Two Ways to Replace a Shoulder, and Which Problem Each Solves

What each implant actually is

An anatomic shoulder replacement rebuilds the joint the way it was drawn. A metal ball replaces the worn head of the arm bone; a smooth cup resurfaces the socket on the shoulder blade. The parts sit where the originals sat. A reverse replacement inverts that: the ball is anchored to the shoulder blade and the cup to the arm bone. Same joint, mirrored.

That inversion is not a stylistic preference. Moving the ball to the socket side shifts the joint's centre of rotation inward and slightly down, which lengthens the leverage the deltoid — the big muscle capping the outside of the shoulder — has over the arm. The reverse design is built so the deltoid can drive an arm the rotator cuff no longer can. Nearly everything else follows from that single idea.

Anatomic and reverse describe the geometry of the implant. Neither word is a judgement about the severity of your problem or the quality of the surgery.

Why the rotator cuff decides which one you are offered

An anatomic replacement is essentially a resurfacing job. It gives the joint a new bearing surface and hands the movement back to your own tendons. If those tendons are intact, that works as designed. If they are torn beyond repair, a new bearing surface changes nothing about the arm you cannot lift — the joint is smooth, and still will not go overhead.

Rotator cuff tears are a common cause of shoulder pain, accounting for nearly two million visits a year in the United States, and most tears do not heal on their own 1Ref 1American Academy of Orthopaedic Surgeons (OrthoInfo) (2024).Rotator Cuff Tears.That rotator cuff tears are a common cause of shoulder pain (nearly 2 million US visits per year), that many are managed nonsurgically with anti-inflammatories, injections, and physical therapy, and that most tears do not heal on their own..

Rotator cuff tears account for nearly 2 million US visits a year, and most do not heal on their own 1Ref 1American Academy of Orthopaedic Surgeons (OrthoInfo) (2024).Rotator Cuff Tears.That rotator cuff tears are a common cause of shoulder pain (nearly 2 million US visits per year), that many are managed nonsurgically with anti-inflammatories, injections, and physical therapy, and that most tears do not heal on their own..

That second clause is the one that matters here. A torn tendon does not knit itself back together with time. It can be repaired, compensated for, or lived with — but the tissue does not restore itself, and a cuff that has been gone for years stays gone. A shoulder in that state is exactly the shoulder the reverse design was drawn for: put an anatomic implant into it and you have resurfaced a joint that still has no engine.

A torn cuff is not, by itself, a reason to replace anything

Here the reasoning turns counterintuitive. The same tendon tear that can eventually make a reverse replacement the sensible option is, in most people at most points in time, not a surgical problem at all. Many cuff tears are managed nonsurgically with anti-inflammatory medication, injections, and physical therapy 1Ref 1American Academy of Orthopaedic Surgeons (OrthoInfo) (2024).Rotator Cuff Tears.That rotator cuff tears are a common cause of shoulder pain (nearly 2 million US visits per year), that many are managed nonsurgically with anti-inflammatories, injections, and physical therapy, and that most tears do not heal on their own. — and where that has been tested head to head, the results are sobering.

In a randomised trial of nontraumatic supraspinatus tears followed for two years, physiotherapy alone, acromioplasty plus physiotherapy, and cuff repair plus physiotherapy produced no significant clinical difference, and conservative care was a reasonable initial option 2Ref 2Kukkonen J, Joukainen A, Lehtinen J, et al. (2015).Treatment of Nontraumatic Rotator Cuff Tears: A Randomized Controlled Trial with Two Years of Clinical and Imaging Follow-up.That physiotherapy alone, acromioplasty plus physiotherapy, and rotator cuff repair plus physiotherapy produced no significant clinical difference at two years for nontraumatic supraspinatus tears, and that conservative care is a reasonable initial option.. A Cochrane review of the same question landed in a similar place: rotator cuff repair, with or without subacromial decompression, probably provides little or no clinically important benefit over non-operative exercise-based treatment for pain and function 3Ref 3Karjalainen TV, Jain NB, Heikkinen J, et al. (2019).Surgery for rotator cuff tears.That rotator cuff repair, with or without subacromial decompression, probably provides little or no clinically important benefit over non-operative exercise-based treatment for pain and function..

A tear on a scan is a finding, not a plan. What decides anything is what the shoulder does, how much it costs you to live with, and what has genuinely been tried — not the word "tear" in a report.

Which matters here in a specific way: a cuff tear found on imaging is not a countdown clock, and it is not a reason to book anything.

The shoulder problems that are not replacement problems

Before any of this becomes a replacement conversation, the diagnosis has to be right, and several common shoulder problems produce exactly the same sentence — I cannot lift my arm and it hurts at night — without involving the joint surface at all. Two are worth naming, because both are routinely mistaken for something that needs an implant.

