Frozen Shoulder: What It Is, How Long It Lasts, and What Helps

The glenohumeral joint has the largest range of motion of any joint in the body. It achieves this by trading stability for mobility: the glenoid socket is shallow (it covers only about a third of the humeral head), and the joint capsule, a fibrous envelope that surrounds the joint, is normally loose enough to allow the humeral head to move through approximately 180 degrees of flexion, 180 degrees of abduction, and 90 degrees of external rotation. The capsule has to be loose because the joint needs to move in every direction. That looseness is a design feature, not a weakness.

Frozen shoulder occurs when this capsule becomes inflamed and then progressively fibrotic, thickening and contracting until the shoulder’s range of motion is severely restricted. The clinical name is adhesive capsulitis. It is one of the few shoulder conditions that has a predictable trajectory: it gets worse, it plateaus, and then it gets better. The problem is the timeline.

What Is Happening

Adhesive capsulitis progresses through three recognized stages [5]:

Stage 1: Freezing (2 to 9 months). Pain is the dominant feature. It is often severe, particularly at night, and can be intense enough to prevent sleep on the affected side. Range of motion begins to decrease but pain, not stiffness, is the primary limitation. The capsule is inflamed. Histological studies show that the capsular tissue in this phase has increased vascularity, inflammatory cell infiltration, and early fibroblast activity. This is the stage where aggressive physical therapy can backfire: pushing into pain increases inflammation and may accelerate the capsular fibrosis.

Stage 2: Frozen (4 to 12 months). Pain begins to subside, but stiffness becomes the dominant problem. External rotation and abduction are typically the most restricted movements. The capsule is now fibrotic and contracted, with dense collagen bands replacing the normally loose capsular tissue. The joint volume, normally about 20 to 30 milliliters, may be reduced to as little as 5 to 10 milliliters. This explains the characteristic pattern of restriction: external rotation is lost first because the anterior-inferior capsule tightens first, and that is the capsular region that must stretch for the arm to rotate outward.

Stage 3: Thawing (5 to 24 months). Range of motion gradually returns. This phase is the longest and the most variable. Most patients recover most of their motion, but studies have found that up to 40 percent retain some mild residual restriction, typically in external rotation. Whether this residual stiffness is clinically significant varies: many patients with a few degrees of lost external rotation do not notice it in daily activity.

The risk factors are well documented. Diabetes is the strongest: a meta-analysis found that the prevalence of adhesive capsulitis in diabetic patients is 10 to 36 percent, compared to 2 to 5 percent in the general population [2]. The association is partly explained by the glycation of collagen in diabetic patients, which increases capsular stiffness independently of the inflammatory process. Thyroid disease, prior shoulder surgery or immobilization, and female sex are also risk factors. Immobilization is a particularly modifiable risk factor: after shoulder injury or surgery, early gentle range-of-motion exercise is specifically intended to prevent the cascade of inflammation and fibrosis that leads to adhesive capsulitis.

What the Evidence Supports

The UK FROST trial, the largest randomized trial of frozen shoulder treatment, compared early structured physiotherapy, manipulation under anesthesia with steroid injection, and arthroscopic capsular release in 503 patients. At 12 months, all three groups had improved substantially, with no clinically important difference between them [1]. This is important: it suggests that the condition’s natural course drives most of the recovery, and interventions modulate the speed and comfort of that course rather than fundamentally changing the outcome.

Corticosteroid injection into the glenohumeral joint has the best evidence for short-term pain relief, particularly during the freezing phase. A systematic review found moderate evidence for pain reduction and improved function in the short term (6 to 12 weeks), with benefits diminishing over time [4]. The injection reduces the capsular inflammation that drives the pain in stage 1, and the window of reduced pain often allows the patient to participate more actively in gentle range-of-motion exercises. A single injection is typical; repeated injections carry diminishing benefit and potential harm to local tissues.

Gentle range-of-motion exercises are recommended throughout all three stages, staying within pain tolerance [5]. The goal is to maintain available motion and prevent further stiffness, not to force range that the capsule will not currently allow. The distinction between “maintaining” and “forcing” is the central practical question in frozen shoulder management: the capsule will not stretch until the fibrotic process allows it, and pushing beyond its current tolerance in the freezing phase fuels the inflammation that is causing the fibrosis.

