Table 3 Characteristics of studies included in the review

Bunker, T. [39]

United Kingdom

Sample: N = 35. Convenience sample. Gender, age, symptom duration and stage of frozen shoulder not reported

Control: Nil

Inclusion:“…fitted the criteria for primary frozen shoulder”

Exclusion: Not reported

Arthroscopy

+ Open release

Co-morbidities, previous management and conservative treatment: Not reported.

Tissue extracted from patients who failed to manipulate.

Naïve tissue: No

Appearance: Consistent abnormality of the subscapularis bursa. Abnormal villous fronding (large, finely divided expansion) of the synovium. Nodular appearance of the synovium.

Histology: Tissue consisted of nodules and laminae of dense collagen (mature' type III). Nodules consisted of a collagen matrix containing fibroblasts arranged alongside layers or bundles of dense collagen. The cell population was moderate to high. Increased vascularity (high or moderate) in seven cases.

Immunocytochemistry; Vimentin (a cytocontractile protein) was strongly expressed.

Myofibroblasts present. Scanty Leukocytes and macrophages (white blood cells). Synovium: (where present) entirely normal or showed minimal papillary infoldings without increased cell production.

Carbone et al. [31]

Italy

Sample: N = 50. Convenience sample. Gender not reported. Mean age = 57.9 years (SD = 9) Symptom duration: Greater than 6 weeks. Stage: “In the freezing stage”

Control: N = 65

RC tear N = 50

Inclusion: Painful stiff shoulder (6 weeks), severe pain effecting ADL, specific clinical sign of FS, night pain, painful restriction of active & passive elevation to < 100°& ≥ 50 % restriction of external rotation. Exclusion: age < 40 or > 70 year, wider tear than short-wide RC tear and with subscapularis tear, massive fluid distension of S-A space, concomitant RC tear & FS (full passive ROM), previous treatment/ trauma shoulder girdle/ spine.

MRI

Co morbidities: Not reported

Previous management: Patients excluded if they had received treatment for shoulder pain—including oral pain relief.

Naive tissue: Yes

Appearance: High intensity signal within the superior subscapularis recess, consistent with fluid distension of the bursa, found in 89.95 % of FS patients. The bursa fluid distension was over, in front of and under the coracoid process.

Carrillon et al. [32]

France

Sample: N = 25. Convenience sample. M:F = 3:22. Mean age = 51 year

Symptom duration: 2–10 months (mean = 6 months. Stage: Not reported

Control:

RC tear N = 15

Inclusion: clinical criteria for FS defined by Codman & Lundberg [9]; Gradually increasing shoulder pain, most severe at rest, ≥ 1 month’s duration, range of anterior elevation of the shoulder no greater than 135°; range of external rotation no >20° and normal GHJ X-ray (no joint space loss, osteophytes, or notches). Exclusion: Not reported.

MRI

(Gadolinium enhancement)

Co morbidities and previous management: Not reported.

Naive tissue: Unknown

Appearance: MRI: Thickening & postgadolinium enhancement (signs of inflammation) of joint capsule and synovial membrane (n = 25), RI (n = 25) & axillary recess (n = 22). No posterior enhancement (signs of inflammation) noted. Postgadolinium enhancement seen in the subacromial bursa (n = 18), supraspinatus & infraspinatus tendons (n = 9) and ACJ (n = 17). Normal tendons of subscapularis and LHB in all patients (n = 25). Arthroscopy (n = 2): Major hemorrhagic thickening of the capsule and synovium at the anterior and inferior part of the joint.

Kilian et al. [33]

Germany

Sample: N = 6. Convenience sample. Gender, mean age, symptom duration not reported. Stage: “Stage II” (Neviaser classification)

Control:

Shoulder Instability

N = 6

Dupytrens N = 6

Not reported.

Arthroscopy

Co morbidities: Not reported

Previous management: Not reported.

