If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The aim of this prospective randomized study was to compare the efficacy of 3 injection methods, intra-articular injection, subacromial injection, and hydrodilatation (HD), in the treatment of primary frozen shoulder.
Patients with primary frozen shoulder were randomized to undergo intra-articular injection (n = 29), subacromial injection (n = 29), or HD (n = 28). Evaluations using a visual analog scale for pain, Simple Shoulder Test, Constant score, and passive range of shoulder motion were completed before treatment and 1 month, 3 months, and 6 months after treatment.
Among the 3 injection methods for primary frozen shoulder, HD resulted in a greater range of motion in forward flexion and external rotation, a lower visual analog scale score for pain after 1 month, and better outcomes for all functional scores after 1 month and 3 months of follow-up. However, there were no significant differences in any clinical outcomes among the 3 groups in the final follow-up at 6 months.
Although HD yielded more rapid improvement, the 3 injection methods for primary frozen shoulder resulted in similar clinical improvement in the final follow-up at 6 months.
Frozen shoulder (adhesive capsulitis) is a common disease that restricts passive and active range of motion (ROM) in the glenohumeral (GH) joint. The concept was initially developed by Codman and Neviaser.
SAI is relatively easy to perform, and it does not require radiologic guidance. HD, or arthrographic distention of the shoulder joint, induces capsular rupture by introducing a fluid into the GH joint, resulting in increased shoulder joint motion. It was introduced as an injection treatment modality for frozen shoulder by Andren and Lundberg.
However, the evidence is insufficient to conclude which injection method is superior among IAI, SAI, and HD for the treatment of frozen shoulder.
Thus, we designed a prospective, randomized study to compare treatment outcomes using IAI, SAI, and HD in patients with primary frozen shoulder. The aim of this study was to identify which treatment modality is superior in terms of the visual analog scale (VAS) score for pain as well as functional outcomes, including ROM. We hypothesized that HD would provide superior clinical improvement compared with IAI and SAI.
Sample size calculation and patient allocation
This was a randomized, prospective, controlled study. We conducted this study in accordance with the principles of the Declaration of Helsinki. The reporting of data from this trial complies with the Consolidated Standards of Reporting Trials (CONSORT) statement.
Sample sizes were calculated to detect a 20% difference among the groups in the VAS score for pain on the basis of the pilot study and previous literature.
A sample size of 30 patients in each group was required for a power of 90% at a type I error level of .05, with an expected dropout rate of 20%.
A total of 164 consecutive patients with primary frozen shoulder were prospectively enrolled between June 2012 and September 2013. Patients were diagnosed with frozen shoulder if they had limitations of both active and passive shoulder motion and more severe pain at night than during the day and if findings on radiography of their shoulders were normal.
), and if they remained unresponsive to conservative treatment consisting of medication or physical therapy for at least 6 months. All patients had limited active and passive ROM in at least 2 directions (abduction and forward flexion <100°, external rotation <20°, or internal rotation <L3).
All patients underwent simple radiography and sonography. Patients with secondary causes of frozen shoulder such as rotator cuff tear (n = 32) or calcific tendinitis (n = 9), those with GH arthritis (n = 4), those with a history of surgery on the same shoulder (n = 1), those who received a steroid injection within 6 months before enrollment (n = 23), and those who refused to participate in the study (n = 5) were excluded. No patient had a history of previous shoulder trauma, manipulation under anesthesia, or suprascapular nerve injection, and none had a worker's compensation status. The remaining 90 patients were randomly allocated into the IAI, SAI, or HD group (30 patients in each group). Patients were randomized using a computer-generated block randomization sequence (www.randomizer.org) by an independent researcher, and the group assignment was disclosed to the physician at the time of intended treatment.
Among these 90 patients, 4 (1 from the IAI group, 1 from the SAI group, and 2 from the HD group) were lost to follow-up. Accordingly, 86 patients (26 men, 60 women; mean age, 54.5 years [standard deviation, 8.3]) with primary frozen shoulder (idiopathic adhesive capsulitis) were ultimately enrolled in this study (Fig. 1). The demographic and clinical data did not differ among the groups (all P < .05), and these data are summarized in Table I. During the study period, all patients underwent conventional conservative treatment, including medication and a home-based physical therapy exercise program. The medication included a nonsteroidal anti-inflammatory drug and muscle relaxant, which were administered for approximately 4 weeks. For physical therapy, active assisted ROM exercise, including stick exercise, was performed for approximately 10 weeks, depending on the recovery of ROM. We employed a 4-quadrant stretching program (passive flexion, horizontal adduction, internal rotation behind the back with the unaffected arm, and external rotation at the side using a stick) to stretch the entire capsule at least 3 times a day (10-15 minutes per session). When ROM had recovered, muscle-strengthening exercise was performed on the scapular stabilizers (such as the lower trapezius and serratus anterior muscles) and rotator cuff using a resistance band at least 3 times a day (10-15 minutes per session).
