Revision Shoulder Replacement: When a Replaced Shoulder Has to Be Redone

A revision shoulder replacement is the operation that redoes a replaced shoulder which has worn out, loosened, become unstable, broken around the implant, or become infected, and it is a bigger, harder operation than the first one.¹ Most people never need one: pooled registry and study data put overall survival at around 90% still in place at 10 years, with the risk of revision running at roughly 1% per year after that.²

I did not want to read about this before my own reverse replacement, and then, three weeks after it, I found myself asking the surgeon what happens if it goes wrong in fifteen years. He gave me a straight answer, which is more than I could find written down anywhere. This is that answer set out properly, checked by a consultant shoulder surgeon. If you have not read the overview yet, start with what a shoulder replacement is, and for the durability figures behind all of this see how long a shoulder replacement lasts.

What a revision actually is

A revision means removing some or all of the original components and replacing them, sometimes with a different type of implant, and dealing with whatever caused the failure at the same time.¹ It is not simply a second first operation. The surgeon is working through scar tissue, in bone that has been drilled and prepared once already and may have lost stock around a loose component, and often with soft tissue that is thinner and less forgiving than it was.

That is why the same surgeons who describe a primary replacement as routine get careful when the word revision comes up. The planning is longer, the imaging is more detailed, the operation itself takes more time, and the range of things that might be found once the shoulder is open is wider. Some revisions swap only one part, a worn plastic socket, say. Others take out the whole construct and rebuild it, occasionally in two stages when infection is involved.

Why a replaced shoulder fails

The recognised reasons are loosening or wear of a component, later failure of the rotator cuff or of the socket (glenoid) in an anatomic replacement, a break around the implant (periprosthetic fracture), instability or dislocation, and infection.¹ They are not equally likely, and they arrive on very different timescales.

Loosening and wear are the slow ones. An anatomic total replacement depends on the rotator cuff, so a cuff that gives way years later can leave the components with nothing driving or steadying them, which shows up in registry and cohort analyses of durability and revision as a genuine late failure route.³ Instability is more of a reverse-replacement problem: dislocation is more common after a reverse than after an anatomic replacement.⁴ Fracture around the implant tends to arrive suddenly, usually after a fall. Infection is the one that can appear at any point, affects roughly 1 in 100 primary replacements and somewhat more after a reverse, and is set out on its own in shoulder replacement infection.

How likely is a revision?

In the first ten years, unlikely: pooled UK registry and study data put overall survival at around 90% still in place at that mark, so most replaced shoulders are quietly getting on with the job.² Beyond the decade the picture changes from “most are fine” to a small, steady, accumulating risk of roughly 1% per year.

The arithmetic matters more than the percentage. A shoulder replaced at 75 has to survive perhaps fifteen or twenty years; one replaced at 55 may need to last thirty-five, running through that annual risk many more times. That is the whole reason surgeons weigh a younger patient’s pain now against a greater lifetime chance of a revision later, which is the argument laid out in shoulder replacement at what age. It is also why a first operation done well, by someone who does shoulders in volume, is worth more than any amount of worrying about the second.

Converting a total or a partial to a reverse

One of the commonest revision routes is not replacing like with like but converting a failed anatomic total replacement or a hemiarthroplasty into a reverse replacement, because a reverse works without a functioning rotator cuff.⁴ If the cuff is the thing that failed, putting in another anatomic implant hands the new components the same problem that broke the old ones.

A reverse switches the geometry so the deltoid lifts the arm instead of the cuff, which is exactly why it earns its place as a salvage option as well as a first operation.⁵ The trade-offs are the ones any reverse carries: excellent power to lift the arm, a real risk of limited internal rotation, and a higher overall complication rate. If you are being offered this conversion and the language is unfamiliar, the plain comparison sits in an anatomic versus a reverse shoulder replacement.

Recovery after a revision

The recovery follows the same shape as a first replacement, a sling for a period, then passive movement, then active movement, then strengthening, but surgeons commonly protect a revision for longer and the strength returns more slowly.⁶ For a primary replacement the sling is typically on for about 2 to 6 weeks and the joint keeps gaining movement and strength over 6 to 12 months; after a revision, expect the same stages with more caution built in and a longer tail.⁷

Physiotherapy matters even more here than it does the first time, because the tissue around the joint has been through two operations and stiffens readily. The staged programme is described in full in physiotherapy after shoulder replacement, and the honest warning I would give is the same one that applies to a first replacement, only more so: judge the arm at six months, not at six weeks.

What results to expect, honestly

Revisions are usually good at settling pain and giving a stable, usable arm, and less predictable for reach and strength, with a higher complication rate than a first replacement.¹ That is the trade a surgeon is asking you to accept, and it is a reasonable one when the alternative is a painful, failing joint.

The ceiling is set mostly by what is left to work with: how much bone remains around the socket and the top of the arm bone, and what state the soft tissue is in. A shoulder that has lost a lot of glenoid bone is a harder rebuild than one where a plastic surface has simply worn. This is one of the places where a surgeon’s shoulder volume and revision experience genuinely change the answer, which is the whole argument in choosing a shoulder surgeon. Ask directly how many revisions they do a year, and what they expect your arm to do afterwards rather than what they hope.

What to do if you think yours is failing

New pain in a replaced shoulder that has been comfortable, a loss of movement you had, a sense of the joint giving way, or any redness, fever or wound discharge is a reason to be seen rather than to wait it out.⁷ Pain that arrives years after everything settled is not something to attribute to age or the weather; it is a symptom your surgical team will want to image.

None of this is a reason to be frightened of the first operation. I am three years out and my own shoulder is quiet, and the numbers say it is likely to stay that way for a long while yet. Knowing what a revision is, and that it exists as a real option rather than a dead end, made me easier about the whole thing, not more anxious. A replaced shoulder is a durable joint on a slow clock, and if that clock ever runs out there is a second operation that can be done. Whether it should be, and which one, is a judgement for a surgeon holding your imaging, not a decision anyone can make from a page.