Frozen shoulder usually improves, but the familiar promise of complete spontaneous recovery within a fixed period is too simple. Duration and outcome vary, and a meaningful minority of people report residual pain or restriction at long-term follow-up.
Recovery often takes many months and sometimes several years. Long-term studies disagree on the proportion achieving completely normal movement, partly because populations and outcome definitions differ. Diabetes is a well-supported risk factor for developing frozen shoulder and may be associated with a more difficult course.
Questioning the textbook timeline
The classic freezing, frozen and thawing stages describe a broad pattern, not a clock. Stages overlap and patients may not identify a clear transition. An early prospective study reported an average duration of about 30 months, while other cohorts found shorter or much longer courses.
Some long-term studies report near-normal function and movement for most participants. Others identify persistent, usually mild, pain or loss of motion in a substantial group. These findings are not necessarily contradictory: “recovered” can mean pain-free, functionally satisfied, or objectively symmetrical range.
Risk factors and associations
Diabetes has the strongest and most consistent association. A 2023 systematic review of longitudinal studies estimated substantially higher odds of developing frozen shoulder in people with diabetes. Thyroid disease and other metabolic or cardiovascular factors are also reported, but association does not prove direct causation and effect estimates vary.
Secondary frozen shoulder can follow trauma, surgery or prolonged immobilisation. It should be distinguished from primary frozen shoulder because the context, precautions and prognosis may differ.
What may influence outcome?
- Severity and functional loss at presentation
- Diabetes and metabolic health
- Duration before assessment and whether the diagnosis is correct
- Sleep, distress, work demands and ability to modify aggravating activity
- Access to appropriate information, symptom control and rehabilitation
These are not deterministic. They help shape monitoring and communication; they should not be used to tell an individual that recovery is impossible.
Communicating prognosis honestly
A useful explanation combines reassurance with uncertainty: most people improve substantially, change is often slow and uneven, and exact timing cannot be predicted. Short flares do not necessarily mean structural deterioration. Conversely, an atypical or worsening course should not be dismissed as merely another stage.
Outcome measurement
Research and practice should look beyond degrees of motion. Pain, sleep, dressing, work, confidence, meaningful activity and patient-rated recovery are all relevant. Follow-up intervals need to be long enough to distinguish a treatment’s short-term acceleration of recovery from the underlying clinical course.