Aggressive stretching during the freezing phase is not supported by evidence and may worsen symptoms. This is one of the few conditions where “no pain, no gain” is actively counterproductive. A patient in the freezing phase who sees a physiotherapist who pushes aggressively into external rotation may leave the session with significantly increased pain and no improvement in motion. Clinicians experienced with frozen shoulder typically prescribe exercises at or just below the pain threshold.

Hydrodilatation (injecting fluid, usually saline with corticosteroid, into the joint capsule under pressure to stretch it) has some trial support, primarily for the frozen phase when the capsule is contracted. A Cochrane review found limited evidence for short-term benefit [3]. It is not widely used as a first-line treatment.

Manipulation under anesthesia and arthroscopic capsular release are options for patients who have not improved after 6 to 12 months of conservative management. Manipulation involves moving the shoulder through its full range while the patient is under general anesthesia, breaking the adhesions mechanically. Arthroscopic capsular release involves cutting the contracted capsule with surgical instruments under direct visualization. The FROST trial found they produce faster short-term improvement but similar long-term outcomes to physiotherapy alone [1]. Both carry risks: manipulation can cause humeral fracture (rare but documented, particularly in osteoporotic patients), and capsular release is an invasive procedure with its own recovery period.

Exercises to Try

These should be performed gently, within pain tolerance, not pushed to the point of sharp pain. The goal changes by stage: in the freezing phase, maintain what you have; in the frozen phase, begin gently exploring the boundary; in the thawing phase, progressively recover range.

Pendulum exercise. Lean forward, supporting your weight with the uninvolved arm on a table. Let the affected arm hang. Swing it gently in small circles, forward and backward, and side to side. Let gravity and momentum do the work; the shoulder muscles should be relaxed. Two minutes, two to three times daily. This provides gentle joint mobilization without loading the capsule or provoking the guarding response that limits motion.

Supine passive flexion. Lie on your back. Hold a stick (a broom handle works) with both hands, shoulder-width apart. Use the unaffected arm to push the affected arm into flexion (overhead), keeping the affected arm relaxed. Go only as far as tolerable. Hold for five seconds. 10 repetitions. Lying supine stabilizes the scapula and isolates glenohumeral motion, which makes this a more accurate measure of true shoulder range than standing overhead reach, where scapular compensation can mask the restriction.

Wall walk. Stand facing a wall, arm’s length away. Walk the fingers of the affected arm up the wall, using the fingers to climb rather than the shoulder muscles. Go as high as tolerable. Hold for five seconds at the top. Slowly walk back down. 10 repetitions, twice daily. The wall provides a graded resistance that allows the patient to control the range precisely. Do not lunge into the wall to gain height; the arm should be doing the work at its own pace.

Cross-body stretch. Use the uninvolved arm to pull the affected arm across the body at chest height. Hold for 15 to 20 seconds. This addresses posterior capsule tightness, which contributes to loss of internal rotation and cross-body reach. Perform gently; do not force. If this reproduces anterior shoulder pain, reduce the range or defer it until the freezing phase has passed.

External rotation with a stick. Stand holding a stick with both hands at waist height, elbows bent to 90 degrees and tucked against the sides. Use the unaffected arm to push the stick sideways, rotating the affected arm outward. Go to the edge of tolerance and hold for 5 seconds. 10 repetitions. External rotation is typically the most restricted movement in frozen shoulder, and this exercise addresses it with the leverage and control of the unaffected arm.

When to See a Doctor

See a doctor if:

• Pain is severe enough to disrupt sleep significantly and is not responding to over-the-counter analgesia
• Range of motion is rapidly declining (progression over days rather than weeks may indicate a different diagnosis)
• Stiffness has not begun to improve after 12 months
• There is significant weakness (which may indicate a concurrent rotator cuff tear rather than, or in addition to, adhesive capsulitis)
• Symptoms followed a fall or injury (need to rule out fracture or dislocation)

When to see a doctor. If joint pain follows a significant injury, is accompanied by swelling, giving way, locking, or loss of range of motion, or if you have neurological symptoms (numbness, tingling, weakness), see a doctor or physical therapist. They can assess whether imaging or specialist input is appropriate.