The term “frozen shoulder” was first used by Codman [1] and thereafter Neviaser [2] noted that the pathology of the condition was actually located at the capsule of the shoulder joint and therefore called it “adhesive capsulitis”. The typical findings are pain and a global restriction of movement, with limited passive external rotation being the most notable [3]. Frozen shoulder management presents the clinicians with an opportunity to use all skills to alleviate pain and restore function of the shoulder.

Many treatment options for adhesive capsulitis have been described, including rest, NSAIDs, active and passive mobilization, physiotherapy, intra-articular corticosteroids, intra-articular hyaluronate injection, manipulation under anaesthesia when conservative treatment fails, and finally arthroscopic capsular release [4-7]. One of the main goals of treatment is to restore shoulder function through manipulation and therapeutic exercises in which the patient must cooperate and take an active part. The most important factor limiting patients’ cooperation in exercise is pain. Hence, regional nerve block, attributable to its role in pain relief, can be used before the exercise program [8]. Among various nerve block techniques, suprascapular nerve block (SSNB) is an effective and simple method for the management of shoulder pain, with no significant complications reported in over 2000 procedures apart from rare vasovagal episodes [9-12].

Although frozen shoulder is believed to be a benign self-limiting disorder, which tends to be resolved over 1-2 years, authors suggested that patients with significant stiffness are good candidates for manipulation under anaesthesia rather than conventional treatment because conventional treatment of intensive physiotherapy must be carried out for months to years in order to regain the range of motion (ROM) [13,14].

It must be emphasized that even after manipulation of shoulder, a regular supervised physiotherapy is critical to ensure a mobile painless shoulder otherwise significant stiffness quickly return. Multiple shoulder manipulation techniques have been described, including manipulation with steroid injection and manipulation under general or local anaesthesia. Fracturing the humerus during shoulder manipulation is a common complication, in addition to shoulder dislocation, post-manipulation pain, hemoarthrosis, tearing of the joint capsule or rotator cuff, and traction injury to nerves [15].

The Codman’s manipulation refers to a specific pattern of motion at the shoulder joint leading to an indirect humeral rotation without placing a rotational torque on the humerus, thereby reducing fracture risk during manipulation. This is achieved when the arm performs a closed-loop motion by three consecutive 90° rotations defined as Codman’s rotations, each around the respective coordinate axis. Such rotations will lead to an apparently indirect 90° rotation around the longitudinal axis of the humerus [16,17].

The prevalence of adhesive capsulitis is 2-5% in a normal population [18,19]. It is more common in females and between the ages of 40 and 60 years [1,20]. A genetic component is reported although the direct mechanisms by which genes influence soft tissue disorders are still unknown [21]. Contra lateral shoulder involvement shoulder involvement reported in up to 20-30% of patients and recurrence in ipsilateral shoulder is unusual [18].

The natural history of idiopathic frozen shoulder syndrome is considered benign. Codman [1] and Grey [22] stated that frozen shoulder is a self-limiting condition with complete resolution of pain and recovery of range of motion within a maximum of 2 years from the onset of symptoms.

Deplama [23] reported on three patients who had remained symptomatic five, six and eight years after the onset of symptoms with no indication of improvement.

Murnaghan [24] stated that “the time course of return of shoulder motion is quite unpredictable”. The long period of pain and disability reported in cases of frozen shoulder has been the reason for different interventions management.

The pathophysiological process is believed to involve synovial inflammation and fibrosis of the shoulder joint capsule [25]. Cytokines such as Transforming Growth Factor-beta (TGF-β) and Platelet Derived Growth Factor (PDGF) may contribute to the inflammatory process [26]. Hand et al. found a chronic inflammatory response with a chronic inflammatory response with a fibroblastic proliferation suggesting the process to be immunomodulated [25].

Four arthroscopic stages described by Naviaser as inflammatory, freezing, frozen, and thawing [2]. In the inflammatory stage, passive ROM is increased with anaesthesia, indicating that ROM is pain limited. Histologically, there are inflammatory infiltrates and hypervascular synovitis with a normal underlying capsule. The freezing stage differs in that passive ROM is similar with or without anaesthesia and histologically shows hypertrophic, hypervascular synovitis with capsular scaring. In the frozen stage, pathological specimens show reduced synovitis and dense scar formation in the underlying capsule. The thawing stage represents resolution and no pathological specimens have been described [27,28]. On the contrary Lundberg [29] documented periarticular inflammatory changes and thickening of the joint capsule without intra-articular adhesions. Rizk et al. [30] discovered thickening and constriction of the capsule. Ozaki [31] found a contracted and hypertrophied coracohumeral ligament.

The diagnosis is made on the basis of medical history and clinical examination. In 1934 Codman [1] proposed the following diagnostic criteria for frozen shoulder:

• Shoulder pain of slow onset.

