What Is Frozen Shoulder? Causes, Symptoms, and Treatment

Frozen shoulder is a condition in which the tissue surrounding the shoulder joint becomes inflamed and then progressively scarred, leading to severe pain and a dramatic loss of movement. Clinically known as adhesive capsulitis, it typically passes through three recognized stages and resolves in most cases within one to two years, though the reality of long-term recovery is more complicated than that neat timeline suggests.¹ It most commonly strikes people between their forties and sixties, affects women more than men, and has well-established links to diabetes and thyroid disease.

The Three Phases

Frozen shoulder is classically described in three overlapping stages, each with a different character. In the first, called the “freezing” phase, pain tends to be the dominant problem. It often starts gradually, sometimes after a minor injury or for no obvious reason at all, and builds over weeks to months. The shoulder becomes increasingly painful, especially at night, and range of motion starts to shrink. This phase can last anywhere from about two to nine months.

The second stage is the “frozen” phase, where pain may actually ease somewhat but stiffness becomes the defining feature. The shoulder capsule has tightened enough that raising your arm, reaching behind your back, or rotating your shoulder outward all become severely limited. Even passive movement, where someone else tries to move your arm for you, meets a hard stop. This phase can persist for four to twelve months.

In the third stage, “thawing,” range of motion gradually returns. How quickly and completely this happens varies widely. Some people regain nearly full movement within months; others are left with measurable restrictions years later.

What Is Actually Happening Inside the Joint

Your shoulder joint sits inside a capsule made of connective tissue. In frozen shoulder, that capsule becomes the problem. Early on, the lining of the capsule (the synovium) becomes inflamed. Researchers have found elevated levels of inflammatory signaling molecules in the joint capsule tissue, along with increased numbers of immune cells that shift from a normal mix toward one dominated by certain types that drive ongoing inflammation.² Growth factors that promote scarring, particularly one called TGF-beta, are also elevated. Studies have found that TGF-beta can cause significant joint scarring when injected into a joint, and its levels are consistently high in frozen shoulder tissue.³

As the condition progresses, inflammation gives way to fibrosis. The capsule thickens with excess collagen, new blood vessels grow into it, and the tissue develops characteristics similar to what’s seen in Dupuytren’s contracture, a fibrotic condition of the hand. Tissue biopsies from frozen shoulder patients show dramatically increased cellularity and vascularity compared to healthy capsule tissue.⁴ The process is not a simple one-way progression from inflammation to scarring. Researchers have found that genes involved in collagen production and breakdown are active throughout all three phases, suggesting that the capsule is constantly remodeling rather than just laying down scar tissue in a single wave.²

One area that gets particularly affected is the rotator interval, a gap between two rotator cuff tendons at the front of the shoulder. The coracohumeral ligament, which runs through this area, contracts and thickens. This is considered a hallmark of the condition and explains why external rotation (turning the arm outward) is often the most restricted movement.

Who Gets Frozen Shoulder and Why

Frozen shoulder is divided into primary (idiopathic) and secondary forms. Primary frozen shoulder appears without an obvious trigger. Secondary frozen shoulder develops after something identifiable: surgery, a fracture, prolonged immobilization of the arm, or a rotator cuff injury. The biology looks similar in both cases, but secondary frozen shoulder is often more predictable and, in theory, more preventable through early mobilization after shoulder injuries or surgery.

Among the established risk factors, diabetes stands out. A systematic review and meta-analysis found that hyperglycemia promotes a process in which sugars bond to collagen proteins, creating cross-links that stiffen the capsule. These advanced glycation end products accumulate over time, and the persistent high blood sugar in diabetes accelerates the process.⁵ People with diabetes not only develop frozen shoulder more frequently but tend to have more stubborn cases that respond less well to treatment. One study found that having type 2 diabetes roughly doubled the odds of developing frozen shoulder compared with controls without diabetes.⁶

Thyroid disease is another recognized risk factor. A study comparing frozen shoulder patients with both healthy controls and people with rotator cuff tears found that individuals with thyroid disorders had roughly 2.7 times the risk of developing frozen shoulder, with women at particular risk.⁷ Other associated conditions include cervical spondylosis and hyperlipidemia, both of which were identified as independent risk factors in at least one study.⁸ There’s also research linking frozen shoulder to poor sleep quality and even constipation, though these associations are less well established.

The typical patient is between 40 and 60 years old. Women are affected more often than men, making up roughly 57 to 75 percent of frozen shoulder patients depending on the study.⁹ Having frozen shoulder in one shoulder raises the risk of developing it in the other, though the two sides don’t always occur at the same time.