Table IPatients' demographic variables by injection group
All data were prospectively collected by a clinical researcher (A.-S.C.) who was blinded to the study design. The patients' demographic data and other characteristics, including age, sex, symptom duration, dominant shoulder, underlying disease (diabetes mellitus, hypertension, heart disease, and thyroid disease), and smoking habits, were recorded. The level of sports activity was defined as high (dynamic or contact sports, such as boxing, basketball, rugby, and tennis), medium (static sports, such as yoga and jogging), or low (mild or no sports activity).
For IAI, an anterior approach was used with a 10-mL syringe and 21-gauge needle, and the procedure was performed with the patient in the supine position. Povidone sterilization was performed around the injection site, and the skin was anesthetized with 2% lidocaine before injection, followed by IAI under sonographic guidance. The needle was placed immediately medial to the head of the humerus and approximately 1 cm lateral to the coracoid process, and it was directed posteriorly as well as slightly superiorly and laterally under sonographic guidance. If the needle struck against bone, it was retracted and redirected at a slightly different angle. When resistance was felt on penetrating the joint capsule, aspiration was performed to ensure that the needle was not placed in a blood vessel. Then, a mixture of 1 mL of triamcinolone (40 mg), 4 mL of 2% lidocaine, and 5 mL of normal saline was injected into the GH joint space slowly and with consistent pressure under sonographic guidance. We chose 40 mg of triamcinolone for evaluating the treatment effect on the basis of previous literature.
For SAI, a posterior approach was used with a 10-mL syringe and 21-gauge needle, as described for the IAI group, and the procedure was performed with the patient in the sitting position with the arm resting comfortably at the side and with the clinician standing behind the patient. After povidone application and skin anesthesia as described for the IAI group, the needle was inserted 2 to 3 cm inferior to the posterolateral corner of the acromion and directed anteriorly in the direction of the coracoid process under sonographic guidance. After verification by aspiration that the needle was not placed in a blood vessel, a mixture of 1 mL of triamcinolone (40 mg), 4 mL of 2% lidocaine, and 5 mL of normal saline, as described for the IAI group, was injected slowly and with consistent pressure under sonographic guidance.
For HD, an anterior approach was used with a 50-mL syringe and 21-gauge needle, and the procedure was performed with the patient in the supine position under overhead fluoroscopy in the operating room. Under fluoroscopy, the GH joint was identified and marked on the skin with a pen. After povidone sterilization, draping, and skin anesthesia, the needle was placed immediately medial to the head of the humerus and approximately 1 cm lateral to the coracoid process, and it was directed posteriorly, slightly superiorly, and laterally, as described for the IAI group. The needle position was verified by fluoroscopy. After aspiration and the injection of 4 mL of contrast medium for joint space confirmation, a mixture of 1 mL of triamcinolone (40 mg), 4 mL of local anesthetic (2% lidocaine), and 40 mL of normal saline was injected into the GH joint slowly and with pressure. When resistance was encountered, the injection was stopped momentarily and then continued. During the injection, the joint was gradually distended, making the axillary and subscapular recesses more visible. The injection was continued until rupture of the capsule occurred. The capsule primarily ruptured in the wall of the subscapular recess or sometimes in the wall of the bicipital or axillary recess.
The capsular rupture appeared as a loss of resistance, with leakage of contrast material on fluoroscopy.
All patients in the 3 groups received the same nonsteroidal anti-inflammatory drug for pain relief after injection. A stretching exercise program, as described previously, was started from the first day after injection.
