Manual therapy and exercise for adhesive capsulitis (frozen shoulder)

Matthew J Page; Sally Green; Sharon Kramer; Renea V Johnston; Brodwen McBain; Marisa Chau; Rachelle Buchbinder

doi:10.1002/14651858.CD011275

. 2014 Aug 26;2014(8):CD011275. doi: 10.1002/14651858.CD011275

Manual therapy and exercise for adhesive capsulitis (frozen shoulder)

Matthew J Page ¹, Sally Green ¹, Sharon Kramer ¹, Renea V Johnston ², Brodwen McBain ³, Marisa Chau ⁴, Rachelle Buchbinder ^2,^✉

Editor: Cochrane Musculoskeletal Group

PMCID: PMC10882424 PMID: 25157702

Abstract

Background

Adhesive capsulitis (also termed frozen shoulder) is commonly treated by manual therapy and exercise, usually delivered together as components of a physical therapy intervention. This review is one of a series of reviews that form an update of the Cochrane review, 'Physiotherapy interventions for shoulder pain.'

Objectives

To synthesise available evidence regarding the benefits and harms of manual therapy and exercise, alone or in combination, for the treatment of patients with adhesive capsulitis.

Search methods

We searched the Cochrane Central Register of Controlled Trials, MEDLINE, EMBASE, CINAHL Plus, ClinicalTrials.gov and the WHO ICTRP clinical trials registries up to May 2013, unrestricted by language, and reviewed the reference lists of review articles and retrieved trials, to identify potentially relevant trials.

Selection criteria

We included randomised controlled trials (RCTs) and quasi‐randomised trials, including adults with adhesive capsulitis, and comparing any manual therapy or exercise intervention versus placebo, no intervention, a different type of manual therapy or exercise or any other intervention. Interventions included mobilisation, manipulation and supervised or home exercise, delivered alone or in combination. Trials investigating the primary or adjunct effect of a combination of manual therapy and exercise were the main comparisons of interest. Main outcomes of interest were participant‐reported pain relief of 30% or greater, overall pain (mean or mean change), function, global assessment of treatment success, active shoulder abduction, quality of life and the number of participants experiencing adverse events.

Data collection and analysis

Two review authors independently selected trials for inclusion, extracted the data, performed a risk of bias assessment and assessed the quality of the body of evidence for the main outcomes using the GRADE approach.

Main results

We included 32 trials (1836 participants). No trial compared a combination of manual therapy and exercise versus placebo or no intervention. Seven trials compared a combination of manual therapy and exercise versus other interventions but were clinically heterogeneous, so opportunities for meta‐analysis were limited. The overall impression gained from these trials is that the few outcome differences between interventions that were clinically important were detected only up to seven weeks. Evidence of moderate quality shows that a combination of manual therapy and exercise for six weeks probably results in less improvement at seven weeks but a similar number of adverse events compared with glucocorticoid injection. The mean change in pain with glucocorticoid injection was 58 points on a 100‐point scale, and 32 points with manual therapy and exercise (mean difference (MD) 26 points, 95% confidence interval (CI) 15 points to 37 points; one RCT, 107 participants), for an absolute difference of 26% (15% to 37%). Mean change in function with glucocorticoid injection was 39 points on a 100‐point scale, and 14 points with manual therapy and exercise (MD 25 points, 95% CI 35 points to 15 points; one RCT, 107 participants), for an absolute difference of 25% (15% to 35%). Forty‐six per cent (26/56) of participants reported treatment success with manual therapy and exercise compared with 77% (40/52) of participants receiving glucocorticoid injection (risk ratio (RR) 0.6, 95% CI 0.44 to 0.83; one RCT, 108 participants), with an absolute risk difference of 30% (13% to 48%). The number reporting adverse events did not differ between groups: 56% (32/57) reported events with manual therapy and exercise, and 53% (30/57) with glucocorticoid injection (RR 1.07, 95% CI 0.76 to 1.49; one RCT, 114 participants), with an absolute risk difference of 4% (‐15% to 22%). Group differences in improvement in overall pain and function at six months and 12 months were not clinically important.

We are uncertain of the effect of other combinations of manual therapy and exercise, as most evidence is of low quality. Meta‐analysis of two trials (86 participants) suggested no clinically important differences between a combination of manual therapy, exercise, and electrotherapy for four weeks and placebo injection compared with glucocorticoid injection alone or placebo injection alone in terms of overall pain, function, active range of motion and quality of life at six weeks, six months and 12 months (though the 95% CI suggested function may be better with glucocorticoid injection at six weeks). The same two trials found that adding a combination of manual therapy, exercise and electrotherapy for four weeks to glucocorticoid injection did not confer clinically important benefits over glucocorticoid injection alone at each time point. Based on one high quality trial (148 participants), following arthrographic joint distension with glucocorticoid and saline, a combination of manual therapy and supervised exercise for six weeks conferred similar effects to those of sham ultrasound in terms of overall pain, function and quality of life at six weeks and at six months, but provided greater patient‐reported treatment success and active shoulder abduction at six weeks. One trial (119 participants) found that a combination of manual therapy, exercise, electrotherapy and oral non‐steroidal anti‐inflammatory drug (NSAID) for three weeks did not confer clinically important benefits over oral NSAID alone in terms of function and patient‐reported treatment success at three weeks.

On the basis of 25 clinically heterogeneous trials, we are uncertain of the effect of manual therapy or exercise when not delivered together, or one type of manual therapy or exercise versus another, as most reported differences between groups were not clinically or statistically significant, and the evidence is mostly of low quality.

Authors' conclusions

The best available data show that a combination of manual therapy and exercise may not be as effective as glucocorticoid injection in the short‐term. It is unclear whether a combination of manual therapy, exercise and electrotherapy is an effective adjunct to glucocorticoid injection or oral NSAID. Following arthrographic joint distension with glucocorticoid and saline, manual therapy and exercise may confer effects similar to those of sham ultrasound in terms of overall pain, function and quality of life, but may provide greater patient‐reported treatment success and active range of motion. High‐quality RCTs are needed to establish the benefits and harms of manual therapy and exercise interventions that reflect actual practice, compared with placebo, no intervention and active interventions with evidence of benefit (e.g. glucocorticoid injection).

