A note on this column: This article is general medical information, not a description of a service offered at this practice. Dr. Ishiguro's surgical practice is focused on hip and knee conditions — he does not treat or perform surgery for frozen shoulder. This column is written to help readers understand a common condition and know when and how to seek the right specialist, typically a shoulder-focused orthopaedic surgeon or physiatrist.
What frozen shoulder actually is
Frozen shoulder — the medical term is adhesive capsulitis — is a condition in which the capsule surrounding the shoulder joint becomes inflamed, thickens, and gradually contracts, restricting both active and passive movement. This last point is a key diagnostic clue: unlike a rotator cuff tear, where a patient may be able to move the arm passively (with help) even if they cannot lift it themselves, frozen shoulder restricts motion no matter who or what is moving the arm.
It affects roughly 2–5% of the general population, most commonly between the ages of 40 and 60, and is somewhat more common in women. It typically affects one shoulder at a time, though a meaningful minority of patients — reported in various series at around 6–17% — go on to develop it in the other shoulder as well, usually within five years.
The three stages
Frozen shoulder is not a static problem — it evolves through a recognisable clinical course, traditionally described in three (sometimes four) overlapping stages. Understanding which stage a patient is in matters, because the right treatment approach is different at each point.
| Stage | Typical duration | What's happening |
|---|---|---|
| Freezing | ~3–9 months | Progressive pain, often worse at night, with gradually worsening stiffness as synovial inflammation develops |
| Frozen | ~4–12 months | Pain often eases somewhat, but stiffness is at its worst; dense fibrous tissue has formed within the capsule |
| Thawing | ~12–24 months (up to 42) | Gradual, spontaneous improvement in range of motion as the capsule slowly remodels |
On average, the entire process can span one to three years. It has traditionally been described as "self-limiting," and for many patients that's true — but this reputation can be misleading. Studies following patients over the longer term have found that a substantial proportion, estimated at roughly 20–60% depending on the series, still have some residual stiffness or discomfort years after onset, and up to about 40% report restricted motion at the four-year mark. In practice, "it will go away on its own" is only part of the story, which is one reason a structured approach to treatment is worth pursuing rather than simply waiting it out unmonitored.
Who gets it — and why diabetes matters
Frozen shoulder can occur without any clear trigger (primary or "idiopathic" frozen shoulder), or it can develop after a period of shoulder immobility — following a fracture, a stroke, cardiac surgery, or any prolonged period of not using the arm normally (secondary frozen shoulder). But among the risk factors identified in the literature, one stands out clearly: diabetes.
What the evidence shows: A systematic review and meta-analysis of six case-control studies covering 5,388 people found that people with diabetes had 3.69 times the odds of developing frozen shoulder compared with people without diabetes (95% CI 2.99–4.56); two additional cohort studies in the same review found diabetes associated with a 32–67% relatively increased hazard of developing the condition (Dyer et al., BMJ Open, 2023). Separate genome-wide association work has gone further, finding evidence that diabetes is not just correlated with frozen shoulder but is likely a causal risk factor, working through glycaemic rather than purely mechanical pathways.
The practical implication is straightforward: anyone with diabetes who develops new shoulder stiffness should have a lower threshold for seeking evaluation, and anyone diagnosed with frozen shoulder without an obvious cause may reasonably be screened for undiagnosed diabetes or prediabetes if they haven't been checked recently. Thyroid disease, Dupuytren's contracture, and Parkinson's disease have also been associated with higher rates of frozen shoulder, though the evidence linking these is generally less robust than for diabetes.
How the diagnosis is made
Frozen shoulder is primarily a clinical diagnosis. The hallmark finding is a global, progressive loss of both active and passive range of motion — particularly external rotation with the arm at the side, which is often the most severely and earliest affected movement. Reaching behind the back or overhead becomes difficult, and pain is often felt diffusely around the shoulder and upper arm rather than in one precise spot.
Because rotator cuff tears, calcific tendinitis, glenohumeral osteoarthritis, and referred pain from the cervical spine can all mimic aspects of frozen shoulder, imaging is commonly used — not to make the primary diagnosis, but to rule out these other causes and confirm the clinical impression, particularly when a patient's presentation is atypical (sudden onset, a clear traumatic mechanism, or asymmetric findings).
What the evidence shows: A 2023 review of the diagnostic imaging literature concluded that plain radiographs remain useful as a first step mainly to exclude other pathology (such as arthritis or calcific deposits), since frozen shoulder itself often looks unremarkable on X-ray. MRI and ultrasound are more sensitive for identifying the capsular thickening and inflammatory changes characteristic of adhesive capsulitis, and are particularly helpful when the clinical picture is ambiguous or a coexisting rotator cuff problem needs to be excluded (Picasso et al., Diagnostics, 2023).