Naive tissue: Unknown

Histology: Quantitative Reverse Transcription Polymerase Chain Reaction (Q RT-PCR) Used for quantification of DNA sequences: A significant increase (P < 0.05) of α1(I) mRNA chains in FS. The quantity of α2(I) mRNA chains between FS, Dupuytren and normal capsular tissue showed no difference. The α1(III) mRNA transcription rate was similar in FS, Dupuytren and normal capsular tissue capsule. Immunohistochemistry: Decreased numbers of fibroblast-like cells with intracellular procollagen I staining recognisable in FS. Weak staining of collagen I in FS and Dupuytren’s tissue when compared to normal capsular tissue. Collagen III staining revealed a corresponding distribution pattern in all 3 groups.

Lho et al. [19]

South Korea

Sample: N = 14. Convenience sample. Gender, age, symptom duration and stage of frozen shoulder not reported

Control:

Shoulder Instability

N = 7

Inclusion: Global restriction shoulder PROM. Arthroscopic confirmation of of hypervascular synovitis& thickened RI &capsule. MRI confirmed no pathology in RI, labrum, LHB or ACJ. Exclusion: Not reported

Arthroscopy

Co morbidities, previous management: Not reported

Naive tissue: No

Histology: Elevated IL-1α (Interleukin 1 alpha cytokine) in RI capsule (1.5 +/− 0.15, P < 0.05) and SAB (2.3 +/− 0.24, P < 0.05), compared to control gp (1.0 +/− 0.01 in joint capsule & 2.0 +/− 0.06 in SAB). Elevated IL-1β (interleukin 1 beta cytokine) in RI capsule only (4.3 +/− 0.3, P < 0.05), compared to control gp (3.1 +/− 0.2). Stimulated levels of Tumor necrosis factor alpha cytokine (TNF- α) found in RI capsule (3.1 +/− 0.35, P < 0.05) & SAB (3.5 +/− 0.41, P < 0.01). Elevated levels of IL-6 (Interleukin 6 cytokine) in SAB only (2.2 +/− 0.3, P < 0.01). Cycloogenase COX-1 (enzyme) was increased in the RI capsule only (4.0 +/− 0.14, P < 0.05). Cycloogenase COX-2 (enzyme) was increased in the RI capsule (5.0 +/− 0.15, P < 0.05) and SAB (6.9 +/− 0 .94, P < 0.05) (but not in controls). TNF-α and IL-6 were increased in joint fluid: TNF-α level higher in FS (16.0 +/− 4.04 pg/mL (picograms per millilitre) than controls (10.0 +/− 1.76 pg/mL) (P < 0.05). Increased production of IL-6 in FS (21.8 +/− 4.63 pg/mL) compared to controls (3.7 +/− 0.42 pg/mL) (P < 0.05).

Li et al. [34]

China

Sample: N = 72. M:F = 22:50. Convenience sample. Mean age = 53.5 years

Symptom duration: 15 weeks—18 months (mean = 9.1 months). Stage: Not reported.

Control: N = 120

Inclusion: “Clinical evidence of FS”. Insidious onset pain & dysfunction. Clinical criteria; increasing pain &stiffness >15 weeks, most severe at rest with restriction of PROM > 30° for 2 or more planes of movement. Exclusion: Previous trauma or shoulder surgery, tumours, RC tear, Calcium deposit on radiography, rheumatoid Arthritis, osteoarthritis, diabetes mellitus, thyroid/heart/ pulmonary/cervical disease, stroke.

MRI

Co morbidities: Excluded. Previous management: All had undergone medical treatment including anti-inflammatory medication, +/−physiotherapy followed up for 24 months.

