Frozen Shoulder: Why It Happens, How Long It Lasts, and When You Need Help

A man in his early fifties sat across from me last month, cradling his right arm as if it were on loan from someone else. He could not reach the back pocket of his trousers. He could not fasten his seatbelt without wincing. Combing his hair had become a small, daily humiliation. He had not fallen, not lifted anything heavy, not done a thing he could point to. “Doctor,” he said, “it just froze.” That word, frozen, is exactly right, and it is the name we give the condition: frozen shoulder, or in the textbooks, adhesive capsulitis.

It is one of the most misunderstood problems I see. Patients are frightened by how stiff and painful the shoulder becomes, and they are often told contradictory things: that it will vanish on its own, that they need urgent surgery, that an injection will cure it overnight. The truth is more interesting and, in some ways, more reassuring. Let me walk you through what is actually happening inside the joint, how long it tends to last, why diabetes matters so much here in North Kerala, and what the evidence says about the treatments that help.

What is actually going on inside a frozen shoulder?

Your shoulder joint sits inside a soft envelope called the capsule. In a healthy shoulder that envelope is loose and roomy, which is why the shoulder is the most mobile joint in the body. In a frozen shoulder, that capsule becomes inflamed and then thickened, tight and scarred. It shrinks around the joint like cling film pulled too tight. That is why the shoulder both hurts and refuses to move, and why no amount of pushing through the pain makes it loosen any faster. The stiffness is mechanical, built into the tissue itself.

In most cases it appears for no clear reason at all, which we call primary or idiopathic frozen shoulder. It typically arrives in middle age, most often between the late forties and the sixties, and it affects women somewhat more often than men. In one large long-term study the average age at onset was about 53, with women outnumbering men by roughly three to two, and about one in five people eventually developed it in both shoulders, though usually not at the same time.³ Sometimes it follows a period when the arm was kept still, after another injury or an operation, and then we call it secondary. It is a genuinely common condition, with national guidance estimating it affects on the order of a couple of people per thousand each year.⁹

The three phases, and why this matters for your expectations

The single most useful thing I can teach a patient is that a frozen shoulder is not one thing but a sequence. It moves through three overlapping phases, and what helps in one phase is not the same as what helps in another. Understanding which phase you are in changes everything about what to expect and what to do.

Add those phases together and you can see why this is a marathon, not a sprint. The freezing phase alone can run from six weeks to nine months. The frozen phase typically lasts four to six months. Thawing, the slow return of movement, can take anywhere from six months to two years. The American Academy of Orthopaedic Surgeons describes these as the broad durations of the three stages,¹ and clinical reviews note that with good conservative care the majority of patients see their symptoms gradually resolve over roughly twelve to eighteen months.¹ But I want to be honest with you: those are typical ranges, not a calendar I can hold you to. Some shoulders move faster, some considerably slower.

Why does diabetes matter so much here?

If you live in North Malabar, there is a fair chance diabetes runs in your family or in your own life, and this is directly relevant. The connection between diabetes and frozen shoulder is one of the strongest and best-established in shoulder medicine. A systematic review and meta-analysis pooling several studies found that people with diabetes have roughly three to four times the odds of developing a frozen shoulder compared with people who do not have diabetes; the pooled estimate put the odds at about 3.7 times higher.⁵ Thyroid problems show a similar association.²

This is why, when a patient walks into my clinic at Aster MIMS Kannur with a stiff, painful shoulder and no injury to explain it, one of the first things I do is ask about and often check their blood sugar. It is also why I am candid that a frozen shoulder in a person with diabetes can be more stubborn and slower to settle than one in a person without it. None of this means the shoulder will not recover. It means we go in with realistic expectations and, ideally, with the blood sugar reasonably controlled, which is good for the whole body in any case.

What the evidence actually says about treatment

Here is where I separate what genuinely helps from what is merely sold as helping. The most important principle, endorsed by the British Elbow and Shoulder Society in its 2025 frozen shoulder pathway, is a step-up approach: start with the simplest, least invasive measures, and escalate only if the shoulder does not respond.⁹ You do not jump to surgery for a condition that, in most people, settles with patience and the right support.

Physiotherapy and pain control come first. Keeping the shoulder gently moving within the limits of comfort, controlling pain so you can sleep and function, and working with a physiotherapist form the backbone of care. The aim is not to force the joint open but to protect what movement you have and coax it gently as the capsule allows.

The steroid injection has a real but time-limited role. An injection of corticosteroid into the joint is one of the more useful tools I have, and the evidence is reasonably clear about exactly what it does and does not do. A meta-analysis of randomised trials found that an intra-articular corticosteroid injection gives significantly better pain relief than placebo in the early window, roughly the first eight weeks, but that this pain advantage is no longer evident by around nine to twenty-four weeks; range of motion, though, improved more in the injection group both early and later.⁶ A randomised trial in patients specifically in the early freezing phase confirmed that a well-placed steroid injection improved pain, movement and function.⁷ So the injection is at its best early, when inflammation is at its peak, and it is most valuable when it buys you a calmer shoulder that can actually do physiotherapy. It is a head start, not a cure, and there is good trial evidence that in the short term an injection outperforms hands-on manual therapy and exercise alone.¹¹

Hydrodilatation is an option, but I am honest about its limits. This procedure involves injecting fluid into the joint under pressure to stretch the tight capsule. It is offered within the step-up pathway,⁹ and some patients do feel it helps. But I will not oversell it: a systematic review and meta-analysis concluded that, across the available trials, hydrodilatation had only a small effect on pain and movement that did not reach what most would call a clinically meaningful difference.⁸ I discuss it as a reasonable intermediate step for the right patient, not as a guaranteed solution.

When does a frozen shoulder actually need surgery?

Rarely, and only after the simpler measures have been given a genuine chance. The shoulders I consider for surgery are the stubborn minority that remain painfully stuck despite a fair trial of conservative care, generally after about six to nine months without adequate improvement.^1,2 Escalating to a procedure before that, in most cases, means intervening on a shoulder that may well have settled on its own.

When surgery is warranted, there are two main approaches. The first is manipulation under anaesthesia, where, with you asleep and your muscles relaxed, the surgeon moves the shoulder firmly through its range to break down the tight, scarred capsule. The second is arthroscopic capsular release, a keyhole operation in which the surgeon precisely cuts the thickened capsule under direct vision. Studies comparing the two find they give broadly similar results.¹⁰ The trade-off is that manipulation, because it relies on force, carries a small but real risk of complications such as a fracture of the arm bone or a tear of the rotator cuff,² which is why many surgeons, myself included, often favour the control of a keyhole release for the right patient. Either way, the operation is only the start; what makes it succeed is the structured physiotherapy that follows.

So what should you take from all this?

If your shoulder has stiffened and become painful without an obvious cause, the odds are good that this is a frozen shoulder and that, with patience and sensible management, it will improve. Understand which phase you are in. Get your pain under control so you can sleep and so you can do physiotherapy. If you have diabetes, take the blood sugar seriously, both for your shoulder and for everything else. Use a steroid injection for what it is genuinely good at, early relief that lets rehabilitation happen, rather than expecting it to be a magic cure. And keep surgery in reserve for the small number of shoulders that truly refuse to thaw.

What I would gently warn against is the opposite extreme: doing nothing, suffering in silence for a year because someone said it would “sort itself out,” and never getting the pain controlled or the diagnosis confirmed. A frozen shoulder deserves to be assessed properly, if only to rule out the other things that can mimic it.