INTRODUCTION
Frozen shoulder, or adhesive capsulitis, is characterized by inflammation of the glenohumeral joint capsule and thickening of the coracohumeral ligament [1]. Clinically, it presents as a global restriction of shoulder movement, affecting both active and passive motions [1]. Without timely intervention, it may cause prolonged functional impairment and diminished quality of life [2]. The condition affects 2%–5% of the general population and up to 30% of individuals with diabetes [3]. Management begins with conservative treatments including analgesics, anti-inflammatories, physical therapy, corticosteroid injections, and hydrodilatation. If symptoms persist beyond 3–6 months, surgical options such as manipulation under anesthesia or arthroscopic capsular release may be warranted [4,5].
Shoulder hydrodilatation is a widely used outpatient, non-surgical treatment for frozen shoulder [6,7]. Although its mechanism remains unclear, it may reduce glycosaminoglycan levels in the joint capsule, inhibiting myofibroblast activity and alleviating stiffness [8]. A recent meta-analysis found that hydrodilatation with corticosteroids significantly improves pain, range of motion (ROM), and shoulder function compared to physical therapy or corticosteroid injections alone [9]. Despite its efficacy, hydrodilatation is typically reserved for patients with symptoms persisting beyond 6 months after failed physical therapy, by which time most have progressed from the painful “freezing” stage to the stiffer “frozen” phase [3,10].
Several studies have examined predictors of shoulder hydrodilatation efficacy in frozen shoulder. Positive outcomes are linked to female sex, prior analgesic use, and coracohumeral ligament thickness less than 3 mm [11], whereas diabetes and bilateral involvement predict poorer long-term results [12]. The role of disease duration remains inconclusive due to variability in study designs and outcome measures; some report no association [6,10,11], while others suggest a correlation with clinical improvement [12].
Given these uncertainties, this study investigates the association between symptom duration and shoulder function following ultrasound-guided hydrodilatation in frozen shoulder. Patients were categorized into early (<6 months) and late (≥6 months) groups. This classification corresponds to the transition from the freezing stage—typically occurring between 2 and 6 months and characterized by predominant pain and partially restricted ROM—to the frozen stage, during which pain gradually decreases while joint stiffness increases [3]. Secondary aims include identifying additional functional predictors and evaluating procedural adverse effects. Findings seek to optimize hydrodilatation timing and guide patient selection to enhance clinical outcomes.
METHODS
The study protocol was approved by the Institutional Review Board of Faculty of Medicine, Chulalongkorn University (No. 605/65) and was registered in the clinical trial registry as TCTR20230201001. Written informed consent was obtained from all participants.
This was a prospective observational analytical study investigating patients with frozen shoulder who underwent ultrasound-guided shoulder hydrodilatation at the Department of Rehabilitation Medicine, King Chulalongkorn Memorial Hospital from November 2022 to October 2024.
The manuscript was prepared in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines (Supplementary Material 1).
Participants
Participants comprised individuals aged 18 to 70 years who were clinically diagnosed with frozen shoulder by rehabilitation physicians who were uninvolved with the research and deemed suitable for ultrasound-guided hydrodilatation. The diagnosis was based on both medical history and physical examination. Key historical features included spontaneous shoulder pain, stiffness, and/or night pain, while physical examination revealed a significant reduction in shoulder ROM in at least two planes, assessed through both active and passive movements [13]. Patients were classified into early (<6 months) and late (≥6 months) intervention groups based on symptom duration. Eligible participants were required to provide informed consent to undergo the hydrodilatation procedure. Patients were excluded if they had a history of shoulder joint infection, trauma (including fractures), previous shoulder surgery, dislocation or subluxation, osteoarthritis of the shoulder, full-thickness tendon tears, rheumatologic disease, or hemiplegic shoulder resulting from neurological conditions. Those who had received shoulder hydrodilatation within the previous 3 months were also excluded.
Research Procedure
All participants with frozen shoulder were referred to our clinic by physiatrists not involved in the study. Baseline assessments were conducted prior to the hydrodilatation procedure, which was performed by one of our researchers with at least 5 years of experience in musculoskeletal ultrasound. The standard injection protocol consisted of 40 mg of triamcinolone (1 mL), 5 mL of 1% lidocaine, and 24 mL of normal saline, for a total volume of 30 mL. Under a sterile technique, the injection was administered into the posterior glenohumeral joint using a posterior approach guided by real-time ultrasound imaging.