Clinical symptoms were evaluated at 4 time points for all patients: before treatment and 1 month, 3 months, and 6 months after treatment. We set 6 months as the final follow-up evaluation period because this period should be adequate to assess the result of the injection treatments and to decide on a subsequent treatment plan for a frozen shoulder. The clinical outcome was evaluated using the VAS score for pain (range, 0-10, with 10 indicating the worst pain), Simple Shoulder Test (SST), Constant score, and passive shoulder ROM.
Passive shoulder ROM was measured using a goniometer by a clinical researcher (A.-S.C.) who was blinded to the study. Forward elevation was measured in degrees between the arm and thorax in the scapular plane. External rotation at the side was measured in degrees between the thorax and forearm, with the arm held in an adducted position at 90° of elbow flexion. Internal rotation at the back was measured by the vertebral level reached with the tip of the thumb and numbered serially 1 to 12 for the first to twelfth thoracic vertebrae, 13 to 17 for the first to fifth lumbar vertebrae, and 18 for any level below the sacral region.
Statistical analysis was performed using SPSS 22.0 software (IBM, Armonk, NY, USA). Repeated-measures 1-way analysis of variance and post hoc Scheffé multiple comparison tests were used to identify significant differences in continuous variables among the groups; the χ 2 or Fisher exact test was used to identify any significant differences in categorical variables. Before the analysis of variance test, the Kolmogorov-Smirnov test for normality and Levene test for homogeneity of variances were performed, and the assumption of normality and homogeneity of variances was fulfilled, all with P values > .05. The primary end point was the VAS score for pain, and the secondary end points were clinical scores and shoulder ROM. Values of P < .05 were considered statistically significant.
At the initial presentation, no intergroup differences were observed in the VAS score, SST, Constant score, and any ROM measure among the 3 groups. Each group displayed significant improvements from baseline to 6 months of follow-up (all P < .001). Every patient in each group was satisfied with his or her results at 6 months, and no patients required an additional treatment, such as manipulation or capsular release.
The VAS score for pain after 1 month of follow-up was significantly improved in the HD group compared with the IAI group (P = .035). However, at 3 and 6 months of follow-up, there were no statistically significant differences among the groups in the VAS score for pain (Table II).
The HD group showed a significantly lower VAS score for pain than the IAI group (P = .035) only at 1 month, a higher Simple Shoulder Test score than both the IAI group (P = .02 at 1 month and P = .05 at 3 months) and SAI group (P = .04 at 1 month and P = .004 at 3 months), and a higher Constant score than the IAI group (P = .039 at 1 month and P = .032 at 3 months). No differences between groups were found at 6 months.
* Statistically significant difference (P < .05) between the HD group and IAI group.
† Statistically significant difference (P < .05) between the HD group and SAI group.
For functional scores, at 1 month, the HD group exhibited a significant improvement in the SST compared with the IAI and SAI groups (P = .02 and .04, respectively) and a significant improvement in the Constant score compared with the SAI group (P = .039). At 3 months, the HD group had a significantly improved SST compared with the IAI and SAI groups (P = .05 and .004, respectively) and a significantly improved Constant score compared with the SAI group (P = .032). However, at 6 months, there were no significant differences in the SST or Constant score among the groups (all P > .05).
For ROM, at 1 month, the HD group experienced greater improvement in forward flexion (P = .009 and .007, respectively) and external rotation (P = .005 and .005, respectively) than the IAI and SAI groups (Table III). However, at 3 and 6 months, there were no significant differences among the groups in any ROM measure.
Table IIIRange of shoulder motion by 3 injection methods
The HD group had a greater degree of forward flexion than the IAI and SAI groups (P = .009 and .007, respectively) and greater external rotation than the IAI and SAI groups (P = .005 and .005, respectively) only at 1 month; however, there were no significant differences between groups in all ROMs after 3 months.
* Statistically significant difference (P < .05) between the HD group and IAI group.
† Statistically significant difference (P < .05) between the HD group and SAI group.
No patients exhibited steroid injection–related major complications, such as joint infection, permanent neurogenic symptoms, or steroid-induced arthritis. Several mild adverse effects were reported. Two patients in the IAI group and 1 patient in the SAI group complained of temporary mild dizziness and nausea after the injection. One patient in the HD group reported transient loss of sensation and motor control in the injected arm for a few hours after the injection, but these symptoms recovered without sequelae. One other patient in the HD group reported transient hypotensive syncope immediately after the injection, but the patient fully recovered after several minutes.
