CLINICAL EVIDENCE REVIEW · DIAGNOSIS
Frozen shoulder remains primarily a clinical diagnosis. The difficult part is not recognising a very typical presentation; it is deciding how much uncertainty remains, identifying conditions that can imitate or coexist with capsular stiffness, and using imaging only when it will change the next decision.
A progressive history of pain and functional loss together with restriction of both active and passive glenohumeral movement - often most evident in external rotation - creates the core diagnostic pattern. No single symptom, movement measure or scan should be treated as a stand-alone confirmation. Plain radiographs are used principally to exclude important structural alternatives. Ultrasound and MRI are supplementary when the history, examination or progress raises a meaningful differential question.
Why diagnosis is more than recognising stiffness
“Frozen shoulder” is often applied loosely to any painful shoulder that has become difficult to move. That shortcut matters because pain-limited movement, a substantial rotator cuff tear, glenohumeral osteoarthritis, calcific tendinopathy, post-traumatic stiffness and neurological disease can all produce a shoulder that appears “frozen” at first contact. Conversely, a person with true capsular restriction may also have age-related cuff changes or cervical symptoms. The task is therefore probabilistic: assemble a coherent clinical pattern, test plausible alternatives and keep the working diagnosis open to review.
The distinction between primary and secondary frozen shoulder also matters. Primary frozen shoulder develops without a clear local precipitating event. Secondary stiffness follows or accompanies another context, such as trauma, surgery, prolonged immobilisation or another shoulder disorder. Research studies do not always define these groups consistently, which is one reason prevalence, prognosis and treatment-effect estimates vary.
A diagnostic hierarchy for practice
The most defensible approach is hierarchical. Start with features that establish the overall clinical pattern, then use targeted tests and imaging to resolve remaining uncertainty. Reversing that order risks allowing incidental imaging findings to drive the diagnosis.
History: establishing probability and urgency
A typical history often describes gradually increasing shoulder pain followed by progressively restricted reach. Night pain and difficulty lying on the affected side are common in more irritable presentations. Dressing, fastening a bra, reaching a back pocket, washing hair, reaching a shelf and putting on a coat may become disproportionately difficult. These features are useful because they describe both elevation and rotational function, but none is diagnostic in isolation.
Clarify whether onset was spontaneous or followed injury, surgery, immobilisation, vaccination or another shoulder episode. Ask about diabetes and thyroid disease because both are associated with frozen shoulder; this changes prior probability, not the diagnostic criteria. Previous contralateral frozen shoulder also increases clinical suspicion. Medication use, anticoagulation, occupation, sleep disruption and the effect on self-care help determine management even when they do not determine diagnosis.
The history must also look beyond the shoulder. Neck pain, paraesthesia, numbness, pain extending below the elbow, progressive weakness, systemic illness, fever, unintentional weight loss, chest symptoms, a cancer history or an inflammatory pattern can redirect assessment. Severe unremitting pain is not automatically sinister, but pain that is non-mechanical, rapidly worsening or accompanied by systemic or neurological change deserves a lower threshold for medical review.
A patient can have frozen shoulder and another clinically important disorder. New neurological deficit, major trauma, deformity, acute systemic illness, a hot swollen joint, suspected infection, malignancy or visceral referral should be managed through appropriate urgent pathways.
Examination: active and passive restriction are not interchangeable
Active movement reflects pain, confidence, strength, motor control and joint mobility. Passive movement reduces some of those demands and therefore helps distinguish a painful or weak shoulder from a mechanically restricted one. Frozen shoulder becomes more likely when both active and passive glenohumeral movement are restricted, particularly when passive external rotation with the arm by the side is clearly limited relative to the other shoulder.
The phrase “capsular pattern” is clinically convenient but should not be interpreted as a rigid ratio that every patient must reproduce. The magnitude and direction of restriction vary with symptom irritability, stage, measurement position, scapular stabilisation and the force used by the examiner. A highly irritable shoulder may stop because of pain and guarding before a firm capsular end-feel is reached. A later, less painful presentation may allow stiffness to be characterised more confidently.
Useful examination components
- Observation: resting posture, protective behaviour, swelling, deformity and muscle wasting.
- Cervical and neurological screen: selected when the history or symptom distribution suggests a non-shoulder contribution.
- Active movement: range, pain behaviour, scapular substitution and confidence.
- Passive movement: external rotation at side, elevation and other directions measured with a consistent position and tolerable force.
- Strength: interpreted cautiously because pain inhibition and restricted test position can mimic weakness.
- Functional tasks: patient-selected activities recorded in a repeatable way.
Special tests developed for rotator cuff or subacromial pain become difficult to interpret in a globally painful, restricted shoulder. A “positive” response may reproduce pain without identifying its source. Testing should therefore answer a defined differential question rather than accumulate labels.
