Toward a common definition of frozen shoulder: a scoping review of randomized controlled trials

Frozen shoulder (FS) is a common condition characterized by pain and restricted range of motion (ROM); however, its definition varies widely across randomized controlled trials (RCTs). This variability in eligibility criteria may compromise comparability across studies and hinder evidence synthesis. In this scoping review, we aimed to systematically map how FS has been defined in RCTs and identify key elements requiring standardization.

We searched the CENTRAL, MEDLINE, and Embase databases from their inception to December 2024. RCTs involving adults with FS or adhesive capsulitis were included. Two independent reviewer pairs screened the studies based on eligibility criteria. We extracted data from the included studies and categorized them into 5 domains based on definitional components, including age-related criteria, pain characteristics, ROM limitations, symptom duration, and imaging requirements. Reporting patterns were summarized descriptively and visualized using frequency plots and heat maps. This scoping review was performed in accordance with the Extended Preferred Reporting Items for Systematic Reviews and Meta-Analyses Statement for Scoping Reviews.

Of 4,426 records identified, 310 RCTs were included. Use of terminology was inconsistent across studies, with 57.7% employing the term "adhesive capsulitis" and 42.3% using "frozen shoulder." Across domains, age-related criteria showed the greatest convergence; most trials targeted middle-aged adults, typically setting a lower age limit of approximately 40 years and an upper limit between 60 and 70 years. Conversely, the pain criterion was poorly aligned; nearly half of the studies omitted pain thresholds, and those that used diverse visual analog scale/numerical rating scale cutoffs infrequently specified night pain and functional disturbances. For ROM, most studies required loss in at least 2 or 3 planes (commonly flexion, abduction, and external rotation); however, quantitative cutoffs and movement combinations varied, and many trials provided ambiguous thresholds. Symptom duration was unreported in approximately 40% of RCTs; among those that reported it, a minimum duration of ≥3 months was the most common, but it was not universal. Similarly, imaging requirements were inconsistent; plain radiographs were the most frequently used modality, with magnetic resonance imaging and ultrasound used in a minority of trials, and >60% of the studies did not cite reference literature to justify their diagnostic criteria.

The definition of FS in RCTs differs substantially across diagnostic domains, which may affect the comparability of study populations and the interpretation of treatment outcomes. The development of standardized, clearly described eligibility criteria, particularly regarding terminology, disease stage, pain and ROM criteria, symptom duration, and exclusion of other shoulder pathologies, will help strengthen methodological rigor and improve the interpretability of future clinical trials.