Naive tissue: No

Appearance: Findings in the FS group, but not in control group:1. High-signal intensity soft tissue in the rotator cuff interval. 2. A thickened inferior glenohumeral ligament (axillary recess).3. A low-signal intensity thickened CHL. The CHL was not visualised in 10 out of 120 shoulders in the control group (8.3 %), and 15 out of 72 shoulders in the frozen shoulder group (20.8 %) (P < 0.05). The CHL thickness in FS (3.99+/−1.68 mm) was significantly > control group (3.08+/−1.32 mm), (P < 0.001).

Manton et al. [35]

United States of America

Sample: N = 9. M:F = 7:2.

Convenience (retrospective) sampling

Mean age = 40 year

Symptom duration and stage: Not reported

Control:

Suspected RC or labral pathology N = 19

Inclusion: Arthrographic diagnosis of ≥2 of: Joint volume < 10 ml, poor /absent filling of axillary recess of the joint or biceps tendon sheath, irregularity of capsule insertion, pain after injection of <10 ml of contrast material, or extravasation of contrast material prior to injection of 10 ml or more. Exclusion: Not reported

Direct MRA

(Intra-articular Gadopentetate Dimeglumine)

Co morbidities: Not reported.

Previous management: No distention or anti- inflammatory injection performed before MRI.

Naive tissue: No

Appearance: No SD in amount of fluid in the biceps tendon sheath (P = 0.45) or the axillary recess (P = 0.37) between FS and controls. No corrugation of the synovium in FS, (In controls n = 7). No abnormalities of the rotator interval capsule in FS (In controls n = 7). The average thickness of the synovium and capsule at the axillary recess was 4.1 mm (FS) and 5.1 mm (controls) (P = 0.11). The mean thickness of the capsule at the humeral head was 3.0 mm (FS) and 4.0 mm (controls) (P = 0.07).

Neviaser, J. [40]

United States of America

Sample: N = 53 Case series (1 case study). Gender, age, symptom duration and stage of frozen shoulder not reported

Control: Nil

Not reported.

Arthrography

(Radiographic examination)

(Intra-articular Diodrast)

Co morbidities and previous management: Not reported

Naive tissue: Unknown

Appearance: Thickening and contracture of capsule with resultant decrease injoint capacity and adherence of the reflected fold causing obliteration of the dependent axillary fold. 42/53 patients had decreased joint capacity, obliteration of the axillary fold and frequently a complete/ almost complete absence of the subscapularis bursa.

In every case there was In some instances the subscapularis bursa was obliterated and could not be visualised. The biceps sheath was outlined in the majority of pts. Only 18 % of the shoulders with proved FS showed failure of visualisation of the biceps sheath by arthrogram.

Sofka et al. [14]

United States of America

Sample: N = 47 M:F = 13:33. Convenience sample

Mean age = 53 years Symptom duration and clinical staging:

Stage 1:(0–3 months) n = 8

Stage 2:(3–9 months) n = 23

Stage 3:(9–15 months) n = 8

Stage 4:(15–24 months) n = 8

Control: Nil

Inclusion:“.....either the presumptive clinical diagnosis of FS or MRI findings suggestive of FS”. Exclusion: Not reported

MRI

Co-morbidities and previous management: Not reported

Naive tissue: Unknown

Appearance: Thickening of the axillary pouch ranged from 2–13 mm (average = 7 mm). All subjects demonstrated RI scarring, (mild n = 16,moderate n = 26, severe = n = 5). No SD between the degree of scarring between gps. Analysis of signal intensity of the capsule included n = 5 with isointensity (the same intensity), 13 with hypointensity, and 29 with hyperintensity relative to the normal signal of shoulder capsule. Capsular and synovial thickening (in the axillary pouch) demonstrated the most correlation with clinical stage of FS with a mean axillary pouch thickness for; stage 2 (7.5 mm), stage 1 (4.1 mm), stage 3 (5.5 mm), and stage 4 (4.1 mm) (P < 0.05). No SD for values for stages 1, 3, and 4 when compared to each other. Evaluation of capsular signal was significant (P = 0.02), with hyperintense signal correlating with stage 2.