• Pain felt at the deltoid insertion.

• Inability to sleep on affected side.

• Atrophy of supra- and infra spinatus muscles.

• Restriction of active and passive ROM.

• Painful and restricted elevation and external rotation.

Most patients with idiopathic frozen shoulder have no history of shoulder trauma. They usually give a history of insidious onset of pain, followed by a loss of motion. Night and rest pain are common in the early stages.

The only sign found in the early stages of the disease process is pain experienced at the end range of shoulder motion. Patients presenting with inflammatory and freezing stages have pain on palpation of the anterior and posterior capsule and describe pain radiating to the deltoid insertion. Later on in the disease process, a mild disuse atrophy of the deltoid and supraspinatus muscles can be found. A diffuse tenderness with palpation over the glenohumeral joint can extend to the trapezius and interscapular area [4]. It has been found that complete loss of external rotation is pathognomonic for frozen shoulder [32]. The disease process least affects extension and horizontal adduction movements [33]. Most of the movements in a severely affected frozen shoulder occur at the scapula-thoracic joint.

Plain x-rays mostly reported as normal but some may show periarticular osteopenia due to disuse [34]. These x-rays can assist in excluding other causes of stiff shoulder, such as rotator cuff disease and glenohumeral arthritis [35]. MRI can be helpful in identifying other causes of a stiff shoulder, such as infection or tumors. Laboratory investigations are useful in patients with other medical issues that may lead to secondary frozen shoulder. These include fasting blood glucose, lipid profile and thyroid stimulating hormone.

The decision regarding the best treatment option must be individualized to each patient depending on the stage of the disease and clinical symptoms, as there is no consensus on a standard management protocol.

Medications

Oral non-steroidal anti-inflammatory drugs can be initiated in patients who present with painful limited ROM during the painful freezing phase [6]. Oral steroids have been proposed as a treatment option for frozen shoulder [7]. However, Bushbinder et al. [36] found that, although it did improve the symptoms initially, the effect did not last beyond six weeks. In light of its adverse reactions, some authors suggest that it should not be routinely used for this condition.

Intra-articular steroids

A corticosteroid intra-articular injection has been extensively used in different ways and with different success rates ranging from 44 to 80% [30,37]. A cochrane database review showed that it might be beneficial in the short term and that the effect will not maintained [38]. However, it is more effective when used in combination with other therapies. Carrette et al. [39] found that intra-articular steroids combined with physiotherapy were more effective in improving shoulder ROM than when each of these was used individually. Jacobs et al. [40] also showed that a combination of steroids and distension had the same outcome at two years as manipulation under anaesthesia.

Physiotherapy alone is an effective treatment but is a complement to other therapies, especially to improve the range of movement in external rotation [41,42]. The goal should be to stretch the capsule sufficiently to allow normal glenohumeral biomechanics. Diercks et al. [41] compared the outcome of 77 patients after some received intensive physiotherapy (passive stretching and manual mobilization) and other supervised neglect (active exercises within pain-free range and pendulum exercises). The supervised –neglect group showed the best results with 89% of patients’ having normal painless shoulders compared to the intensive group with only 63% of patients achieving the same results.

Hydrodilation was first described by Andren and Lundberg [43] in 1965, appears to be another good therapeutic intervention for achieving rapid symptomatic relief from adhesive capsulitis [44,45]. It consists of an injection of a solution causing rupture of the capsule by hydrostatic pressure. The solution could be saline solution or combined with corticosteroids [45]. Quraishi et al. [46] showed better results with hydrodilation than manipulation under anaesthesia. They reported that at 6 months follow-up the Constant score showed a statistically significant improvement. However, the ROM had not improved.

Manipulation under anaesthesia

Duplay [47] initially recommended this kind of manipulation as a treatment option for adhesive capsulitis in 1872. It is generally indicated in patients with persistent functional disability in spite of adequate non-operative treatment for 4-6 months. However, opponents cite the risk for dislocation, fracture, nerve palsy, and rotator cuff tears as limitations to this technique [15]. During this procedure, the synovium, the joint capsule especially the inferior axillary pouch of capsule are ruptured, but tears have also been observed to involve the intra-articular long head of biceps and the subscapularis tendon [48]. Some authors [35,49] recommend that an arthroscopic examination be performed before a closed manipulation as they have shown that it helps to reduce stiffness and pain. Physiotherapy is recommended for two to six weeks post-surgery.

