The Shoulder Pain That Stopped Me Sleeping: How I Knew It Was Time

Shoulder pain bad enough to stop you sleeping is one of the clearest signs that arthritis or a failing rotator cuff has moved past what painkillers, injections and physiotherapy can hold, and it is the single commonest reason people end up considering a replacement¹. This is the account of the pain itself, and how I finally knew it was time, that I could not find when I was the one lying awake.

For the best part of two years I did everything except deal with it. I blamed the mattress, then my age, then my golf swing, and I rearranged my whole night around a shoulder I would not name as the problem. The tidy medical pages told me what the operation was; none of them described the thing that actually pushed me to it, which was months of not sleeping. If you want the calm overview of the surgery alongside this, the pillar on shoulder replacement lays it out; this piece is the honest run-up to the decision.

Why does a worn shoulder hurt most at night?

A worn shoulder hurts most at night because lying down loads the damaged joint and removes all the small, unconscious ways you rest a painful arm during the day, so there is simply no position that lets it settle. Surgeons treat pain that limits sleep as one of the more telling features when they weigh up whether a joint has reached the end of non-surgical management¹.

In the daytime I could cheat. I held the arm close, I stopped reaching, I let my other hand do the work without noticing I was doing it. Lying down took all of that away. Every time I rolled towards the bad side a grinding ache ran from the joint down towards my elbow and I was awake again. I ended up sleeping upright in an armchair with the arm propped on cushions, which I told myself was temporary for a year and a half. The pain of shoulder arthritis and when replacement earns its place is exactly this: not a dramatic injury, but a slow theft of ordinary rest.

What I tried before anyone mentioned surgery

Before surgery is ever suggested, the standard non-surgical measures come first, in roughly this order: simple painkillers and anti-inflammatories, activity changes and physiotherapy, and a steroid injection into the joint, because a replacement is offered only once those no longer control the pain². Surgeons operate mainly for pain, not stiffness on its own.

I went through all of it. The painkillers took the edge off and then stopped touching it. Physiotherapy helped my posture and did nothing for the grinding. The injection bought me a genuinely good six weeks, and when it wore off the disappointment was worse than before, because I had felt what relief was like and then lost it again. That is a normal path, not a failure, and it matters: the fact that I had honestly tried the alternatives was part of what made me a reasonable candidate, which is the ground covered in am I a candidate for shoulder replacement.

When does the pain become a reason to operate?

Pain becomes a reason to operate when it stops responding to painkillers, injections and physiotherapy, when it is limiting sleep and daily life, and when the changes on imaging match where the pain actually sits. Those three things together, rather than any single one, are what point toward a replacement¹.

That last part surprised me. I had assumed a bad enough scan would decide it, but the surgeon was clear that the picture on the screen only counts when it lines up with the pain I was describing. Plenty of people have worn joints on imaging and little pain, and the operation is for the pain, not the picture. My night pain, my failed injections and a joint that had ground itself smooth all told the same story, and that agreement was the point at which surgery stopped being a threat and started being an option.

The scan that matched the pain, and the cuff that decided the operation

When the imaging finally matched the pain, the detail that mattered most was not the worn joint surface but the state of the rotator cuff, because the cuff, not my age, decides which operation a shoulder needs. A torn, irreparable cuff with arthritis points toward a reverse replacement, where the deltoid muscle lifts the arm, rather than a standard total that depends on a working cuff³.

I had never heard of any of this. I assumed a shoulder replacement was one operation, and that if I needed it, that was that. Instead the surgeon showed me a cuff that was long past repairing and explained that this single fact ruled out the standard total and pointed to a reverse, which is now the most commonly performed type in several national registries⁴. Understanding why is the whole of rotator cuff arthropathy and reverse replacement, and it was the piece of information I most wished I had found while I was still lying awake guessing.

The night I stopped putting the decision off

The decision, when it finally came, was not made in the clinic but on an ordinary bad night, when I accepted that sleeping upright in a chair for eighteen months was not a lifestyle I had chosen but a symptom I had been ignoring. Pain that reliably breaks sleep and no longer answers to the non-surgical measures is precisely the picture surgeons are describing when they say a joint has reached the point of a replacement⁵.

There was no single dramatic moment, just an accumulation. The car park where the arm would not reach the boot. The holiday I spent one-handed. The morning I realised I could not remember the last full night I had slept lying down. What tipped me was not more pain but running out of reasons to call it something else. Once I stopped bargaining with it, the actual choice, to have the operation the surgeon and my cuff had already settled on, felt almost anticlimactic. What came next, the sling weeks and the slow reach back, I have written plainly in my shoulder replacement recovery, honestly.

Was letting it get that bad a mistake?

Letting the pain run that long was a mistake in one sense and not in another: I lost eighteen months of sleep I did not need to lose, but seeing someone earlier would have informed the decision rather than forced it, so nothing was taken off the table by waiting. Most people who go ahead do well, with around 90% reporting good or excellent results and roughly 90% to 95% having less pain and better function afterwards².

That figure is a likelihood, not a promise, and the operation improves movement rather than fully restoring it, so I am careful not to sell it as a cure. What I would tell my earlier self is narrow and practical: night pain that painkillers and physiotherapy no longer touch is worth having looked at, not because a scan forces your hand, but because knowing what is actually wrong is better than eighteen months of blaming the mattress. The pain relief is the part the surgery does best, and for me, after the first restful night lying flat, it was the part that mattered most.