs and Supplements — what is claimed, what is supported, and what interacts.
• Rheumatology — the full category of rheumatic and autoimmune joint conditions.
• Osteoarthritis — the wear-related arthritis that shoulder RA is most often confused with.
• Osteoporosis — thin bone raises the risk of acromial stress fracture after reverse replacement.
• Vitamin D3 — relevant to bone quality before and after shoulder surgery.
• Magnesium — commonly discussed for muscle cramp and sleep quality alongside chronic joint pain.
• Turmeric — the most-studied anti-inflammatory herb taken alongside conventional RA treatment. Evidence tier: small, short, mostly low-quality randomised trials reporting symptom benefit; nothing showing it slows erosion, and it is an adjunct, never a substitute for a DMARD. Documented harm: gastrointestinal upset, a mild antiplatelet effect, and a well-described pattern of turmeric-associated liver injury — which deserves real caution alongside methotrexate, itself hepatotoxic.

Back to Table of Contents

Featured Videos