Keywords: Adult; Female; Humans; Male; Middle Aged; Bursitis; Bursitis/therapy; Exercise Therapy; Exercise Therapy/methods; Glucocorticoids; Glucocorticoids/therapeutic use; Injections, Intra‐Articular; Musculoskeletal Manipulations; Musculoskeletal Manipulations/methods; Randomized Controlled Trials as Topic; Shoulder Pain; Shoulder Pain/therapy

Plain language summary

Manual therapy and exercise for frozen shoulder (adhesive capsulitis)

Background

Frozen shoulder is a common cause of shoulder pain and stiffness. The pain and stiffness can last up to two to three years before going away, and in the early stages it can be very painful.

Manual therapy comprises movement of the joints and other structures by a healthcare professional (e.g. physiotherapist). Exercise includes any purposeful movement of a joint, muscle contraction or prescribed activity. The aims of both treatments are to relieve pain, increase joint range and improve function.

Study characteristics

This summary of an updated Cochrane review presents what we know from research about the benefits and harms of manual therapy and exercise in people with frozen shoulder. After searching for all relevant studies published up to May 2013, we included 32 trials (1836 participants). Among the included participants, 54% were women, average age was 55 years and average duration of the condition was six months. The average duration of manual therapy and exercise interventions was four weeks.

Key results—manual therapy and exercise compared with glucocorticoid (a steroid that reduces inflammation) injection into the shoulder

Pain (higher scores mean worse pain)

People who had manual therapy and exercise for six weeks did not improve as much as people who had glucocorticoid injection—improvement in pain was 26 points less (ranging from 15 to 37 points less) at seven weeks (26% absolute less improvement).

• People who had manual therapy and exercise rated their change in pain score as 32 points on a scale of 0 to 100 points.

• People who had glucocorticoid injection rated their change in pain score as 58 points on a scale of 0 to 100 points.

Function (lower scores mean better function)

People who had manual therapy and exercise for six weeks did not improve as much as people who had glucocorticoid injection—improvement in function was 25 points less (ranging from 15 to 35 points less) at seven weeks (25% absolute less improvement).

• People who had manual therapy and exercise rated their change in function as 14 points on a scale of 0 to 100 points.

• People who had glucocorticoid injection rated their change in function as 39 points on a scale of 0 to 100 points.

Treatment success

31 fewer people out of 100 rated their treatment as successful with manual therapy and exercise for six weeks compared with glucocorticoid injection—31% absolute less improvement (ranging from 13% to 48% less improvement).

• 46 out of 100 people reported treatment success with manual therapy and exercise.

• 77 out of 100 people reported treatment success with glucocorticoid injection.

Side effects

Out of 100 people, three had minor side effects such as temporary pain after treatment with manual therapy and exercise for six weeks compared with glucocorticoid injection.

• 56 out of 100 people reported side effects with manual therapy and exercise.

• 53 out of 100 people reported side effects with glucocorticoid injection.

Quality of the evidence

Evidence of moderate quality shows that the combination of manual therapy and exercise probably improves pain and function less than glucocorticoid injection up to seven weeks, and probably does not result in more adverse events. Further research may change the estimate.

Low‐quality evidence suggests that (1) the combination of manual therapy, exercise and electrotherapy (such as therapeutic ultrasound) may not improve pain or function more than glucocorticoid injection or placebo injection into the shoulder, (2) the combination of manual therapy, exercise, electrotherapy and glucocorticoid injection may not improve pain or function more than glucocorticoid injection alone and (3) the combination of manual therapy, exercise, electrotherapy and oral non‐steroidal anti‐inflammatory drug (NSAID) may not improve function more than oral NSAID alone. Further research is likely to change the estimate.

High‐quality evidence shows that following arthrographic joint distension, the combination of manual therapy and exercise does not improve pain or function more than sham ultrasound, but may provide greater patient‐reported treatment success and active range of motion. Further research is very unlikely to change our confidence in the estimate of effect.

No trial compared the combination of manual therapy and exercise versus placebo or no intervention.

Summary of findings

Summary of findings for the main comparison. Combination of manual therapy and exercise compared with glucocorticoid injection for adhesive capsulitis (frozen shoulder).