The aim of this prospective randomized comparative trial was to evaluate whether 1 of the 3 injection methods for primary frozen shoulder was superior to the others. Steroid injection for the treatment of frozen shoulder has been widely used as a safe and effective treatment modality for pain relief.
It may be used when medication or physical therapy exercise fails to relieve the symptoms of frozen shoulder before consideration of more traumatic or invasive treatments, such as manipulation under anesthesia or surgical capsular release.
48 patients who experienced frozen shoulder for <6 months were randomly assigned to receive weekly shoulder injections into the subacromial bursa or GH joint, but no significant difference in outcome was found. In addition, Oh et al
demonstrated that IAI offered no advantage over SAI in patients with primary frozen shoulder. Our results were similar. However, in this study, the HD group had better outcomes in terms of the VAS score for pain and ROM up to 1 month and functional scores up to 3 months.
Previous studies reported that HD is relatively safe and cost-effective and that it provides direct results.
Although numerous case series have reported favorable results for arthrographic shoulder joint distention, most have included corticosteroids; therefore, it may not be possible to directly attribute the effect to joint distention alone.
The efficacy of HD is believed to be attributable to the combination of injected steroids and capsular distention or rupture, and better outcomes may naturally be expected in the HD group. In this study, we found superior results in the HD group compared with those in the IAI and SAI groups in the early postinjection period, but the difference disappeared at the later follow-up period of 6 months. We believe that the combined effect of capsular distention or rupture in the HD group could lead to favorable results in the early period, but this is not maintained at later periods. Steroid injection alone without HD, into either the intra-articular or subacromial space, may be sufficient to guarantee results comparable to those by injection plus HD 3 months after injection, if appropriate physical therapy exercise is continued. Therefore, clinicians may choose any of the 3 injection methods according to their preference, equipment availability, and patient request. However, we cannot ignore the non-negligible benefit of HD treatment in this study (ie, its rapid treatment effect). Thus, HD would be the most appropriate treatment, especially for patients who expect a quick recovery.
A few studies compared the effect of HD with steroid injection, but the findings were controversial.
reported an increase in shoulder ROM in the HD group compared with the findings in the steroid injection group, which is consistent with our results. We believe this difference may in part be attributable to the number of injections or the stage of frozen shoulder. In contrast to previous studies,
injection was performed only once in all groups. The additional effect of HD may be easier to identify in patients who were injected once rather than in those who received repeated injections, in whom the cumulative effects of steroids may make the difference unclear. Furthermore, the possible mislabeling of different stages of frozen shoulder may affect the results, with no difference in outcomes. In the early inflammatory phase, steroid injection would be more effective, and patients may be unable to tolerate capsular dilatation, with a resultant insufficient injection volume; by contrast, HD is believed to be more effective in the frozen phase than in an earlier phase.
to make the groups more homogeneous by excluding patients in the inflammatory phase, thereby increasing comparability.
Several limitations should be noted in interpreting our findings. First, patients were not blinded to the injection methods, which could have biased the results. Some patients may have had the biased impression that HD is a better treatment method. Second, we used different imaging guidance methods; specifically, IAI and SAI were guided by sonography, whereas HD was guided by an image intensifier and radiopaque contrast material. Third, the volume of injected saline during HD was not consistent because the injection was continued until rupture of the capsule occurred. Fourth, although all participants were instructed to perform a home-based physical therapy exercise program, compliance was not verified individually. Fifth, although the diagnosis of primary frozen shoulder was based on physical examination, radiographs, and ultrasonography, those with labral lesions or small rotator cuff tears could have been missed. These lesions can lead to secondary frozen shoulder, but we did not perform magnetic resonance imaging to confirm its appearance. Sixth, although there was no statistical difference in the incidence of underlying disease among the groups, systemic diseases such as diabetes and hypothyroidism may have affected the treatment results for frozen shoulder. Finally, although we tried to include patients with the same stage of frozen shoulder, the individual natural history of frozen shoulder may have varied.
Among the 3 injection methods of IAI, SAI, and HD for primary frozen shoulder, HD led to better pain scores and ROM at 1 month and better functional scores at 1 month and 3 months. However, the 3 injection methods provided similar clinical outcomes at the final follow-up of 6 months with the numbers available.
The authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.
We thank Ae-Sun Chang (A.-S.C.) for the support in data collection.
Treatment of rigid shoulders by joint distension during arthrography.