• Frozen shoulder. It progresses through freezing, frozen, and thawing stages and usually resolves over one to three years, with range-of-motion physical therapy as the primary treatment 4Ref 4American Academy of Orthopaedic Surgeons (OrthoInfo) (2024).Frozen Shoulder (Adhesive Capsulitis).That frozen shoulder progresses through freezing, frozen, and thawing stages and usually resolves over one to three years, with range-of-motion physical therapy as the primary treatment — used here as the differential that most resembles an arthritic shoulder.. A shoulder that has seized up is not a shoulder that has worn out, and time is doing more of the work than any implant would.
• Subacromial pain. Arthroscopic subacromial decompression provided no clinically important benefit over placebo — an arthroscopy with no decompression performed — or over no treatment for subacromial shoulder pain 5Ref 5Beard DJ, Rees JL, Cook JA, et al. (CSAW) (2018).Arthroscopic subacromial decompression for subacromial shoulder pain (CSAW): a multicentre, pragmatic, parallel group, placebo-controlled, three-group, randomised surgical trial.That arthroscopic subacromial decompression provided no clinically important benefit over placebo (arthroscopy only) or over no treatment for subacromial shoulder pain — used to show that shoulder operations differ sharply in what they have been tested against..

"Shoulder surgery" is not one thing with one track record. One of the commonest shoulder operations of the last generation did not beat its own placebo. That is not an argument against replacement — it is an argument for asking which operation you are being offered, and what that specific one has been compared against.

A painful, stiff, weak shoulder has several possible authors, and they have different answers. Getting the author right is the whole job of the appointment.

How anyone knows whether it worked

Shoulder outcomes are measured with questionnaires rather than with X-rays, because how a joint looks and how an arm works come apart routinely. The DASH — Disabilities of the Arm, Shoulder and Hand — is a validated self-reported measure of symptoms and physical function across upper-extremity musculoskeletal disorders 6Ref 6Hudak PL, Amadio PC, Bombardier C (Upper Extremity Collaborative Group) (1996).Development of an upper extremity outcome measure: the DASH (disabilities of the arm, shoulder, and hand).That the DASH is a validated self-reported measure of symptoms and physical function across upper-extremity musculoskeletal disorders, used here to explain how shoulder outcomes are actually tracked., and it asks about the things people actually came in for.

You will meet its relatives: the oxford shoulder score and the ases shoulder score cover overlapping ground in different words. Which one gets used matters less than that one is used, twice.

The practical point: filling one out before the decision, not only after it, gives you the only before-and-after you will ever have. Without a baseline, "better" is a memory — and memories of pain are famously unreliable. A score from six months ago is a fact. Your recollection of six months ago is a story.

What to ask the surgeon

The comparison you are actually being asked to make is not anatomic versus reverse. It is: what is my joint surface doing, what is my cuff doing, and what does each of those facts rule in or out? A surgeon who has reached the decision from your imaging and your examination will answer these directly and without irritation.

• What is the state of my rotator cuff, and how do you know? Whether a tendon is repairable is a different question from whether it is torn, and only the first one shapes the implant choice.
• Which implant are you planning, and what would change your mind on the day? Some decisions are finalised once the joint is open. Worth knowing in advance whether yours is one.
• What motion should I expect back, and what will I not get back? An honest answer names the losses alongside the gains.
• What has not been tried yet, and why? If a genuine course of physical therapy is missing from the record, that is better surfaced before an implant is ordered than after.
• What does failure look like, and what is the second operation? Every implant has a revision path attached. Asking about yours is planning, not pessimism.
• What will this cost, itemised? The surgeon's fee, the facility fee, the implant, and the anaesthesia are commonly billed separately, so a single quoted figure is rarely what arrives.

When replacement is clearly the right call

Nothing on this page is an argument against replacing a shoulder. That is worth stating plainly, because the evidence quoted above has a boundary and it is very easy to carry it past that boundary. Those trials studied tendon tears — people whose cuff hurt and whose joint surface was intact. They did not study people whose joint has worn through, and their results do not transfer to that person.

This is the single most important sentence here: a replacement is the operation for a surface problem. The cuff evidence is about tendons. They are different targets, and a trial run on one says nothing about the other. Reading "physiotherapy matched surgery" and concluding that a destroyed glenohumeral joint should be managed with exercise is a category error, and it costs people years of a life they did not need to lose.

The picture that moves the conversation decisively toward replacement is one your surgeon should be able to say out loud: the joint surface itself is the problem rather than the tendon around it; nonsurgical care has genuinely been tried rather than nominally mentioned; the shoulder is limiting the life you actually live rather than an activity you might resume; and the cuff status is clear enough to determine which implant, because that is the question the cuff answers.

The question was never whether to avoid the operation. It is whether your problem is the surface, the tendon, or both — because only one of those three is solved by an implant.

The wider indications question — shoulder replacement indications, and the closely argued hip replacement indications alongside it — is its own subject, and better read before the appointment than during it. So is reverse shoulder replacement recovery, which is worth understanding before you plan the months that follow.