After the procedure, participants were monitored for at least 30 minutes for any adverse reactions. Paracetamol was permitted as a rescue medication for severe pain, and the amount of administered medication was recorded. No other painkillers were allowed. Participants received guidance and informational leaflets from the researcher on appropriate practices for managing frozen shoulder, which included avoiding activities that exacerbate pain and performing recommended structured exercises to improve ROM and reduce muscle tightness around the shoulder (Supplementary Material 2). During the study, participants were instructed not to undergo additional physiotherapy beyond the prescribed home exercises or any further injections in the affected shoulder during the study period.
Outcome Assessments
The baseline demographic data of age, sex, history of diabetes mellitus, affected shoulder (dominant or non-dominant side), as well as interested clinical factors including duration of symptoms prior to receiving ultrasound-guided shoulder hydrodilatation, history of physical therapy treatment, stage of frozen shoulder (freezing, frozen, and thawing), baseline Oxford Shoulder Score (OSS), visual analog scale (VAS), and shoulder ROM were assessed. Follow-up outcomes of OSS, VAS, and ROM were repeatedly assessed 6 weeks after the procedure.
The primary outcome measure was the OSS, which consists of a self-administered questionnaire containing 12 items related to shoulder pain and function over the past 4 weeks. Each item is rated on a five-point scale of none, mild, moderate, severe, and unbearable. Total scores range from 0 to 48, with higher scores indicating more severe shoulder problems [14]. The secondary outcomes measures are the VAS for pain severity and shoulder active and passive ROM assessed in terms of forward flexion, abduction, internal rotation, and external rotation in accordance with anatomical planes. Shoulder ROM was assessed using a 360° standard goniometer in standardized positions for each movement. For forward flexion and abduction, participants were positioned either supine or sitting with the elbow extended and arm at the side. The goniometer axis was placed at the midpoint of the lateral shoulder for flexion and at the anterior or posterior shoulder for abduction, with the stationary arm set at 0° and the movement arm aligned parallel to the humerus. Degrees of internal and external rotation were measured with the participant in a supine position, shoulder abducted to 90° and elbow flexed at 90° in a pronated position. The goniometer axis was placed at the elbow, aligned with the humerus, with the stationary arm at 0° and the moving arm parallel to the forearm [15].
Adverse events occurring during or following the hydrodilatation were observed and documented. Additionally, the frequency of shoulder exercises was recorded through a self-reported shoulder exercise log. Participants who engaged in exercises for more than 80% of the scheduled follow-up days were classified as compliant. The frequency of analgesic use was also recorded through a log detailing the consumption of paracetamol over the 6-week period.
Sample Size Calculation
The sample size for this study was calculated using the formula for two independent proportions. Based on the findings of Clement et al. [16], the success rate of shoulder hydrodilatation, as evaluated by the OSS at 1 month post-procedure, was 55%. The research team anticipated that the success rate would be significantly higher, at least 85%, in patients who underwent hydrodilatation within 6 months of symptom onset. The sample size calculation assumed a significance level (α) of 0.05, a power of 0.80 (β=0.2), and an equal classification ratio of patients between the late and early treatment groups (1:1). Consequently, 72 patients were recruited for the study, with 36 patients receiving hydrodilatation within 6 months of symptom onset (early group) and the remaining 36 patients undergoing the procedure at least 6 months after symptom onset (late group).
Statistical Analysis
Data analysis was conducted using IBM SPSS version 22.0 (IBM Corp.). The distribution of data was assessed for normality using box plots and skewness values. Baseline characteristics were compared between early and late groups using an unpaired t-test for continuous data and a chi-square test for categorical data. Pre- and post-ultrasound-guided shoulder hydrodilatation evaluations were compared using a paired t-test. Effect sizes were calculated using Cohen’s d for repeated measures (denoted as dRepeated Measure; dRM) to quantify the magnitude of treatment effects in each group [17].
Associations between various factors and changes in the OSS score were analyzed, including those of symptom duration, age, sex, diabetes status, dominant-side symptoms, physical therapy history (received/not received), baseline shoulder ROM, and baseline VAS. Univariate regression analysis was performed using a linear regression model, with results presented as means and 95% confidence intervals. Factors with a P-value <0.20 in univariate regression analysis were further examined using multivariate regression, with statistical significance set at P<0.05. Additionally, the relationship between pain severity (classified as mild, moderate, or severe) [18] and change in the OSS score was explored using one-way analysis of variance.
RESULTS
Seventy-two patients were enrolled in this study and divided into two groups according to duration of symptoms: the early injection group (<6 months), consisting of 36 patients with an average symptom duration of 3.5 months before receiving ultrasound-guided shoulder hydrodilatation, and the late injection group (≥6 months), comprising 36 patients with an average symptom duration of 11.1 months. Baseline characteristics of the participants are summarized in Table 1.