Combination of manual therapy and exercise compared with glucocorticoid injection for adhesive capsulitis (frozen shoulder)
Patient or population: patients with adhesive capsulitis (frozen shoulder)  Settings: general practices in high‐income countries  Intervention: manual therapy plus exercise for six weeks  Comparison: glucocorticoid injection
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No. of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	Glucocorticoid injection	Manual therapy plus exercise
Participant‐reported pain relief   ≥ 30%	See comment	See comment	Not estimable	‐	See comment	No studies reported this outcome
Overall pain (change from baseline)   0‐100 visual analogue scale (lower score = less pain)  Follow‐up: 7 weeks	Mean overall pain (change from baseline) in the control group was  58 points	Mean overall pain (change from baseline) in the intervention group was  26 points lower   (36.8‐15.2 lower)		107  (1 study)	⊕⊕⊕⊝  moderate^a	Absolute risk difference 26% (37% to 15% fewer); relative percentage change 30% (43% to 17% fewer) NNTB 3 (2 to 4)
Function (change from baseline)   Shoulder Disabilty Questionnaire 0‐100 (lower scores = better function)  Follow‐up: 7 weeks	Mean function (change from baseline) in the control group was  39 points	Mean function (change from baseline) in the intervention group was  25 points lower   (35.24‐14.76 lower)		107  (1 study)	⊕⊕⊕⊝  moderate^a	Absolute risk difference 25% (35% to 15% fewer); relative percentage change 38% (51% to 22% fewer) NNTB 3 (2 to 4)
Global assessment of treatment success   Complete recovery or much improvement (self‐rated)  Follow‐up: 7 weeks	Study population^b		RR 0.6   (0.44‐0.83)	108  (1 study)	⊕⊕⊕⊝  moderate^a	Absolute risk difference 30% (48% to 13% fewer); relative percentage change 40% (56% to 17% fewer). NNTB 4 (3 to 8)
	769 per 1000	462 per 1000   (338‐638)
Active shoulder abduction	See comment	See comment	Not estimable	‐	See comment	No studies reported this outcome
Quality of life	See comment	See comment	Not estimable	‐	See comment	No studies reported this outcome
Adverse events	Study population		RR 1.07   (0.76‐1.49)	114^c   (1 study)	⊕⊕⊕⊝  moderate^a	Absolute risk difference 4% (15% fewer to 22% more); relative percentage change 7% (24% fewer to 49% more) NNTH not applicable Adverse events recorded included pain after treatment < 1 day or > 2 days, facial flushing, irregular menstrual bleeding, fever and skin irritation
	526 per 1000	563 per 1000   (400‐784)
The basis for the assumed risk* (e.g. median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; NNTB: number needed to treat for an additional beneficial outcome; NNTH: number needed to treat for an additional harmful outcome; RR: risk ratio.
GRADE Working Group grades of evidence.  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Combination of manual therapy, exercise, electrotherapy and placebo injection compared with glucocorticoid injection for adhesive capsulitis (frozen shoulder)
Patient or population: patients with adhesive capsulitis (frozen shoulder)  Settings: outpatient rheumatology clinic and general practices in high‐income countries  Intervention: manual therapy plus exercise plus electrotherapy for four weeks plus placebo injection  Comparison: glucocorticoid injection
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No. of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	Glucocorticoid injection	Manual therapy plus exercise plus electrotherapy plus placebo injection
Participant‐reported pain relief ≥ 30%	See comment	See comment	Not estimable	‐	See comment	No studies reported this outcome
Overall pain (change from baseline)   Shoulder Pain and Disability Questionnaire (SPADI, 0‐100) (lower scores = less pain)^a   Follow‐up: mean 6 weeks	Mean overall pain (change from baseline) in the control groups was  39.1^b	Mean overall pain (change from baseline) in the intervention groups was  3.78 lower   (19.26 lower‐11.7 higher)^c		86  (2 studies)	⊕⊕⊝⊝  low^d,e,f	Absolute risk difference 4% (19% fewer to 12% more);   relative percentage change 5% (27% fewer to 17% more). NNTB not applicable SMD 0.21 (‐0.65 to 1.07)
Function (change from baseline)   Shoulder Pain and Disability Questionnaire (SPADI, 0‐100) (lower scores = better function)^g   Follow‐up: mean 6 weeks	Mean function (change from baseline) in the control groups was  34.2^h	Mean function (change from baseline) in the intervention groups was  8.56 lower   (0.56‐16.56 lower)ⁱ		86  (2 studies)	⊕⊕⊝⊝  low^d,j	Absolute risk difference 9% (1% to 16% fewer);   relative percentage change 14% (1% to 26% fewer) NNTB 5 (3 to 68) SMD 0.46 (0.03 to 0.89)
Global assessment of treatment success	See comment	See comment	Not estimable	‐	See comment	No studies reported this outcome
Active shoulder abduction	See comment	See comment	Not estimable	‐	See comment	No studies reported this outcome
Quality of life (change from baseline) SF‐36 Physical Component Score (0‐100) (lower scores = worse quality of life) Follow‐up: 6 weeks	Mean quality of life (change from baseline) in the control group was 4.4	Mean quality of life (change from baseline) in the intervention group was 3.3 lower (8.57 lower‐1.97 higher)		49 (1 study)	⊕⊕⊝⊝  low^k,l	Absolute risk difference 3% (9% fewer to 2% more);   relative percentage change 9% (23% fewer to 5% more) NNTB not applicable
Adverse events	See comment	See comment	Not estimable	‐	See comment	No studies reported this outcome
The basis for the assumed risk* (e.g. median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; NNTB: number needed to treat for an additional beneficial outcome; SD: standard deviation; SF: Short Form; SMD: standardised mean difference; SPADI: Shoulder Pain and Disability Index; VAS: visual analogue scale.
GRADE Working Group grades of evidence.  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Study ID	Manual therapy component(s)	Exercise component(s)	Duration of session (minutes)	No. sessions (per week)	No. weeks treatment	Total no. sessions
Buchbinder 2007	Both passive and self‐executed muscle stretching techniques to stretch muscles passing over the glenohumeral joint; cervical and thoracic spine mobilization, glenohumeral joint passive accessory glides; glenohumeral joint passive physiologic mobilization including rotation.	Supervised: strength and coordination exercises for rotator cuff and scapular stabilizers; proprioceptive challenge	30	2 per week in first 2 weeks; 1 per week in next 4 weeks	6	8
Bulgen 1984	Maitland's mobilisations (no other details provided)	NA	NR	3	6	18
Carette 2003	Mobilisation techniques (no other details provided)	Supervised: active ROM exercises (for acute adhesive capsulitis); active and auto‐assisted ROM exercises and isometric strengthening exercises (for chronic adhesive capsulitis)	60	3	4	12
Celik 2010	NA	Supervised: scapulothoracic strengthening (serratus anterior, middle and lower trapezius, latissimus dorsi), upper trapezius stretching, and postural exercises. Home: active assistive ROM exercises (flexion, scapular elevation, and internal and external rotation exercises), posterior and inferior capsule stretching exercises, and self‐stick exercises	Dependent on participants pain and muscle strength	5	6	30
Chan 2010	Passive mobilisation (Grade A and Grade B mobilisation techniques, as advocated by Cyriax for treatment of stage II capsulitis)	Home: active and active‐assisted ROM exercises, capsular stretching exercise, postural correction and scapular stabilising work	30	1	6	6
Chauhan 2011	Deep transverse friction massage of the two tendon supraspinatus and subscapularis as laid by Cyriax 1983, followed by inferior capsular stretching. Deep friction was given transverse to the fiber direction	Supervised: passive ROM exercises. Home: Not specified	60	3	2	6
Cheing 2008	NA	Home: exercises in four directions: (i) forward flexion; (ii) external rotation; (iii) horizontal adduction; and (iv) internal rotation	NR	2 to 3	4	10
Dacre 1989	Mobilisation (no other details provided)	NA	NR	1	4 to 6	4 to 6
Dundar 2009	NA	Supervised: active stretching and pendulum exercises	60	5	4	20
Ghosh 2012	NA	Supervised: active and passive shoulder mobilisation exercises plus shoulder wheel and pulley exercise	NR	NR	NR	NR
Guler‐Uysal 2004	Cyriax approach of deep friction massage and manipulation	NA	60	3	2	6
Harsimran 2011	Either anterior or posterior glide mobilisation (Kaltenborn grade III)	NA	NR	5	1	5
Johnson 2007	Either anterior or posterior glide mobilisation (Kaltenborn grade III)	NA	NR	2 to 3	2 to 3	6
Ma 2006	Mobilisation (no other details provide)	Supervised: active exercises (not specified)	20	5	4	20
Maricar 1999	Mobilisation of shoulder quadrant, shoulder capsular stretch, shoulder flexion, shoulder abduction, shoulder external and internal rotation using Maitland Grade III+ and IV	NA	NR	1	8	8
Maryam 2012	NA	Supervised: active ROM exercises	NR	NR	NR	10
Nellutla 2009	NA	Supervised: either proprioceptive neuromuscular facilitation (PNF) movement patterns for exercises or conventional free exercises, such as finger ladder exercises, Codman's pendulum exercises, and overhead shoulder pulley and shoulder wheel	NR	5	3	15
Nicholson 1985	Passive mobilisation. Generally, in the early sessions gliding and distractive mobilisation techniques were performed with the joint near its neutral position, progressing in the later sessions to mobilisation towards the end of the range of motion	NA	NR	2 to 3	4	8 to 12
Pajareya 2004	Passive mobilisation (no other details provided)	Supervised: passive glenohumeral joint stretching exercises up to the participant's tolerance, based on Cyriax. Home: pulley exercises (actively assisted exercises) and active non‐assisted exercises using a towel and wall	NR	3	3	9
Rainbow 2008	Either high‐velocity, low‐amplitude chiropractic manipulative therapy to the cervical and thoracic spine and glenohumeral joint or Grade 4 mobilisation of the glenohumeral joint according to the supine glenohumeral mobilisation technique	NA	NR	2	6	12
Ryans 2005	Maitland mobilizations which were progressed as the condition improved, and proprioceptive neuromuscular facilitation	Supervised: active exercise therapy with gym equipment	NR	2	4	8
Samnani 2004	NA	Passive ROM exercises	15	6	6	36
Sharad 2011	End range mobilisation techniques. Initially a few minutes of warming up was given using mid‐range mobilization with the patient positioned supine, after which intensive end rage mobilisation techniques, Grades 3 and 4 as described by Maitland in all the movement planes were delivered, interspersed with accessory movements (glides)	NA	Dependent on participants presentation and tolerance	5	3	15
Shrivastava 2011	Either Maitland's Graded Oscillations Technique or Mulligan's Mobilisation with Movement technique	NA	Dependent on participants symptoms	6	2	12
Sirajuddin 2010	Either anterior or posterior glide mobilisation (Kaltenborn grade III)	NA	15	2	3	6
Tanaka 2010	Mobilization techniques used by Vermeulen 2000, which are performed in the end‐ranges of limited joint mobility	NA	40	2	18	36
van der Windt 1998	Passive mobilisation (no other details provided)	Supervised: not specified	30	2	6	12
Vermeulen 2006	Either intensive passive mobilisation techniques in end‐range positions of the glenohumeral joint (applied with intensities according to Maitland grades III and IV) or assive mobilisation techniques within the pain‐free zone (applied with intensities according to Maitland grades I and II)	NA	30	2	12	24
Wen 2009	NA	Supervised: hand exercises, Codman hanging and swinging movement, flexion in front of the body at 90 degrees, hanging the limb and swinging it backwards and forwards, inside and out, and in circles, and wall climbing the 'ladder' exercises (if duration of symptoms was 1‐2 months); passive range of motion exercises, active exercises with gym equipment, active exercises with a wooden stick, and shoulder wheel exercises (if duration of symptoms was 2‐3 months); pulley exercises (if durations of symptoms was 3‐6 months)	NR	7	2	15
Yan 2005	NA	Supervised: either gymnastics training exercises with dumbbells weighing 2‐5 kg or bare‐handed exercises	5 to 10	7	12	84
Yang 2007	Either end‐range mobilisation or mid‐range mobilisation or mobilisation with movement	NA	30	2	3	6
Yang 2012	End‐range mobilisation and scapular mobilisation	NA	NR	2	8	16