How It Feels Day to Day

The numbers on frozen shoulder’s daily impact are striking. In a study of 96 patients, about 60 percent reported pain at 7 out of 10 or higher, and three-quarters showed significant functional disability. Nearly all participants, 99 percent, had high levels of kinesiophobia, which is an excessive fear of movement driven by the expectation that it will cause pain. Over half showed signs of clinical anxiety, and more than a third reported poor sleep quality.¹⁰

Sleep disruption is one of the most common complaints. The pain often worsens at night, especially when lying on the affected side, and the resulting sleep deprivation feeds into anxiety and low mood. Research has confirmed that pain, anxiety, and disturbed sleep cluster tightly together in frozen shoulder patients, and that addressing the psychological side may improve overall outcomes.¹¹ The pain itself can extend well beyond the shoulder. The most commonly affected area is the front and side of the shoulder, but the spatial spread of pain has been shown to correlate with pain intensity and with measures of central sensitization, where the nervous system amplifies pain signals.

Tasks that seem trivial become difficult or impossible: reaching for a seatbelt, putting on a coat, washing your hair, reaching a high shelf. The combination of physical limitation, chronic pain, and sleep loss can make frozen shoulder genuinely debilitating, even though it doesn’t involve a structural break or tear.

Conservative Treatment

Most people with frozen shoulder begin with non-surgical management, and most will not need to go beyond it. The main pillars are pain control, corticosteroid injections, and carefully dosed physical therapy.

Anti-inflammatory medications and over-the-counter painkillers are usually the first line for managing discomfort. They don’t change the underlying disease course, but they can make it more tolerable, especially during the painful freezing phase.

Corticosteroid injections into the shoulder joint are one of the best-supported treatments in the short term. A systematic review of randomized trials found that corticosteroid injection was superior to both placebo and physiotherapy alone for up to about 12 weeks.¹² A meta-analysis of randomized controlled trials found that steroid injections significantly reduced pain scores at four to six weeks and at three to four months compared to controls, with improvements in external rotation, abduction, and overall shoulder function at all measured time points.¹³ The benefit tends to fade by six months, though, which is consistent with the idea that injections help manage the inflammatory window rather than altering the overall disease trajectory. Where the injection is placed within the joint also matters. A network meta-analysis found that injections targeting the rotator interval, the area most involved in frozen shoulder pathology, offered the best results with moderate to high certainty.¹⁴

Physical therapy is standard but requires some nuance. One important and somewhat counterintuitive finding is that aggressive stretching can actually slow recovery. A prospective study compared intensive physical therapy (including passive stretching and manual mobilization) with “supervised neglect,” where patients did only gentle exercises within their pain limits. After two years, 89 percent of the supervised neglect group had normal or near-normal painless shoulder function, compared with just 63 percent of the intensive therapy group. The gentle group also reached that threshold faster.¹⁵ Rehabilitation models now generally recommend matching exercise intensity to the patient’s irritability level rather than pushing through pain.¹⁶

Procedures for Stubborn Cases

When months of conservative treatment haven’t produced enough improvement, two main procedural options are available: hydrodilatation and manipulation under anesthesia.

Hydrodilatation, sometimes called hydrodistension, involves injecting a large volume of fluid (usually saline mixed with steroid and local anesthetic) into the shoulder joint under image guidance. The goal is to stretch and even rupture the contracted capsule through hydraulic pressure. The anterior capsule tends to give way first, which may explain why patients often see the biggest gains in external rotation. Studies have found significant improvements in function, pain, and disability after the procedure, particularly when followed by physiotherapy.¹⁷ The technique was first described in 1965 and has remained in use largely because it is relatively low risk and can be performed as an outpatient procedure.¹⁸

Manipulation under anesthesia (MUA) is more forceful. While the patient is under general anesthesia, the surgeon manually breaks through the capsular adhesions by moving the shoulder through its full range of motion. It’s effective at restoring mobility quickly, and a review of 11 studies reported an overall complication rate of just 0.4 percent across nearly 700 patients. However, about 14 percent of patients in those studies needed a second intervention afterward.¹⁹ The procedure does carry risks that aren’t fully captured by complication rates. When researchers performed arthroscopy immediately after manipulation, they found capsular ruptures in every patient (that’s the intended effect), but also fresh partial rotator cuff tears, labral detachments, and other unintended damage in a subset of cases.²⁰ Most of these injuries didn’t seem to affect clinical outcomes in the study’s follow-up window, but they raise questions about long-term consequences.