When comparing changes before and after ultrasound-guided shoulder hydrodilatation between the early and late injection groups, both groups demonstrated significant improvement in the OSS, reduction in the VAS, and increase in shoulder ROM (P<0.001) (Table 2). The success rates of interventions—defined by an OSS improvement exceeding the minimal clinically important difference (4.3 points) [19]—did not differ significantly between the early and late injection groups (89% vs. 86%, P=0.722).
Univariate analysis of factors influencing changes in OSS post-injection revealed significant factors of symptom duration exceeding 6 months, female sex, history of physical therapy, pre-injection VAS scores, and pre-procedure external rotation angle of the shoulder. However, multivariate analysis indicated that only the pre-injection VAS score had a significant relationship with the reduction in OSS (P=0.006) (Table 3). When pain levels were categorized as mild, moderate, and severe[18], there was no significant difference in the reduction of OSS scores between the groups; however, a linear trend of improvement was observed (P=0.161), as shown in Fig. 1.
In terms of shoulder exercise compliance, 72% of volunteers (26 individuals) in the early injection group showed good adherence, compared to 75% of participants (27 individuals) in the late injection group (P=0.789). Paracetamol usage showed no significant difference between groups (P=0.148). In the early injection group, 81% of patients did not use analgesics, 19% used them occasionally, and none used them regularly. In the late injection group, 81% did not use analgesics, 11% used them occasionally, and 8% used them regularly. No severe adverse events were reported. Minor adverse effects, such as temporary pain and swelling lasting 1–2 days, occurred in 12 cases and were resolved with cold compresses or analgesics.
DISCUSSION
This study examines the association between the duration of symptoms prior to ultrasound-guided shoulder hydrodilatation and post-injection shoulder function, as measured by the OSS score, in patients with frozen shoulder. The results demonstrate that the symptom duration before injection does not significantly correlate with changes in shoulder function in the short-term, similar to other factors such as sex, age, diabetes status, history of physical therapy, and pre-injection shoulder ROM. Notably, the only significant independent factor influencing changes in OSS was the baseline pain score; patients with higher pain severity before the injection exhibited greater improvements in OSS scores after the procedure.
Frozen shoulder can be clinically categorized into three stages: stage I (freezing), characterized by pain with minimal restriction in ROM; stage II (frozen), marked by persistent pain and progressive limitation in ROM; and stage III (thawing), characterized by pronounced stiffness with a reduction in pain severity [13]. The duration of symptoms prior to seeking medical consultation may provide insight into the clinical stage of the condition. Pandey and Madi [3] recommend considering surgical intervention for frozen shoulder if conservative treatment is ineffective for 6–9 months, suggesting the 6-month mark as a reasonable timeframe for reevaluating treatment options. However, the findings of the present study indicate no correlation between symptom duration or timing of injection (early vs. late within 6 months) and changes in OSS after 6 weeks. This aligns with Sinha et al. [10], who reported no relationship between symptom duration and shoulder activity outcomes (OSS and Quick Disabilities of Arm, Shoulder and Hand score) at 4 weeks post-hydrodilatation. Although 85% of patients in their study had symptom durations exceeding 6 months, our study included patients with much earlier symptoms, averaging 3.5 months in the early intervention group. Nevertheless, the comparable short-term results suggest that symptom duration may not be associated with functional improvement following shoulder hydrodilatation. Our findings are consistent with Yang et al. [11], who found that symptom duration did not affect Shoulder Pain and Disability Index score changes 2 months after shoulder hydrodilatation.
Additionally, this study found a significant relationship between higher pre-injection pain scores and improved shoulder function post-injection, especially in those with moderate to severe pain. Previous research has elucidated the pathophysiology underlying the development of frozen shoulder, identifying key processes such as inflammation, increased vascularity, fibroblast proliferation, and thickening of the synovial membrane in the capsular tissues [13]. The reduction in pain can be attributed to the mechanism of action of corticosteroids, which suppress fibroblast proliferation [20] and counteract inflammation [21]. These effects contribute to pain relief, improved ROM, and enhanced shoulder function.