OUTCOME	INTERVENTION			CONTROL			EFFECT ESTIMATE
OUTCOME	Mean	SD	n	Mean	SD	n	Mean difference (95%CI)
Overall pain (VAS 0‐100) change from baseline to 3 weeks	17	21	55	32	26	52	‐15.00 [‐23.99, ‐6.01]
Overall pain (VAS 0‐100) change from baseline to 7 weeks	32	29	56	58	28	51	‐26.00 [‐36.80, ‐15.20]
Overall pain (VAS 0‐100) change from baseline to 6 months	54	33	54	63	31	51	‐9.00 [‐21.24, 3.24]
Overall pain (VAS 0‐100) change from baseline to 12 months	59	30	55	70	24	48	‐11.00 [‐21.44, ‐0.56]
Function (Shoulder Disability Questionnaire 0‐100) change from baseline to 3 weeks	6	22	55	19	27	52	‐13.00 [‐22.36, ‐3.64]
Function (Shoulder Disability Questionnaire 0‐100) change from baseline to 7 weeks	14	27	56	39	27	51	‐25.00 [‐35.24, ‐14.76]
Function (Shoulder Disability Questionnaire 0‐100) change from baseline to 6 months	33	34	54	45	30	51	‐12.00 [‐24.25, 0.25]
Function (Shoulder Disability Questionnaire 0‐100) change from baseline to 12 months	38	34	55	42	33	48	‐4.00 [‐16.96, 8.96]
Night pain (VAS 0‐100) change from baseline to 3 weeks	9	23	55	21	26	52	‐12.00 [‐21.32, ‐2.68]
Night pain (VAS 0‐100) change from baseline to 7 weeks	22	30	56	36	28	51	‐14.00 [‐24.99, ‐3.01]
Night pain (VAS 0‐100) change from baseline to 6 months	33	41	54	34	36	51	‐1.00 [‐15.74, 13.74]
Night pain (VAS 0‐100) change from baseline to 12 months	35	39	55	37	33	48	‐2.00 [‐15.91, 11.91]
Passive abduction (degrees) change from baseline to 3 weeks	‐3	13	55	2	12	52	‐5.00 [‐9.74, ‐0.26]
Passive abduction (degrees) change from baseline to 7 weeks	‐1	14	56	4	11	51	‐5.00 [‐9.75, ‐0.25]
Passive abduction (degrees) change from baseline to 6 months	7	17	54	9	12	51	‐2.00 [‐7.60, 3.60]
Passive external rotation (degrees) change from baseline to 3 weeks	‐3	12	55	6	14	52	‐9.00 [‐13.95, ‐4.05]
Passive external rotation (degrees) change from baseline to 7 weeks	‐2	14	56	13	16	51	‐15.00 [‐20.72, ‐9.28]
Passive external rotation (degrees) change from baseline to 6 months	7	21	54	16	18	51	‐9.00 [‐16.47, ‐1.53]
	Events	Total		Events	Total		Risk ratio (95%CI)
Total adverse events	32	57		30	57		1.07 [0.76, 1.49]
Adverse events: pain after treatment lasting <1 day	17	57		9	57		1.89 [0.92, 3.88]
Adverse events: pain after treatment lasting >2 days	13	57		16	57		0.81 [0.43, 1.53]
Adverse events: facial flushing	1	57		9	57		0.11 [0.01, 0.85]
Adverse events: irregular menstrual bleeding	0	57		6	57		0.08 [0.00, 1.33]
Adverse events: fever	1	57		4	57		0.25 [0.03, 2.17]
Adverse events: skin irritation	2	57		1	57		2.00 [0.19, 21.44]
Adverse events: any other adverse event	4	57		6	57		0.67 [0.20, 2.24]
Global assessment of treatment success ("complete recovery or much improvement) at 7 weeks	26	56		40	52		0.60 [0.44, 0.83]