Arthroscopic capsular release

The first arthroscopic release was described by Conti in 1979. It is especially recommended in diabetic patients or in patients with postoperative or post-fracture frozen shoulder [50]. Arthroscopy has been considered useful to confirm the diagnosis, to exclude other significant pathology, to classify the stage of the disease and to treat the stiff shoulder with or without manipulation [51]. Potential risks of arthroscopic capsular release include recurrent stiffness, post-operative anterior dislocation and axillary nerve injury at the 6 o’clock position [3]. Pain pumps are suggested to assist in early pain-free mobilization in the first few days. These should be placed in the subacromial space; as some complications have been reported if placed intra-articular [3]. Patients can be started on physiotherapy in hospital and discharged on home exercises that are both passive and active-assisted. Continuous passive motion (CPM) can be helpful in refractory cases [3].

Open surgical release should be considered in patients for whom arthroscopy is contraindicated or has failed [51]. Traditionally, nonoperative management of adhesive capsulitis is recommended for a minimum of six to twelve months before considering operative intervention [52]. However, patients with persistent symptoms and those who have risk factors such as diabetes mellitus or are affected bilaterally might benefit from earlier surgical [53].

Anatomical background

The shoulder joint is supplied primarily by axillary nerve and suprascapular nerve with small branches from the subscapular and lateral pectoral nerves. SSN originates from the upper trunk with contribution from C5-6 and some variable contribution from C4. It travels anterior to the trapezius and parallel to omohyoid, crosses the posterior triangle to enter the suprascapular notch. The superior articular branch comes off 4.5 cm proximal to transverse scapular ligament and continues along with the main nerve beneath the ligament [54]. The SSN then travels towards the spine where it sends a branch to the supraspinatus muscle and winds around the spinoglenoid notch to supply the infraspinatus muscle. In its course along the scapular spine, the inferior articular branch separates from the main nerve and courses obliquely to supply the posterior shoulder joint [55]. SSN supplies 70% of the sensory fibers to the superior and postero-superior shoulder joint, the acromio- clavicular joint, capsule and overlying skin variably [56].

Suprascapular nerve block techniques

Traditionally, SSN blockade has been performed via the use of anatomical landmarks. More recently, the use of imaging guidance to more accurately guide needle placement has been described [56]. Various landmark approaches have been described and can be categorized into posterior, superior and lateral approaches. The posterior approach attempts to block the SSN at the level of suprascapular notch [57-61], while the superior approach aims to block the SSN by surrounding the nerve with local anaesthetic on the floor of supraspinous fossa [62,63]. A lateral approach to localize the SSN has also been described [64,65]. Disadvantages of the posterior approach are the potential absence of suprascapular notch in some individuals and the potential risk of pneumothorax. The superior approach may negate these disadvantages. Dangoisse et al. described an indirect SSN block, using anatomical landmarks [63]. This type of approach is easy and decreases the risk of pneumothorax. It can be performed by most trained specialists.

A 25-G needle has to be introduced through the skin 2 cm cephaloid to the midpoint of the spine of the scapula, with the patient sitting and the upper limbs bending beside the body. Anatomic landmarks must be palpated, such as clavicle, acromioclavicular articulation, acromion, scapular spine, and coracoid process. The entire area must be sterilized with alcohol, and then the needle to be advanced parallel to the blade of the scapula until bony contact is made in the floor of the suprascapular fossa (Figure 1). The needle must be aspired before infusion of anaesthetic solution so that there is no risk this solution enters the blood stream directly. This technique has previously been demonstrated to be safe, and it effectively blocks the articular branches of the suprascapular nerve [63]. For treatment of chronic shoulder conditions, injectable steroids usually are added to the local anaesthetic solution (10 ml solution of 0.5% bupivacaine hydrochloride and 40 mg of methyl prednisolone acetate) [65-70]. Local steroid injection blocks transmission through nociceptive C fibers, thus prolonging the effect of the local anaesthetic through alteration of the function of K channel on the excitable tissue [71,72].

Figure 1: Posterior view of the suprascapular nerve block using the Dangoisse technique. Landmarks are indicated as follows: acromion and lateral end of the scapular spine (a), medial end of the scapular spine (b), midpoint of the scapular spine (c), inferior angle of the scapula (d), and lateral border of the scapula (e). The needle is aligned 2 cm superior to the midpoint of the scapular spine parallel to the blade of the scapula.

SSNB is a safe procedure with a generally low rate of complications. The largest study retrospectively analyzed 1,005 SSNBs performed by multiple clinicians in multiple centers over a 6-year period reported no major complications [73]. There were only 6 minor adverse events which included transient dizziness (n=3), transient arm weakness (n=2), and facial flushing (n=1) [73].

Codman’s manipulation includes three consecutive 90° rotations called elevation, swing, and descending movements.

(1) Starting position: The patient hangs his or her arm along the side with the thumb pointing forward and fingers pointing toward the ground.

(2) Elevation (first move): The arm is elevated 90° in the sagittal plane without rotation about the humeral shaft axis (i.e., thumb points upward and fingers point forward).