Surgery

Arthroscopic capsular release is the main surgical option for frozen shoulder that hasn’t responded to other treatments. Using a camera and small instruments inserted through keyhole incisions, the surgeon cuts through the thickened capsule in a controlled fashion, releasing the contracture without the brute force of manipulation. It offers the advantage of direct visualization, meaning the surgeon can see exactly what they’re cutting and avoid damaging surrounding structures.²¹

Outcomes from capsular release are generally good. A study of 255 patients found significant improvements in both shoulder function scores and range of motion at five years after surgery, with no cases of dislocation, nerve injury, or recurrence.²² When comparing capsular release head-to-head with manipulation under anesthesia, a systematic review and meta-analysis found that the two procedures performed similarly in terms of pain relief, functional improvement, and range of motion over the short term. Capsular release showed a small statistical advantage in long-term pain relief beyond 12 months, though the difference didn’t cross the threshold considered clinically meaningful. The capsular release group did have a higher rate of severe complications.²³ The choice between MUA and arthroscopic release often comes down to the surgeon’s assessment of the individual case, the patient’s other shoulder pathology, and local practice patterns.

Does It Really Go Away on Its Own?

The traditional teaching is that frozen shoulder is “self-limiting,” meaning it eventually resolves without intervention. This claim is deeply embedded in medical education and patient information leaflets, but the evidence supporting full resolution is weaker than most people realize. A systematic review specifically examining the natural history of frozen shoulder found that the idea of progression to full recovery was not supported by the literature. Instead, the pattern that emerged was one of early improvement that slows over time, often leaving people with measurable limitations that persist for years.²⁴

A long-term follow-up study, tracking patients for up to 27 years, painted a more optimistic picture: about 94 percent of untreated patients eventually regained range of motion equal to their unaffected shoulder. But even in that study, the average duration of symptoms was 15 months for untreated patients and 20 months for those treated non-operatively, with some individual cases lasting up to five years.²⁵ So yes, for most people, the condition eventually runs its course. But “eventually” can mean years, and there’s a meaningful minority who don’t get back to normal. The idea that you can simply wait it out for 12 to 18 months and expect a full recovery overpromises. It’s more accurate to say that most people improve substantially, many recover fully, and a smaller group carries residual stiffness or mild pain long term.

Frozen Shoulder and the Naming Problem

One source of confusion around this condition is that it has been called different things over 150 years. The French physician Duplay described “peri-arthritis” of the shoulder in 1872. In 1934, Codman coined the more evocative “frozen shoulder” to describe insidious stiffness with restricted elevation and rotation. Later, the term “adhesive capsulitis” became the clinical standard, implying that adhesions inside the capsule were the cause. The problem is that adhesions are not consistently found at surgery, and the term may overstate one mechanism while ignoring the broader inflammatory and fibrotic picture. Modern classification distinguishes between primary (idiopathic) frozen shoulder, which arises without a known trigger, and secondary forms tied to trauma, surgery, or systemic disease. Clinicians use all of these terms somewhat interchangeably, which can make navigating your own diagnosis confusing. If a doctor tells you that you have adhesive capsulitis, they are talking about the same condition as frozen shoulder.

Emerging Approaches

Most treatment for frozen shoulder still involves some combination of steroid injections, careful exercise, and time. But a few newer approaches are being studied. Ultrasound-guided pulsed radiofrequency targets the nerves supplying the shoulder capsule with short bursts of electrical energy, aiming to reduce pain signaling without destroying tissue. A randomized controlled trial found that this approach outperformed a sham procedure at both 6 and 12 weeks, with significant improvements in pain, shoulder function, and quality of life.²⁶ The evidence base is still thin, consisting of individual trials rather than large meta-analyses, but the results are encouraging for patients who can’t tolerate steroid injections or haven’t responded to them.

Research into the molecular biology of frozen shoulder has also opened up potential drug targets. The inflammatory cytokines and growth factors found at elevated levels in frozen shoulder tissue, including TGF-beta and various interleukins, are the same molecules targeted by biologic drugs already approved for other conditions like rheumatoid arthritis. No biologic therapy has been tested specifically for frozen shoulder in a large clinical trial yet, but the identification of these molecular pathways has prompted calls for such studies.²⁷ If frozen shoulder really does share biological features with other fibrotic conditions like Dupuytren’s disease, treatments developed for one might eventually prove useful for the other. That remains speculative for now, but it gives the field a direction it didn’t have when frozen shoulder was simply a condition you waited out.