Ma 2006: Mobilisation plus exercise (intervention) versus acupuncture (control)
Outcome	Intervention	Control
Outcome	Mean; SD not reported
Static pain (VAS 0‐10) at 4 weeks	0.4	0.7
Dynamic pain (VAS 0‐10) at 4 weeks	3.7	4.3
Active flexion (degrees) at 4 weeks	129.9	129.3
Active abduction (degrees) at 4 weeks	98.3	115.7
Active internal rotation (degrees) at 4 weeks	43.9	51.7
Ma 2006: Mobilisation plus exercise plus acupuncture (intervention) versus acupuncture (control)
Outcome	Intervention	Control
Outcome	Mean; SD not reported
Static pain (VAS 0‐10) at 4 weeks	0.7	0.7
Dynamic pain (VAS 0‐10) at 4 weeks	3.3	4.3
Active flexion (degrees) at 4 weeks	137.2	129.3
Active abduction (degrees) at 4 weeks	110.1	115.7
Active internal rotation (degrees) at 4 weeks	54.4	51.7
Chan 2010: Passive mobilisation plus home care plus glucocorticoid injection (intervention) versus home care programme plus glucocorticoid injection (control)
Outcome	Intervention	Control
Outcome	Mean change; SD not reported
Overall pain (VAS 0‐10) change from baseline to 10 weeks	5.7	6.2
Active abduction (degrees) change from baseline to 10 weeks	40.7	80
Active internal rotation (cm) change from baseline to 10 weeks	28.4	38.9
Dacre 1989: Mobilisation (intervention) versus glucocorticoid injection (control)
Outcome	Intervention	Control
Outcome	Mean extracted from Figures; SD not reported
Day pain (VAS 0‐100) at 6 weeks	23.9	25.9
Day pain (VAS 0‐100) at 6 months	18.7	17.2
Pain on motion (VAS 0‐100) at 6 weeks	35.4	35
Pain on motion (VAS 0‐100) at 6 months	21.5	28.5
Passive abduction (degrees) at 6 weeks	128.4	127.2
Passive abduction (degrees) at 6 months	133.1	140.5
Passive internal rotation (mm) at 6 weeks	259.6	290.8
Passive internal rotation (mm) at 6 months	247.5	235.4
Dacre 1989: Mobilisation plus glucocorticoid injection (intervention) versus glucocorticoid injection (control)
Outcome	Intervention	Control
Outcome	Mean extracted from Figures; SD not reported
Day pain (VAS 0‐100) at 6 weeks	21.7	25.9
Day pain (VAS 0‐100) at 6 months	22.6	17.2
Pain on motion (VAS 0‐100) at 6 weeks	34.2	35
Pain on motion (VAS 0‐100) at 6 months	18.1	28.5
Passive abduction (degrees) at 6 weeks	124.8	127.2
Passive abduction (degrees) at 6 months	138.6	140.5
Passive internal rotation (mm) at 6 weeks	276.9	290.8
Passive internal rotation (mm) at 6 months	240.6	235.4
Maricar 1999: Manual therapy plus exercise (intervention) versus exercise (control)
Outcome	Intervention	Control
Outcome	Mean; SD not reported
Active flexion (degrees) at 3 weeks	151.8	146.2
Active flexion (degrees) at 5 weeks	159.5	155.6
Active flexion (degrees) at 8 weeks	170.9	164
Active internal rotation (degrees) at 3 weeks	56.1	50
Active internal rotation (degrees) at 5 weeks	67.4	58.9
Active internal rotation (degrees) at 8 weeks	78.3	70.3
Samnani 2004: Passve ROM exercises and therapeutic activity program (intervention) versus therapeutic activity program (control)
Outcome	Intervention	Control
Outcome	Mean (SD); n not reported
Active functional‐hand‐to‐back score (0‐10 scale, higher scores=increased ROM) at 6 weeks	5.8 (1.9)	2.6 (0.9)
Sharad 2011: End‐range mobilisation plus ultrasound plus exercises (intervention) versus ultrasound plus exercises (control)
Outcome	Intervention	Control
Outcome	Mean; SD not reported
Overall pain (VAS 0‐10) at 3 weeks	0.9	1.09
Active abduction (degrees) at 3 weeks	40.26	22
Active flexion (degrees) at 3 weeks	36.82	21.18
Harsimran 2011: Anterior glide mobilization (intervention) versus posterior glide mobilisation (control)
Outcome	Intervention	Control
Outcome	Median change; SD not reported
Overall pain (VAS 0‐10) change from baseline to 5 days	2.5	3
Johnson 2007: Anterior glide mobilization (intervention) versus posterior glide mobilisation (control)
Outcome	Intervention	Control
Outcome	Median (range); SD not reported
Night pain (0‐4 ordinal scale) at 3 weeks	3 (1‐4)	4 (1‐4)
Overall function (0‐4 ordinal scale) at 3 weeks	2.5 (1‐3)	2 (0‐3)
Shrivastava 2011: Maitland's mobilisation (intervention) versus Mulligan's mobilisation with movement (control)
Outcome	Intervention	Control
Outcome	Mean; SD not reported
Overal pain (VAS 0‐10) at 4 weeks	4.05	3.6
Function (SPADI 0‐100) at 4 weeks	40	42
Flexion (degrees) at 4 weeks (unclear if active or passive)	121.25	122
Abduction (degrees) at 4 weeks (unclear if active or passive)	91.25	99.5
Internal rotation (degrees) at 4 weeks (unclear if active or passive)	46.5	41.25

Date	Event	Description
1 May 2008	Amended	Converted to RM5. CMSG ID C067‐R
24 February 2003	New citation required and conclusions have changed	Substantive amendment
24 February 2003	Amended	This review is based on the original review, 'Interventions for shoulder pain.' Please see published notes for further details

