Frozen Shoulder in Menopause: It's Not Just Hormones

What is frozen shoulder, and how is it different from ordinary shoulder pain?

It's a thickening and contracture of the capsule that surrounds the shoulder joint, which physically shrinks the space the joint has to move in. That matters because it isn't a muscle problem or a wear-and-tear problem — and it has one specific sign that separates it from almost everything else.

That sign is the loss of passive external rotation. In plain terms: someone else holds your elbow at your side and rotates your forearm outward, away from your body. With most shoulder problems, that movement is available even if it hurts. With frozen shoulder, it is physically blocked.

This is worth knowing before your appointment for one practical reason: it's a test someone has to do to you, not one you can do on yourself. If you lift your own arm, other muscles compensate and the restriction hides. So “I can still move it” isn't evidence either way, and it's not the thing to report.

The UK trial I'll come to later used “restriction of passive external rotation” of at least 50% as its entry criterion. That's how central this one movement is to the diagnosis.

Why does it land on women in their forties and fifties?

Nobody knows, and the honest position in the literature turns out to be more interesting than a confident guess would have been. The condition is still formally described as idiopathic — meaning of unknown cause — while at the same time showing a markedly narrow age-and-sex pattern that nobody has managed to explain.

A 2026 study in Climacteric opens by stating the puzzle in a single sentence: adhesive capsulitis “is considered idiopathic, yet typically affects women aged 40–60 years.” That's the whole mystery. A condition with no known cause that nonetheless picks out women at exactly the age the menopause transition happens.

It is reasonable to look at that and think: hormones. Researchers have thought the same thing, which is why somebody finally went and checked. What they found is the most useful part of this article, and it is not what you'd expect from how the internet discusses it.

Does hormone therapy protect against frozen shoulder?

The honest answer is that we don't yet know. The most direct attempt to test it that I could find did not show a significant effect, and I want to walk through its numbers carefully — because this is a study whose headline is very easy to misuse in either direction.

Look at what's tempting there. “Twice the odds” is a real number from a real dataset, and a headline writing itself around it would say hormone therapy halves your risk of frozen shoulder. The authors say the opposite of that, in their own plain-language summary: the difference “was not significant. A larger sample size is needed.”

And look at the study's own title. It describes itself as “a preliminary pilot study to address design issues” — its purpose was partly to work out how many women a proper study would need to recruit. It was built to size the next study, not to answer the question.

So that's the hormone question: open, being investigated properly, and currently not a reason to do anything differently. Which would be a frustrating place to end an article — except that while researchers are unsure about estrogen, they are not remotely unsure about something else.

What's the link with diabetes, and why does it matter to you?

This is the part I'd want a friend to read. The association between frozen shoulder and diabetes turns up across different study designs — its size is disputed, its existence much less so — and it runs in a direction most people don't consider. Not just “diabetes raises your risk,” but the reverse.

Read that 30% slowly. It means that if you line up ten women who have just been told they have a frozen shoulder, around three of them have diabetes — and some of those three do not yet know it. The shoulder got there first.

I want to be careful about the size of the risk, though, because it varies a lot between study designs — and the biggest numbers come from the least reliable designs. A more recent and more rigorous review makes that visible.

Those two stat-cards disagree, and I'd rather show you that than pick the dramatic one. Case-control studies start with people who already have the condition and look backwards, which tends to inflate associations. Cohort studies follow people forward and are harder to fool.

So the truthful summary is: diabetes clearly raises the risk, probably by something between a third more and a few times more, and the researchers themselves say better studies are needed to pin it down.

But here's why the exact multiplier doesn't change what you should do. The 30% figure isn't about how much diabetes raises your risk — it's about how common diabetes turns out to be among people who already have a frozen shoulder. That number is what justifies the test, and it holds whichever risk estimate you prefer.

What else should be checked?

Your thyroid, and this one comes with a strongly worded instruction rather than a suggestion. The same endocrine thread that connects frozen shoulder to diabetes also connects it to thyroid dysfunction, and both are conditions that become more common in exactly this decade of life.

A 2021 review puts it plainly: frozen shoulder is “a condition often associated with diabetes and thyroid dysfunction, and which should always be investigated in patients with a primary stiff shoulder.” Not considered. Always investigated.

That phrasing matters because of what usually happens instead. A frozen shoulder is an orthopaedic-feeling problem, so it gets an orthopaedic-feeling response: rest it, move it, injection, physio. The two blood tests that the literature says should always be done can slip past the entire appointment, because nobody in the room is thinking about blood.

And if either comes back abnormal, you have gained something far more valuable than a shoulder explanation. Undiagnosed type 2 diabetes and an untreated thyroid are both worth finding on their own terms, years before they would otherwise have surfaced.