(3) Swing (second move): The arm is moved 90° to the coronal plane without rotation about the humeral shaft axis (i.e., fingers now point to the right or left for the right and left shoulders, respectively).

(4) Descending (third move): Finally, the arm is lowered 90° downward (i.e., fingers point to the ground). After these three rotations, the patient will notice that the thumb points to the right or left (for the right and left shoulders, respectively), which means that the arm has rotated by 90° [16,17].

A general law of motion was proposed to answer the question of Codman’s paradox, which is stated as when the long-axis of the arm performs a closed-loop motion by three sequential rotations known as Codman’s rotations, it produces an equivalent axial rotation angle about the long-axis. The equivalent axial rotation angle equals the angle of swing. Validity of the proposed law of motion is demonstrated by computer simulation of various Codman’s rotations [17].

We studied a combined approach including SSNB followed by Codman’s manipulation of the glenohumeral joint and a home program of ROM exercises, pendulum exercises for the arm and stretching techniques for the shoulder joint in patients with idiopathic frozen shoulder [74]. We found active range of motion increased significantly for flexion, abduction, internal rotation and external rotation. A significant decrease of visual analog scale and shoulder disability Questionnaire scores between baseline and follow-up assessments at 1, 6 and 12 weeks post manipulation was also observed [74]. Extension of pain relief for 12 weeks post injection is beyond the pharmacological effect of the drug. There are many possible explanations, including a decrease in central sensitization of dorsal horn nociceptive neurons. In addition, depletion of substance P and nerve growth factor in the synovium and afferent C fibers of the glenohumeral joint after the blockade may also contribute to the longterm relief. Furthermore, a ‘wind down’ (a reduction in peripheral nociceptive input) has been suggested [56,75,76]

In this combined approach, instead of manipulating the shoulder under general anaesthesia in the operating room, Codman’s manipulation following SSNB was used in the outpatient clinic, thus reducing the risk of general anaesthesia, patient discomfort, and treatment cost. Furthermore, no complications were encountered and patients tolerated the procedure well.

Our results were comparable to those of Hollis et al. [77] who performed Codman’s manipulation under general anaesthesia in patients with frozen shoulder in terms of reduction of pain and disability and improvement of ROM. In a previous study, Khan et al. [78] performed manipulation for the glenohumeral joint following SSNB and intra-articular local anaesthesia in patients with idiopathic frozen shoulder, showing a significant decrease in VAS and increase in ROM; however, shoulder disability was not assessed. Our results were similar to those of Khan and colleagues, although we used a different type of manipulation, no intra- articular anaesthesia was used and shoulder disability was assessed using the Shoulder Disability Questionnaire. An additional study was performed by Mitra et al. [79] on patients with frozen shoulder in whom SSNB was performed followed by intra-articular shoulder injection with steroid and local anaesthetic, and finally manipulation was performed in flexion and abduction movements only. The results of our study are in accordance with those of Mitra and colleagues, although our patients were not subjected to the risk of intra-articular injection and the manipulation technique used in our study included rotational movements, thus improving ROM in internal and external rotations, in addition to flexion and abduction, in contrast to the study by Mitra and colleagues in which only flexion and abduction movements showed improvement. Ozkan et al. [80] reported an improvement in shoulder pain following SSNB. Their study varied from ours, as they included only 10 patients with frozen shoulder secondary to diabetes mellitus, which was excluded from our study; no manipulations were performed and shoulder disability was not assessed. Yet, the results of Ozkan and colleagues support our results in the efficacy of SSNB and provide a hope for the management of pain in frozen shoulder.

In a recent meta-analysis of randomized trials, eleven randomized controlled trials that compared SSNB with physical therapy, placebo, and intra-articular injections were included, comprising 591 patients. Regarding pain relief, SSNB provided better pain relief for 12 weeks compared with physical therapy and placebo injections, but was not superior to intra-articular injections. Differences in patient populations and use of pulsed radiofrequency did not cause a significant variation in therapeutic efficacy, but guidance using ultrasound showed consistently better effectiveness than guidance using surface landmarks and fluoroscopy [81].

Combined approach of SSNB followed by Codman’s manipulation and home exercises proved to accelerate the recovery of idiopathic frozen shoulder. This combined approach is effective and safe to be administered in outpatient clinics by a well-trained physician, offering clear advantages (ease of application, low cost, rare side effects) and considering that the top priority of a pain control program is restoration of function to perform usual ADL. It may prove to be a useful treatment for patients who are unfit or unwilling to consider manipulation under anaesthesia. Further, there are economic benefits as patients are able to return to work sooner without the need for hospitalization or spending time in physical therapy sessions.

Conflict of interest

There are no competing interests (financial/potential influence of the contents/ other relationships or activities) involved in this work.

The authors received no financial support for the research, authorship, and/or publication of this article.