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Overall pain	2		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
1.1 Change from baseline to 6 weeks	2	86	Std. Mean Difference (IV, Random, 95% CI)	0.21 [‐0.65, 1.07]
1.2 Change from baseline to 6 months	2	78	Std. Mean Difference (IV, Random, 95% CI)	‐0.08 [‐1.10, 0.93]
1.3 Change from baseline to 12 months	1	49	Std. Mean Difference (IV, Random, 95% CI)	0.24 [‐0.33, 0.80]
2 Function	2		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
2.1 Change from baseline to 6 weeks	2	86	Std. Mean Difference (IV, Random, 95% CI)	0.46 [0.03, 0.89]
2.2 Change from baseline to 6 months	2	78	Std. Mean Difference (IV, Random, 95% CI)	0.27 [‐0.18, 0.72]
2.3 Change from baseline to 12 months	1	49	Std. Mean Difference (IV, Random, 95% CI)	0.11 [‐0.45, 0.67]
3 Quality of life	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.1 SF‐36 PCS change from baseline to 6 weeks	1	49	Mean Difference (IV, Fixed, 95% CI)	‐3.30 [‐8.57, 1.97]
3.2 SF‐36 PCS change from baseline to 6 months	1	49	Mean Difference (IV, Fixed, 95% CI)	‐1.0 [‐6.27, 4.27]
3.3 SF‐36 PCS change from baseline to 12 months	1	49	Mean Difference (IV, Fixed, 95% CI)	‐1.70 [‐6.97, 3.57]
3.4 SF‐36 MCS change from baseline to 6 weeks	1	49	Mean Difference (IV, Fixed, 95% CI)	0.5 [‐5.60, 6.60]
3.5 SF‐36 MCS change from baseline to 6 months	1	49	Mean Difference (IV, Fixed, 95% CI)	‐2.8 [‐8.90, 3.30]
3.6 SF‐36 MCS change from baseline to 12 months	1	49	Mean Difference (IV, Fixed, 95% CI)	‐1.1 [‐7.20, 5.00]
4 Active range of motion (degrees)	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
4.1 Sum of amplitudes of movement in flexion, abduction and external rotation (degrees) change from baseline to 6 weeks	1	49	Mean Difference (IV, Fixed, 95% CI)	‐12.0 [‐37.09, 13.09]
4.2 Sum of amplitudes of movement in flexion, abduction and external rotation (degrees) change from baseline to 6 months	1	49	Mean Difference (IV, Fixed, 95% CI)	‐10.20 [‐35.29, 14.89]
4.3 Sum of amplitudes of movement in flexion, abduction and external rotation (degrees) change from baseline to 12 months	1	49	Mean Difference (IV, Fixed, 95% CI)	2.5 [‐22.59, 27.59]
5 Active range of motion (cm)	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
5.1 Hand‐behind‐back (cm) change from baseline to 6 weeks	1	49	Mean Difference (IV, Fixed, 95% CI)	0.80 [‐5.72, 7.32]
5.2 Hand‐behind‐back (cm) change from baseline to 6 months	1	49	Mean Difference (IV, Fixed, 95% CI)	‐0.70 [‐7.22, 5.82]
5.3 Hand‐behind‐back (cm) change from baseline to 12 months	1	49	Mean Difference (IV, Fixed, 95% CI)	‐2.90 [‐9.42, 3.62]
6 Passive range of motion	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
6.1 External rotation (degrees) change from baseline to 6 weeks	1	37	Mean Difference (IV, Fixed, 95% CI)	2.40 [‐6.86, 11.66]
6.2 External rotation (degrees) change from baseline to 4 months	1	29	Mean Difference (IV, Fixed, 95% CI)	‐1.10 [‐13.59, 11.39]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Overall pain (VAS 0‐10)	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.1 At 3 weeks	2	48	Mean Difference (IV, Random, 95% CI)	0.23 [‐1.00, 1.46]
2 Active range of motion	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
2.1 Internal rotation at 3 weeks	1	30	Mean Difference (IV, Fixed, 95% CI)	8.26 [4.58, 11.94]
2.2 External rotation change from baseline to 3 weeks	1	18	Mean Difference (IV, Fixed, 95% CI)	‐28.3 [‐36.73, ‐19.87]