Measurement: precision should serve the decision
Range-of-motion values are only comparable when position and method are comparable. Record whether movement was active or passive, the position of the arm and patient, how the scapula was controlled, the instrument used and why the movement stopped. An apparently precise angle is misleading if the method changes between visits.
| Element | Why it matters | Practical wording |
|---|---|---|
| Movement and position | External rotation at the side is not equivalent to rotation in abduction. | Passive external rotation, arm at side, supine |
| Scapular control | Scapulothoracic substitution can conceal glenohumeral restriction. | Scapula manually stabilised / not stabilised |
| Limiting factor | Pain, guarding and mechanical resistance have different implications. | Stopped by pain before firm resistance |
| Symptom response | A flare after testing may be more important than a few additional degrees. | Returned to baseline within two hours |
| Functional anchor | Range does not fully represent participation or confidence. | Can reach hairline; cannot reach top shelf |
Imaging: exclusion and clarification, not automatic confirmation
The 2025 clinical practice guideline for primary frozen shoulder concluded that several ultrasound and MRI findings are associated with the condition, but their diagnostic performance does not surpass clinical diagnosis. Imaging findings should therefore be interpreted as supplementary. This distinction is important: a scan can demonstrate capsular or rotator-interval changes, but it cannot decide whether those findings explain the person’s presentation.
| Modality | Most useful clinical question | What it can contribute | Important limitation |
|---|---|---|---|
| Plain radiographs | Is there arthritis, fracture, dislocation or another bony explanation for stiffness? | Excludes major structural alternatives and may identify calcification. | A normal radiograph does not by itself confirm frozen shoulder. |
| Ultrasound | Is a substantial cuff tear, calcific tendinopathy or another soft-tissue disorder plausible? | Dynamic, accessible assessment of cuff and bursal structures; some features can support frozen shoulder. | Operator dependent; incidental abnormalities are common and capsular assessment is incomplete. |
| MRI | Does persistent diagnostic uncertainty justify broader soft-tissue and joint assessment? | May show rotator-interval or capsular thickening, signal change and other pathology. | Associated findings are not sufficiently specific to replace clinical reasoning. |
| Laboratory tests | Is there a systemic, inflammatory, endocrine or infectious question? | Targeted investigation based on history and examination. | Not routine confirmation tests for primary frozen shoulder. |
The decision to image should state the unresolved question and the management consequence. “MRI because symptoms persist” is weaker reasoning than “MRI because marked weakness and active-passive disparity raise concern about a clinically significant cuff tear that would alter referral and rehabilitation.”
Differential diagnosis: pattern recognition with explicit alternatives
| Possible diagnosis or context | Features that may point away from uncomplicated primary frozen shoulder | Potential next step |
|---|---|---|
| Glenohumeral osteoarthritis | Older age, crepitus, previous injury and radiographic joint-space or bony change. | Plain radiographs and management appropriate to arthritis. |
| Large rotator cuff tear | Traumatic onset, marked weakness, active range much worse than passive range. | Targeted imaging or specialist assessment where it would alter care. |
| Calcific tendinopathy | Acute severe pain may dominate; restriction can be pain-mediated rather than persistently capsular. | Radiograph or ultrasound when clinically indicated. |
| Cervical radiculopathy or peripheral nerve disorder | Neck-related symptoms, paraesthesia, neurological deficit or pain distribution beyond the shoulder. | Neurological assessment and appropriate referral/investigation. |
| Inflammatory arthropathy or infection | Systemic features, multiple joints, prolonged morning stiffness, warmth, swelling or fever. | Medical assessment; urgent escalation when infection is possible. |
| Post-traumatic or post-operative stiffness | Clear precipitating event, structural injury, altered precautions or surgical context. | Use the relevant trauma or post-operative pathway, not a primary-frozen-shoulder protocol. |
| Malignancy or referred visceral pain | Non-mechanical pain, systemic change, cancer history, chest or abdominal symptoms, atypical progression. | Appropriate urgent medical pathway. |
Irritability and stage describe management context, not diagnostic certainty
Stage labels - freezing, frozen and thawing - can help communicate broad symptom patterns, but real presentations overlap. A person may have severe night pain and substantial stiffness at the same time; another may improve in pain while function remains restricted. Stage should not be inferred from duration alone.
Irritability is often more actionable. Consider resting and night pain, pain before end range, post-activity response, sleep disturbance and functional tolerance. A high-irritability presentation supports gentler examination and a lower initial loading dose. Lower irritability may permit more complete movement assessment and progressive rehabilitation. This is a dosage decision, not proof that the diagnosis has changed.
Reassessment is part of diagnosis
A working diagnosis should generate expectations without becoming a promise. Record the expected direction of change and the circumstances that should trigger review. Failure to improve does not automatically mean the initial diagnosis was wrong, because frozen shoulder can be prolonged and variable. It does mean the clinician should reconsider adherence, dosage, comorbidity, psychosocial and occupational barriers, competing pathology and whether further investigation is now justified.
Implications for professional practice
- Document a working diagnosis, not simply a label. State the findings that support frozen shoulder, relevant alternatives considered and why further investigation is or is not indicated.
- Use active-passive comparison thoughtfully. Restriction in both is central, but pain, guarding and measurement method affect interpretation.
- Make imaging answer a question. Incidental cuff or degenerative findings should not displace a coherent clinical presentation without good reason.
- Track outcomes beyond range. Sleep, pain, function, confidence and meaningful activity may change on different timelines.
- Build review points into the plan. Diagnostic reasoning continues as the trajectory becomes clearer.
Selected evidence and guidance
- Clinical Practice Guidelines for Diagnosis and Non-Surgical Treatment of Primary Frozen Shoulder (2025).
- British Elbow & Shoulder Society patient care pathways and guidelines.
- Shoulder Pain Diagnosis, Treatment and Referral Guidelines for Primary, Community and Intermediate Care.
- The diagnosis and management of the stiff shoulder.
- Diabetes as a risk factor for the onset of frozen shoulder: systematic review and meta-analysis.
Evidence checked: August 2026. Review earlier if a major guideline or high-quality study changes diagnostic recommendations.