Song et al. [36]

Korea

Sample: N =35. M:F = 14:21. Convenience sample. Mean age = 50.1 year Symptom duration: At least 4 weeks. Stage: Not reported

Control: N = 45

Inclusion: Clinical Diagnosis: painful stiff shoulder for ≥ 4 weeks, severe shoulder pain affecting ADL/work, night pain, painful restriction of active and passive elevation to < 100°, 50 % restriction of external rotation, normal radiologic appearance, no secondary causes. Exclusion: RC tear, calcium deposition on radiograph. Bony abnormalities, such as # of clavicle/ greater tuberosity of the humerus and bony Bankart lesion, shoulder surgery, or > than specified ROM.

Indirect MRA

(Intra-venous Gadobutrol)

Co-morbidities and previous management: Not reported.

Naive tissue: Unknown

Appearance: FS patients had a significantly thicker joint capsule (5.9 +/− 1.7) in the axillary recess and a significantly thicker enhancing portion (6.5 +/− 2.5) of the axillary recess and of the RI (8.3 +/− 3.4) than control gp (4.2 +/− 1.7; 2.1 +/− 3.0; 3.0 +/− 3.6) (P < 0.001). 5 pts with FS (14 %) and 7 controls (16 %) had subacromial bursitis (P = 1.0). 3 pts with FS (9 %) and 7 controls (16 %) had OA of the ACJ (P = 0.5). No glenohumeral joint effusion was observed in 29 of 35 patients with FS (83 %).

Uhthoff & Boileau [41]

France

Sample: N = 4 . M;F = 0:4.

Convenience sample.

Mean age = 60 year

Symptom duration: 12 months.

Stage: Not reported

Control: Nil

Not reported

Arthroscopy

Dupuytren’s (n = 1)

Previous management: Not reported

Naive tissue: Unknown

Appearance: Marked synovial reaction of the GHJ.

Histology: Little difference in histological findings in synovial tissue & the extracellular matrix of the posterior & anterior structures. Site of biopsy;(1) synovial tissue & capsule from the posterosuperior part of the joint (n = 4); (2) synovial tissue and capsule at the RI (n = 4); (3) tissue from the CHL (n = 4); (4) synovial tissue and capsule from the axillary fold (n = 2); and (5) synovial tissue and inferior capsule in contact with the axillary nerve (n = 1). Vimentin (a cytocontractile protein) expression in synovial and endothelial cells was similar at the level of the posterosuperior site and the RI. Vimentin was strongly expressed in cells and extracellular matrix of the capsule at the RI, the CHL, and the axillary fold.

No expression for vimentin was detected in cells or in the extracellular matrix from posterosuperior capsule specimens. Desmin not expressed in any section. A marked synovial vascular reaction accompanied by formation of villi was found at all sites (intensity varied among different locations). Presence of fibroplasia was evident at all surgically released sites, and areas of spatially nonaligned type III collagen, containing an increased number of fibroblasts, were separated by strands of spatially aligned type I collagen containing the typical fibrocytes in nearly normal numbers. The simultaneous presence of types I and III collagen was similar at all released sites with the exception of the inferior capsule in which little type III collagen was found. Signs of inflammation or perivascular infiltration were not detected in any section.

Xu et al. [37]

Australia

Sample: N = 8. M:F = 5:3. Sample: Unclear. Mean age = 58 years

Symptom duration: 4–9 months (mean = 6.3 months). Stage: Not reported

Control:

RC pathology N = 10

Inclusion: Pain at night and rest. Radiograph = normal. Decreased ROM under anaesthetic. Evidence of synovial fibroblastic proliferation & associated fibrosis on histological examination of biopsy samples. Exclusion: Previous surgery, radiographic signs of shoulder girdle #, Rheumatoid Arthritis, pts with FS & RC tear at same time.

Arthroscopy

Co morbidities and previous management: Not reported.