Methods	Design: parallel‐group, 2‐arm, single‐blind randomised controlled trial (Australia) Interventions: following arthrographic joint distension, manual therapy plus directed exercises or sham ultrasound Sample size calculation: 78 participants per group were estimated to be needed on the basis of detection of a clinically relevant difference of 10 points in the Shoulder Pain and Disability Index (SPADI) (SD = 23.9) at 3 months at the 5% level of statistical significance with 80% power, including a 10% rate of loss at follow‐up Analysis: intention‐to‐treat analysis using all randomly assigned participants who provided any postbaseline data planned Source of funding: Australian National Health and Medical Research Council Project Grant (non‐industry)
Participants	Number of participants: 156 (78 per group) Baseline characteristics Group receiving arthrographic joint distension followed by manual therapy plus directed exercises Mean (SD) age = 55 (9.3) years; male:female = 24:51 Median (range) duration of symptoms: 6 (3‐60) months Group receiving arthrographic joint distension followed by sham ultrasound Mean (SD) age = 55.3 (7.7) years; male:female = 31:43 Median (range) duration of symptoms: 6 (3‐57) months Inclusion criteria Age ≥ 18 years Symptoms of pain and stiffness in predominantly 1 shoulder for ≥ 3 months, and restriction of passive motion ≥ 30° in ≥ 2 planes of movement, measured to onset of pain with a gravity inclinometer Exclusion criteria Severe pain at rest (> 7 of 10 on a visual analogue scale) Systemic inflammatory joint disease Radiological evidence of shoulder osteoarthritis, fracture or calcification; reason to suspect a complete rotator cuff tear (arm elevation weakness, positive drop arm sign, high‐riding humerus on shoulder radiograph or complete rotator cuff tear on ultrasound) Contraindications to arthrogram and/or distension such as current warfarin therapy Allergy to local anaesthetic or iodinated contrast Pregnancy Likely not to attend for treatment or comply with follow‐up Inability to partake in moderate exercise Previous postdistension physiotherapy Lack of written informed consent
Interventions	Before active physiotherapy or placebo interventions, all participants received arthrographic distension of the glenohumeral joint with glucocorticoid and normal saline performed under radiological guidance. Both groups were permitted to use analgesia and non‐steroidal anti‐inflammatory drugs during the study period. At the end of 6 weeks' treatment, participants were instructed to maintain a 10‐minute daily exercise programme Arthrographic joint distension followed by manual therapy plus directed exercises (N = 78) Components of intervention Manual therapy: both passive and self‐executed muscle stretching techniques to stretch muscles passing over the glenohumeral joint; cervical and thoracic spine mobilisation, glenohumeral joint passive accessory glides; glenohumeral joint passive physiological mobilisation including rotation Supervised exercise: strength and co‐ordination exercises for rotator cuff and scapular stabilisers; proprioceptive challenge Dosage: 30 minutes each session Frequency of administration: Twice per week for 2 weeks, then once per week for 4 weeks (8 sessions) Provider: physiotherapist Arthrographic joint distension followed by sham ultrasound (N = 78) Components of intervention: sham ultrasound and application of a non‐therapeutic gel. Participants received no instruction in exercise techniques and no manual therapy Dosage: 30 minutes each session Frequency of administration: Twice per week for 2 weeks, then once per week for 4 weeks (8 sessions)
Outcomes	Outcomes assessed at the end of 6 weeks' treatment, at 12 weeks (post randomisation) and 26 weeks (post randomisation) Primary outcome Shoulder pain and disability index (SPADI) (0‐100 scale, where a higher score indicates worse pain and/or disability) Secondary outcomes Overall pain using a 10‐cm Likert scale comprising a vertical line with 0 (no pain) at the bottom and 10 (maximal imaginable pain) at the top Pain at night using a 10‐cm Likert scale comprising a vertical line with 0 (no pain) at the bottom and 10 (maximal imaginable pain) at the top Activity‐related pain using a 10‐cm Likert scale comprising a vertical line with 0 (no pain) at the bottom and 10 (maximal imaginable pain) at the top Pain at rest using a 10‐cm Likert scale comprising a vertical line with 0 (no pain) at the bottom and 10 (maximal imaginable pain) at the top Active range of motion in total shoulder flexion, total shoulder abduction and external rotation using a gravity inclinometer (in degrees), and internal rotation measured as the distance (in centimetres) from the base of the occiput to how high the hand would reach up behind the back Health‐related quality of life using the SF‐36 (8 subscales ranging from 0 to 100, where a higher score represents better health) Health‐related quality of life using the Assessment of Quality of Life (AQoL) instrument, which comprises 15 items in 5 dimensions (illness, independent living, social relationships, physical senses and psychological well‐being). All item responses are ordinal scales with 4 levels per item, and scores are scaled from 0.00 (death) to 1.00 (perfect health) Participant‐perceived improvement using a 5‐point ordinal scale (from 1 = failure: marked worsening to 5 = success: much improved and/or completely recovered) Blinded assessor‐perceived improvement using a 5‐point ordinal scale (1 = marked worsening to 5 = marked improvement) Adverse events (elicited by open‐ended questions)
Notes	Trial registered in ANZCTR (https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?ACTRN=12605000685617)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Consenting, eligible participants were randomized in permuted blocks of 4 and 6, stratified by treatment center, to receive either active or placebo regimens according to a computer‐generated table of random numbers created by the study biostatistician"  Comment: An adequate method was used to generate the allocation sequence
Allocation concealment (selection bias)	Low risk	Quote: "To ensure treatment allocation concealment, just prior to commencement of treatment, study centers telephoned a central number for the treatment allocation according to the participant’s identification number. Only the telephone receptionist had access to the allocation schedule (and no other role in the trial)"  Comment: An adequate method was used to conceal the allocation sequence
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "Thirty‐nine participants (53%) in the active group correctly identified their treatment group compared with 35 participants (50%) in the placebo group; 31 participants (42%) in the active group were uncertain which treatment they had received compared with 23 participants (33%) in the placebo group. Blinding index was 0.49 (bootstrap 95% CI 0.40, 0.56), interpreted as moderate success of blinding"  Comment: Given the nature of the interventions, personnel could not be blinded to treatment. However, participants were not informed which intervention was expected to be superior
Blinding of outcome assessment (detection bias)   Self‐reported outcomes	Low risk	Quote: "Between March 2002 and April 2005, we performed a randomized, placebo‐controlled, participant and single assessor blinded trial in participants with adhesive capsulitis..."  Quote: "All participants were evaluated by the same blinded outcome assessor (JMY) at baseline (just prior to arthrographic joint distension), 6 weeks (at the conclusion of the physiotherapy or placebo program), 12 weeks, and 26 weeks"  Comment: Participants who completed self‐reported outcome measures were blind to treatment
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Quote: "Between March 2002 and April 2005, we performed a randomized, placebo‐controlled, participant and single assessor blinded trial in participants with adhesive capsulitis..."  Quote: "All participants were evaluated by the same blinded outcome assessor (JMY) at baseline (just prior to arthrographic joint distension), 6 weeks (at the conclusion of the physiotherapy or placebo program), 12 weeks, and 26 weeks"  Comment: An assessor of objective outcomes was blind to treatment
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "We recruited 156 study participants (78 in both groups), and 144 (74 active, 70 placebo; 92.3%) completed the 26‐week trial. Participants moved through the trial as outlined in Figure 2. Seven participants (3 active, 4 placebo) withdrew from the trial prior to completing the allocated intervention. Because there were no postbaseline followup data for these participants, they were excluded from the efficacy analysis. Three participants withdrew prior to the 12‐week followup (1 active, 2 placebo) and 2 participants (2 placebo) withdrew prior to the 26‐week followup. These participants were excluded from the 12‐week and 26‐week efficacy analyses, respectively. Six participants (2 active, 4 placebo) had a second arthrographic joint distension during the trial period but remained in the efficacy analysis. Overall, of those who completed their allocated treatment, 3 (4%) of 75 in the active group and 8 (10.8%) of 74 in the placebo group had further treatment during the 26‐week trial (P = 0.11)"  Quote: "Characteristics of the 7 participants with no followup data are also provided in Table 1. Although comparisons with the 149 participants are limited, these 7 participants appeared somewhat younger and had worse symptom severity"  Quote: "To assess the sensitivity of the results to the exclusion of the 7 participants who provided no postbaseline data and to the missing data from participants lost to followup during the trial period, we performed a single imputation of these participants’ 3‐month postbaseline values using regression modeling. These models predicted the 12‐week data for these participants based on the relationship between the 12‐week data, baseline characteristics, and randomized treatment arm among the 149 participants who did have postbaseline data. All analyses in Table 3 corresponding to the 12‐week time point were then repeated using the complete data set with these imputed values and the results differed very minimally"  Comment: The numbers and reasons for dropouts and losses to follow‐up were relatively small and balanced between groups, so they are unlikely to have biased the results
Selective reporting (reporting bias)	Low risk	Comment: Outcome data were fully reported for all outcomes reported in the trial protocol of the publication
Other bias	Low risk	Comment: No other sources of bias were identified