What the evidence says about treatment
Treatment approaches for frozen shoulder range from doing very little to surgery, and the right choice depends heavily on the stage of the condition, the severity of symptoms, and how much the stiffness is affecting daily function.
Commonly used treatment options
- Watchful waiting — structured observation with corticosteroid injections used as needed for pain, without a formal exercise program
- Physical therapy — supervised stretching and mobilisation exercises, typically most helpful in the freezing and early frozen stages for pain control, and in the thawing stage for restoring motion
- Corticosteroid injection — commonly used, particularly in the earlier, more painful stage, often combined with either of the above
- Hydrodilatation (distension) — an image-guided injection that stretches the tightened capsule with fluid, sometimes combined with steroid
- Manipulation under anaesthesia or arthroscopic capsular release — generally reserved for patients whose symptoms haven't improved after 6–12 months of a genuine trial of non-surgical treatment
A notable and somewhat counterintuitive recent finding concerns the comparison between physical therapy and simple watchful waiting.
What the evidence shows: A prospective randomised controlled trial of 61 patients with idiopathic frozen shoulder (31 randomised to watchful waiting, 30 to physical therapy, both groups allowed corticosteroid injections as needed) found that both groups improved significantly on shoulder-specific outcome scores at 6 weeks, 3, 6, and 12 months — but there was no statistically significant difference between the two groups at any time point. Physical therapy, however, cost roughly ten times more than watchful waiting over the study period. The trial was presented as the AAOS 2024 Annual Meeting's best paper in the shoulder and elbow category (Siddiq, Martin, et al., AAOS Annual Meeting, 2024). This is a single trial and doesn't mean physical therapy is never useful — many clinicians still find supervised exercise helps patients manage pain and stay engaged with their recovery — but it does suggest that structured observation is a reasonable, evidence-supported option rather than a lesser one.
Injectable treatments beyond corticosteroids have also been studied. Platelet-rich plasma (PRP), prepared from a patient's own blood, has drawn interest as an alternative or adjunct to steroid injection.
What the evidence shows: A 2024 systematic review and meta-analysis pooling 14 randomised controlled trials (1,024 patients) found that PRP injection produced meaningful improvements in pain and shoulder function in frozen shoulder, with a safety profile comparable to other injectable treatments (Zhang et al., BMC Musculoskeletal Disorders, 2024). Head-to-head comparisons with corticosteroid injection in the literature are mixed, and PRP protocols vary considerably between studies (preparation method, number of injections, concentration), so it's best understood as a reasonable option under specialist guidance rather than a clearly superior one at this stage.
For the minority of patients who don't improve with a genuine trial of non-surgical care, manipulation under anaesthesia and arthroscopic capsular release are both established options, with published series generally reporting good improvement in pain and motion. The choice between them, and the decision of when to move to surgery at all, is one best made by a shoulder specialist after a proper trial of conservative treatment — usually a minimum of several months — has been given a fair chance to work.
When to see a specialist
Because frozen shoulder shares symptoms with several other shoulder conditions — most importantly rotator cuff tears, which are managed very differently — new shoulder pain and stiffness that doesn't improve within a few weeks is worth a proper clinical evaluation rather than self-diagnosis. This is especially true for anyone with diabetes, a history of prolonged shoulder immobilisation, or a prior frozen shoulder in the other arm, given the recognised risk of bilateral involvement.
References (PubMed / Journal)
Dyer BP, Rathod-Mistry T, Burton C, van der Windt D, Bucknall M. Diabetes as a risk factor for the onset of frozen shoulder: a systematic review and meta-analysis. BMJ Open. 2023;13(1):e062377. DOI: 10.1136/bmjopen-2022-062377
Picasso R, Pistoia F, Zaottini F, Marcenaro G, Miguel-Pérez M, Tagliafico AS, Martinoli C. Adhesive Capsulitis of the Shoulder: Current Concepts on the Diagnostic Work-Up and Evidence-Based Protocol for Radiological Evaluation. Diagnostics (Basel). 2023;13(22):3410. DOI: 10.3390/diagnostics13223410
Zhang WB, Ma YL, Lu FL, Guo HR, Song H, Hu YM. The clinical efficacy and safety of platelet-rich plasma on frozen shoulder: a systematic review and meta-analysis of randomized controlled trials. BMC Musculoskelet Disord. 2024;25:706. DOI: 10.1186/s12891-024-07629-1
Siddiq BS, Martin SD, Cherian NJ, et al. Watchful Waiting Achieves Higher Value Than Physical Therapy for the Treatment of Idiopathic Frozen Shoulder: A Prospective Randomized Controlled Trial. Presented at the AAOS 2024 Annual Meeting (Paper 401); reported in AAOS Now, Feb 12, 2024. aaos.org/aaosnow — Randomized Study Compares Nonoperative Treatments for Frozen Shoulder