When categorized by pain severity, greater improvement in OSS scores was observed in the higher severity group, following a linear trend, although the between-group differences were not significant. Nonetheless, all pain severity groups demonstrated clinically significant improvements in OSS scores, exceeding the previously established threshold of 4.3 points [19]. This aligns with the findings of Lee et al. [22], who reported improvements in shoulder function following shoulder hydrodilatation in patients with pre-injection pain scores greater than 50 mm on the VAS. Similarly, Buchbinder et al. [23] observed that patients who experienced improved shoulder function had an average pre-injection pain score of 6 of 10 on a Likert scale. A recent network meta-analysis examined pharmacological interventions in early-stage frozen shoulder, defined by a pain-predominant inflammatory phase or symptom duration less than 6 months [6,9]. Intra-articular corticosteroid and platelet-rich plasma injections demonstrated superior short-term efficacy in reducing pain and improving function compared to physical therapy alone [24]. Although intra-articular corticosteroid injections alone can be effective in treating frozen shoulder, combining it with shoulder hydrodilatation has shown greater improvement, particularly in short- to medium-term outcomes such as pain relief and ROM [6,9]. In particular, patients with early-stage frozen shoulder demonstrated greater improvement in external rotation with shoulder hydrodilatation compared to steroid injection alone in medium-term assessment [6,7]. These benefits may contribute to improved shoulder function, as observed in our study. An additional advantage of this combined intervention is its ability to facilitate self-exercise by reducing pain and partially restoring ROM.
Our findings indicate that even late-stage injections can enhance short-term functional outcomes and may help postpone or reduce the need for surgical intervention in patients with frozen shoulder. Although spontaneous recovery often occurs within 2 years, up to 40% of patients may continue to experience persistent pain and residual stiffness [1]. Early management can support exercise adherence, minimize disability, and improve overall quality of life. Notably, Kim et al. [12] found that longer symptom duration significantly affected ASES scores over a 2.7-year follow-up following conservative treatments, highlighting the contrast between short- and long-term outcomes. These findings address the need for further investigation into factors influencing long-term outcomes following shoulder hydrodilatation.
Other factors, including sex, age, diabetes, and history of physical therapy, did not correlate with shoulder function after injection in this study. This aligns with Yang et al. [11], who found no relationship between sex and shoulder function post-injection, and Kim et al. [12], who reported similar findings regarding diabetes and shoulder function. However, Sinha et al. [10] noted that younger patients exhibited better shoulder function compared to older patients, which may be due to their younger population (30% younger than 50 years) compared to the 19% in this study.
Shoulder hydrodilatation is an effective non-surgical treatment for frozen shoulder, yielding favorable outcomes in terms of reduced pain and improved shoulder ROM, both in the short and medium term, compared to physical therapy or corticosteroid injections alone [9]. Hydrodilatation is a cost-effective procedure that can be performed in outpatient settings, avoiding the risks associated with surgery or radiation exposure. Although the procedure requires skilled technique, the results of our study suggest significant benefits for patients with frozen shoulder at all stages of the condition, especially for those with severe pain, who are likely to experience improvements in shoulder function post-injection.
Notably, our study specifically investigates the posterior approach technique, with specific volume and substance used during capsular distension. Prior studies have reported injection volumes ranging from 18 to 47 mL, with higher volumes intended to induce capsular rupture, a strategy associated with inferior clinical outcomes compared to capsule-preserving techniques [25,26]. The mean volume employed in capsule-preserving hydraulic distension for adhesive capsulitis is approximately 25 mL [5]. Additionally, the choice of injectate is a critical determinant of therapeutic efficacy. Triamcinolone has been shown to reduce steroid flare reactions by more than fivefold and to yield significantly superior 6-month outcomes compared to methylprednisolone acetate [27].
There were several limitations in this study. First, not all participants underwent diagnostic imaging beyond ultrasound screening to exclude rotator cuff tears. Second, the observational prospective design may be subject to unmeasured confounding factors. Third, we did not examine procedural variations, such as the needle approach, the volume or type of injected substance, and capsular endpoint, which could have influenced the outcomes [5,28]. Fourth, only short-term effects were evaluated. Longer follow-up could offer deeper insights into the natural course of the disease, which often improves spontaneously over time. Finally, the initial assumption used in the sample size calculation may have been overly optimistic regarding the difference in efficacy between early and late intervention groups. A larger sample size in future studies may help to better identify factors associated with changes in functional outcomes following shoulder hydrodilatation. Future research should involve larger cohorts and long-term follow-up to more accurately assess treatment efficacy and clarify potential differences.
CONCLUSIONS
The duration of symptoms prior to ultrasound-guided shoulder hydrodilatation did not have a significant effect on short-term post-injection shoulder function in patients with frozen shoulder. Groups with symptom duration less than or more than 6 months experienced improvements in shoulder function. Notably, patients with higher pain severity prior to injection tended to show greater functional improvement following the procedure.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement
Print