And if you already know you have diabetes?

Then this section doesn't drop out of your path — it changes shape. Two things are worth doing rather than one.

Ask for a current HbA1c anyway, because the shoulder is a reason to look at how the diabetes is actually running right now. And flag the diabetes before any steroid injection, because an injection into the shoulder joint measurably moves blood glucose.

That is not a reason to refuse the injection — it's the treatment that held its own against surgery in the trial further down this page. It's a reason to tell whoever manages your diabetes that it's happening, and to expect a higher reading the next day rather than be alarmed by one.

Why is it worst at night?

Because of what happens to the joint when you lie still and when you lie on it — and because night pain is not an incidental detail of this condition. It's so characteristic that the UK's National Health Service lists it among the two defining symptoms, with the sleep consequence named explicitly.

The NHS describes the symptoms as “pain in your shoulder which can be worse at night and disturb sleep,” alongside “stiffness in your shoulder which makes it difficult to move.” On a site about midlife sleep, that's the sentence that earns this article its place.

Two things drive it. Lying on the affected side compresses a joint whose capsule is already tight and inflamed, which is why so many women describe being woken the moment they roll over in their sleep. And during the day, movement keeps things loose in a way that stillness doesn't — so the shoulder stiffens overnight and the first movement of the morning is the worst one.

This is also where frozen shoulder quietly becomes a sleep problem rather than a shoulder problem. Months of being woken several times a night has its own consequences, and they compound the ones the menopause transition is already producing. Our guides to menopause fatigue and getting back to sleep at 3 a.m. are about that half of the problem.

How long does it actually last?

Longer than anyone wants to hear, and the honest sources decline to give a tidy figure. The NHS says frozen shoulder “can take months or years to get better,” adding that “the pain and stiffness will usually go away eventually.” Note the two hedges in that sentence: usually, and eventually.

The 2021 review is more specific, and more encouraging: “most patients respond very well to combination of conservative treatment resulting in gradual resolution of symptoms in 12–18 months.” It also notes that the traditional staging of frozen shoulder into phases has a duration that “is not constant and varies with the intervention.”

That last point is worth pulling out, because the three-phase story — freezing, frozen, thawing — gets presented online as a fixed timetable you can count down. The phases are a useful planning tool for clinicians, not a schedule. Yours won't necessarily take the length someone's chart says it should.

What actually works?

Here the evidence is unusually clear, and unusually reassuring if you've been told you might need an operation. A large UK trial compared the two surgical options against the non-surgical one, and the result was not what the people who run shoulder clinics expected.

Sit with that for a second. The least invasive option — twelve sessions of structured physiotherapy with a steroid injection — performed about as well at a year as having your shoulder capsule surgically divided under general anaesthetic. And the surgery that performed marginally best on the score was also the one that hurt the most people.

That doesn't make surgery pointless. The 2021 review describes the usual sequence: resistant cases that haven't responded to conservative treatment for six to nine months may be offered surgery. The trial's message isn't “never operate” — it's that starting with physiotherapy and an injection is not a lesser option you're being fobbed off with.

One caution on the manipulation route, from that same review: manipulation under anaesthesia “could result in unwarranted complications like fractures of humerus or rotator cuff tear.” The trial found it the most cost-effective of the three, but it is not risk-free, and that's a conversation to have rather than a box to tick.

What if it isn't frozen shoulder?

Shoulder pain in your fifties has several causes and they get confused with each other constantly, partly because “frozen shoulder” has become the folk name for any stiff painful shoulder. The distinguishing question is the one from the top of this article: is passive external rotation genuinely blocked, or just painful?

Things that commonly get called frozen shoulder and aren't:

• Rotator cuff problems. Pain on lifting and weakness, but the joint can usually still be rotated outward by someone else. This is a common mix-up.
• Osteoarthritis of the shoulder. Stiffness too, but with a grinding quality and a different pattern on imaging.
• Neck problems referring pain down. If the pain travels past your elbow, or comes with tingling or numbness in the hand, the neck is a likelier source than the shoulder.
• Polymyalgia rheumatica. This one matters because it's a different kind of condition with a different treatment, and it looks superficially similar in exactly this age group. The NHS describes “pain and stiffness in the shoulders, neck and hips, which develops quickly over a few days or weeks,” adds that “both sides of the body are usually affected,” and that the stiffness is worst on waking and “usually starts to improve after about 45 minutes as you become more active.” Frozen shoulder is typically one shoulder, and it doesn't loosen off within the hour the way the polymyalgia stiffness does.
• Secondary frozen shoulder. The same capsular contracture, but following an injury, surgery or a period of immobility. For women in this age group, breast surgery, axillary node surgery and radiotherapy to that area all belong on that list — and if that's your history, say so unprompted, because it changes the diagnosis from “no known cause” to one with a known trigger.

