Hydrodilatation involves injecting a contrast medium, a local anesthetic, cortisone, and up to 40 ml of sterile saline solution intra-articularly in order to stretch the joint capsule and break up adhesions [7, 1]. A 2008 Cochrane Review presented “silver” level evidence for arthrographic distension with saline and steroids [8]. This review of 196 patients in 5 clinical trials provides evidence for short-term benefits in pain, range of movement, and function in frozen shoulder; however, these studies did not compare the efficacy of capsular distention to other treatment interventions. Buchbinder et al. [7] demonstrated significant improvement in pain and shoulder function 3 months following arthrographic distension with a mixture of saline and steroid [7]. In a prospective study by Quraishi et al. [30], 48 patients with frozen shoulder either underwent hydrodilatation or manipulation under anesthesia. Patient satisfaction and Constant scores were greater in the hydrodilatation group at 6 months; however, range of motion did not differ between groups [30]. Koh et al. [22] assessed a capsule-preserving method of hydraulic distension with saline solution and corticosteroid in patients with frozen shoulder [22]. At the time of the injection, pressures were monitored to avoid capsule ruptures. Results revealed a significant decrease in capsule stiffness and an increase in capsular volume following a series of three injections over a 1-month period. Patients also reported a significant reduction in pain and increases in shoulder range of motion. Elleuch et al. [14] have demonstrated that the benefits of capsular distension followed by an intense rehabilitation program are greater than conservative rehabilitation only and are maintained 12 months following the procedure [14].

Hydrodilatation is a safe, effective, and less invasive treatment approach for the management of frozen shoulder [5, 9]. Capsular rupture does not appear to be key to achieving success; however, the use of a steroid is recommended in conjunction with rehabilitation. Reported range of motion improvement has been shown to be 28° in forward flexion, 42° in abduction, 22° in internal rotation, and 26° in external rotation [9, 34]. We have never had great success with this technique.

Manipulation under anesthesia (MUA) involves applying a passive stretch in forward flexion, abduction, and adduction with the scapula stabilized as well as into the extreme ranges of internal and external rotation. Tearing of the capsule may be palpated or even audible by the physician. Closed MUA is typically not indicated in posttraumatic and postsurgical cases of shoulder stiffness due to the increased risk of fracture [19].

Hamdam and Al-Essa prospectively studied 88 patients (98 shoulders) with frozen shoulder that previously failed conservative treatment [18]. Patients received either MUA alone or MUA with one of two types of intra-articular injection (methylprednisolone or a large volume [50–100 cc] of normal saline), followed by physical therapy. Patients who received MUA with an intra-articular normal saline injection had better outcomes at 6–8 months.

Ng et al. [26] reported the outcomes of MUA plus physical therapy in 86 patients with frozen shoulder with a mean duration of symptoms of 13 months [26]. They found a significant improvement in function as measured by the Disabilities of the Arm, Shoulder and Hand (DASH), a significant reduction in VAS pain, as well as improvement in shoulder range of motion (53° in forward flexion, 80° in abduction, and 32° in external rotation) 6 weeks after intervention. Other studies have also demonstrated similar outcomes with longer-term follow-up [35, 15, 18, 25, 28]. Patients with idiopathic or posttraumatic shoulder stiffness have better outcomes compared to patients with postsurgical shoulder stiffness [35].

It is well established that diabetes is a strong risk factor for the development of frozen shoulder; however, the impact of diabetes on intervention remains unclear [2, 6, 16, 33]. Downie et al. [12] compared the outcome of intervention for frozen shoulder in patients with and without diabetes [12]. This retrospective chart review of 148 patients who underwent MUA or arthroscopic release demonstrated similar functional outcomes as measured by DASH and Constant scores. Similar results have been presented by Wang et al., demonstrating no difference in short- or long-term outcomes following MUA in patients with and without noninsulin-dependent diabetes mellitus, [36] though the risk of recurrence is high [20]. On the contrary, the previously mentioned study by Hamdam et al. (2003) reported worse outcomes in diabetic patients under going MUA and intra-articular injection compared to their nondiabetic peers [18] and others have reported no benefit of MUA in patients with diabetes [20].

A 2013 systematic review compared patient outcomes after MUA or arthroscopic capsular release for frozen shoulder [17]; 989 patients in 22 studies were included. At a median follow-up of 35 months (range: 3–189), the differences between the MUA and capsular release groups were minimal for changes in abduction, flexion, and external rotation range of motion and final Constant score. As the quality of evidence is low, it is difficult to provide any strong clinical recommendations based on the existing published literature, and randomized controlled trials are needed to support the use of either modality.

Complications of closed MUA include iatrogenic humeral fractures, glenohumeral dislocation, rotator cuff and labral tears, brachial plexus injuries, and hemorrhagic effusions and hematomas [19, 24]. The senior author encourages the use of closed manipulation in a slow fashion with full cooperation by the anesthetist to ensure complete muscle relaxation prior to commencing. Success rates have been found in the nondiabetic population avoiding arthroscopic capsular release.

Arthroscopic capsular release involves releasing scar tissue formation from the underside of the subscapularis tendon through the rotator interval, followed by release of the anterior then inferior and posterior capsule. Release of these structures improves humeral head translation in the inferior and lateral directions [19]. Jerosch [21] described the 360° capsular release, which involves release of the anterior, posterior, superior, and inferior capsule [21]. Abduction range of motion improved from 76° preoperatively to 167° postoperatively. External rotation with the arm in abduction improved from 4° preoperatively to 85° postoperatively. Internal rotation with the arm in abduction improved from 17° preoperatively to 63° postoperatively. While complications of arthroscopic capsular release may include shoulder instability and dislocation in overly aggressive releases, as well as axillary nerve injury when releasing the inferior capsule, there were no complications reported in this series of 28 patients.

Snow et al. [32] compared anterior and inferior release to anterior, inferior, and posterior capsular release in patients with frozen shoulder [32]. They did not find any added benefit of the posterior capsular release; all groups demonstrated significant improvements in function on the Constant score, improved range of motion, and decreased pain. Eighty-nine percent of patients reported that they felt their condition was better or much better following the arthroscopic release.