Naive tissue: Unknown

Appearance: Capsular tissue from FS patients was thickened and hyperaemic. Subsynovial hypercellularity was noted, with fibroblastic proliferation and associated variable, focally prominent collagen production and fibrosis. Associated prominent small vascular channels and vascular congestion was seen. [In RC tissue, plump connective tissue cells in a loose fibrous stroma were noted, vascular proliferation was not present, and fibroblastic proliferation with fibrosis was not evident.]. PGP9.5 (a pan-neuronal marker) and GAP43 (a neuronal membrane protein, nerve marker) immunoreactions: The immunoreactivity pattern of distribution of the nerve markers PGP9.5 and GAP43 was similar in capsular tissue from FS and from controls– Both were mainly seen in the subsynovial tissue adjacent to blood vessels. In the FS tissue, PGP9.5 nerve fibres were often observed close to small blood vessels and within the fibroblastic tissue. The expression of PGP9.5 and GAP43 was significantly higher in FS samples (2.8 +/− 0.2 and 2.4 +/− 0.4 per field) than in rotator cuff tear samples (1.6 +/− 0.6 and 1.3 +/− 0.4 per field, P < 0.05). CD34 (a blood vessel marker) immunoreactions: CD34 was strongly expressed in the capsular tissue in 6 FS patients (75 %) but in only 1 rotator cuff tear patient (10 %), supporting increased vascularity in the FS samples. Increased subsynovial vascularity and increased numbers of plump fibroblasts were observed in FS compared with RC patients. Vascular proliferation and congestion in the subsynovial fibrous tissue was seen only in FS.P75 (a nerve growth factor (NGF) receptor - neurotrophin receptor) immunoreactions:P75 was expressed in vascular adventitia (the outer most connective tissue) and nerve fibres around blood vessels and was frequently seen in the subsynovial tissue. Although not everywhere, increased expression of P75 was observed in the FS samples compared with RC patients. Moderate to strong staining for P75 antibody was noted in the capsular tissue in 100 % of FS but only in 30 % of RC samples.

Zhao et al. [38]

China

Sample: N = 60 M:F = 24:36. Sample: Unclear. Mean age = 50.2 years

Symptom duration: 15 weeks - 30 months (mean = 12 months)

Stage: “Patients were classified into early or late stage” Further details unclear.

Control: N = 60

Inclusion: Clinically diagnosed with FS, insidious onset of pain and dysfunction. Clinical criteria: increasing pain and stiffness for > 15 weeks, most severe at rest, with restriction of PROM greater than 30° in two or more planes of movement. Exclusion: Previous surgery or trauma. Neurological disorder involving the upper limbs. Clinical history and clinical examination compatible with RC tear. Presence of calcium deposition on radiographic evaluation, Rheumatoid arthritis, Osteoarthritis.

MRI

Co morbidities: Not reported

Previous management: Not reported

Naive tissue: Unknown

Appearance: FS pts had a significantly thicker CHL (4.21 mm +/− 0.97) than control subjects (2.12 mm +/− 0.84, P < 0.001). Mean thickness of the articular capsule at the RC interval > in FS pts (7.20 mm +/− 2.13) than in controls (4.43 mm +/− 1.16, P < 0.05). Partial or complete obliteration of the subcoracoid fat triangle (“subcoracoid triangle sign”) was significantly more frequent in FS pts compared with control subjects (partial obliteration, 22 vs. 2 cases (73 % vs. 13 %); complete obliteration, 8 vs. 1 cases (26 % vs. 1.6 %), P < 0.001. Synovitis-like abnormalities around the long biceps tendon were also markedly more frequent in patients than in control subjects (18 vs. 2 cases (60 % vs. 6 %), P < 0.05. Patients were not significantly different from control subjects with regard to synovitis-like abnormalities at the articular surface of the subscapularis tendon or in the supraspinatus muscle tendon.