What about the other shoulder?

It's one of the first questions people ask, and the one I can give the least satisfying answer to. Frozen shoulder can affect the second shoulder, either at the same time or later, and it's reported more often in people with diabetes — but I couldn't find a figure I'd trust enough to put a number on it here.

What I will say is that both shoulders going stiff together, at the same time, in your fifties, is the pattern that should make polymyalgia rheumatica part of the conversation rather than assuming it's two frozen shoulders at once.

When should you get help now?

Most stiff painful shoulders in midlife are not dangerous, and this list isn't here to make you anxious about an ordinary one. It's that a few specific patterns need seeing today or this week rather than going into a physiotherapy queue, and they're recognisable.

Emergency, today

• Shoulder pain with chest tightness or discomfort, breathlessness, sweating or nausea. The US National Heart, Lung, and Blood Institute lists “pain in the shoulder, back, or arm” among the symptoms women are more likely to have during a heart attack, and notes these “can happen together with chest pain or without any chest pain.” Pain spreading to the jaw or neck is a recognised heart attack symptom in anyone. This is an ambulance question, not an appointment question.
• A shoulder that looks deformed, or that you genuinely cannot move at all, or shoulder pain straight after a fall or injury.
• Pain with fever or feeling systemically unwell — an infected joint is rare but it is urgent.
• New weakness or numbness in the arm or hand, as opposed to pain limiting what you do.

This week, not in six weeks

• You have, or have had, cancer — and this is new night-time shoulder or arm pain. I've spent this article explaining that night pain is the hallmark of frozen shoulder, so I need to say the other half: in someone with a cancer history, new bone pain that is worse at night is also how secondary deposits present. It doesn't mean that's what this is. It means your history changes who should look at it and how soon.
• Pain that doesn't ease in any position, including lying still, or that is steadily worsening week on week rather than settling into a pattern.
• Unexplained weight loss, persistent fever or night sweats alongside the shoulder.
• Stiffness in both shoulders, neck and hips at once, worse in the morning — the polymyalgia pattern, which responds to a completely different treatment.

What should you do first?

Five things, in an order that deliberately puts the urgent and the cheap ahead of the slow. None of this replaces a diagnosis, and a shoulder that has been painful or stiff for months deserves a proper examination rather than a plan you manage on your own from a website.

1. First, rule out the list above. If your shoulder pain comes with chest tightness, breathlessness, sweating or nausea, this list is not for you — that's an emergency assessment now. If it followed a fall, comes with fever, or there's new weakness in the arm, be seen the same week. And if you have a cancer history and this is new night pain, say that first, before anything else on this page.
2. Get the external rotation checked by someone else. It's the sign that separates this from the conditions it's confused with, and it's a thirty-second examination. Say that's what you want looked at — and if your stiffness is in both shoulders, your neck and your hips, ask about polymyalgia rheumatica by name instead — and if you also have a new headache, jaw pain on chewing or any change in your sight, that is today, not an appointment.
3. Ask for the HbA1c and the TSH. Thirty percent of people with this diagnosis have diabetes, and the review language is “should always be investigated.” This is the step most likely to be skipped and the one most likely to change something that matters. If you already have diabetes: still ask for a current HbA1c, and tell them before any steroid injection.
4. If step 2 confirmed the blocked rotation, start the physiotherapy and treat it as the real treatment. On the best trial available it held its own against surgery at a year, and twelve sessions with a steroid injection was the comparator nothing beat clinically. If step 2 pointed somewhere else — neck, cuff, polymyalgia — this isn't your treatment, because that trial only enrolled people with genuine capsular restriction. And if your frozen shoulder followed surgery or an injury, say so, because that history belongs in the plan.
5. Deal with the pain and the nights while you wait. Unaffected side, pillow hugged in front, pillow behind your back. Ask specifically what you should be taking for pain, and for how long — months of untreated night pain is its own injury, and a year-long condition deserves a proper answer on analgesia. That's a question for your doctor or pharmacist rather than something to work out from a website.

And if you're already six to nine months in with proper conservative treatment and nothing has changed, that's the point the surgical conversation genuinely opens up — not a reason to keep waiting out a timetable.

Where to go from here

If the broader joint aches are the bigger problem, menopause joint pain at night covers the generalised version, and hip pain at night the other joint that commonly flares in midlife. If it's the sleep damage that's wearing you down, start with menopause fatigue. And our honest look at HRT covers the hormone decision on its actual merits, which is where it should be made.

For how all of this fits into the bigger picture of midlife sleep, start at the complete perimenopause and menopause sleep guide.

Frequently asked questions