Methods	Design: Parallel‐group, 4‐arm randomised controlled trial (United Kingdom) Interventions: Mobilisation 3 times per week for 6 weeks or intra‐articular glucocorticoid injection every week for 3 weeks or ice therapy plus proprioceptive neuromuscular facilitation (PNF) 3 times per week for 6 weeks or no treatment (all were taught pendular exercises) Sample size calculation: not reported Analysis: intention‐to‐treat analysis Source of funding: Arthritis and Rheumatism Council (non‐industry)
Participants	Number of participants: 42 (11, 11, 12 and 8 in each respective group) Baseline characteristics: Baseline characteristics were not reported by group Mean (range) age: 55.8 (44‐74) years; male:female = 14:28 Mean (range) duration of symptoms: 4.8 (1‐12) months Inclusion criteria Pain in the shoulder for at least 1 month, with sleep disturbance due to night pain and inability to lie on the affected shoulder All active and passive shoulder movements were restricted, with reduction in external rotation of at least 50% Exclusion criteria Sensory symptoms or signs in the affected arm or radiation of pain to the neck Generalised arthritis, fractures or dislocations of the humerus, cervical spondylosis or evidence of referred pain
Interventions	All participants were taught pendular exercises and were advised to do them for 2 to 3 minutes every hour. Non‐salicylate analgesics and diazepam 5 mg at night were available as required Mobilisation (N = 11) Components of intervention Manual therapy: Maitland's mobilisations Dosage: not reported Frequency of administration: 3 times weekly for 6 weeks (18 sessions) Provider: physiotherapist Glucocorticoid injection (N = 11) Components of intervention: methyl prednisolone acetate 20 mg and 1% lignocaine hydrochloride 0‐5 mL injected into the subacromial bursa and a similar amount into the shoulder joint by the anterior route Frequency of administration: weekly, for 3 weeks Provider: rheumatologist Ice therapy plus proprioceptive neuromuscular facilitation (N = 12) Components of intervention: ice packs followed by proprioceptive neuromuscular facilitation (PNF) Frequency of administration: 3 times weekly for 6 weeks (18 sessions) Provider: physiotherapist No treatment (N = 8)
Outcomes	Outcomes assessed at baseline, weekly for 6 weeks, then monthly for 6 months. No primary outcome was stated by trialists Night pain using a 10‐cm visual analogue scale (VAS) and by verbal reporting of whether the type of pain was "better," "the same" or "worse" Pain on movement using a 10‐cm VAS and by verbal reporting of whether the type of pain was "better," "the same" or "worse" Rest pain during the day using a 10‐cm VAS and by verbal reporting of whether the type of pain was "better," "the same" or "worse" Passive range of movement (external rotation, total rotation, glenohumeral abduction, total flexion, total abduction, glenohumeral flexion, internal rotation), all measured with a hydrogoniometer to the nearest 5 degrees
Notes	No outcome data were reported in a way that was suitable for inclusion in a meta‐analysis. In the 2003 version of this review, study authors were contacted, but this attempt was unsuccessful. The trial report states, "All patients reported an improvement in pain during the study, but 17 still had residual pain at the end of it, usually mild, but moderate in four. Residual pain was commoner in the ice group, affecting half the patients, but was equally distributed in the other groups. All types of pain were still experienced—rest pain, night pain, and pain on movement. The maximum improvement in pain was achieved by the fourth week of treatment and then continued slowly after that. This was most obvious in the steroid group, but did not reach statistical significance"
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The patients were randomly allocated to one of four treatment groups"  Comment: No information was reported about how the allocation sequence was generated
Allocation concealment (selection bias)	Unclear risk	Comment: No information was reported about how the allocation sequence was concealed
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Comment: Given the nature of the interventions, participants were not blind to treatment and may have had different expectations about the benefits of each intervention
Blinding of outcome assessment (detection bias)   Self‐reported outcomes	High risk	Comment: Unblinded participants who may have had different expectations about the benefits of the intervention they received self‐reported some outcomes (e.g. pain)
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Quote: "Clinical assessment was performed before treatment, weekly for 6 weeks and monthly for a further 6 months by an independent observer who was not aware of the treatment given"  Comment: Outcome assessor of objective outcomes was blind to treatment
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "The patients were randomly allocated to one of four treatment groups: (1) steroid group (11 patients)…; (2) Mobilization group (11 patients)…; (3) Ice group (12 patients)…; (4) Non‐treatment group (8 patients)"  Quote: "Forty‐two patients, 28 females and 14 males, with previously untreated frozen shoulder completed this study"  Comment: All randomly assigned participants completed the study
Selective reporting (reporting bias)	High risk	Comment: No outcome data were reported in a way that was suitable for inclusion in a meta‐analysis. Trialists did not report any numerical data for pain outcomes and presented only mean scores (with no measures of variation) for total flexion, total abduction, external rotation and total rotation in figure format. No results for glenohumeral flexion or glenohumeral abduction were reported
Other bias	Low risk	Comment: No other sources of bias were identified

Methods	Design: parallel‐group, 4‐arm, single‐blind randomised controlled trial (Canada) Interventions: 12 sessions of supervised physiotherapy plus glucocorticoid injection (triamcinolone hexacetonide 40 mg) or glucocorticoid injection alone or supervised physiotherapy plus saline injection or saline injection alone Sample size calculation: 36 participants per group were estimated to be needed on the basis of detection of a clinically relevant difference ≥ 10 points in the Shoulder Pain and Disability Index (SPADI) (SD ≤ 15) at the 5% level of statistical significance with 80% power Analysis: intention‐to‐treat analysis (analysing all participants randomly assigned, using a last observation carried forward analysis) Source of funding: Arthritis Society of Canada (non‐industry)
Participants	Number of participants: 93 (21, 23, 26 and 23 in each respective group) Baseline characteristics Group receiving supervised physiotherapy plus glucocorticoid injection: Mean (SD) age = 54.9 (10.5) years; male:female = 7:14 Mean (SD) duration of symptoms: 22.1 (14.9) weeks Group receiving glucocorticoid injection alone Mean (SD) age = 55.4 (10) years; male:female = 8:15 Mean (SD) duration of symptoms: 21.2 (11) weeks Group receiving supervised physiotherapy plus saline injection Mean (SD) age = 54.2 (8.3) years; male:female = 14:12 Mean (SD) duration of symptoms: 20.8 (11.2) weeks Group receiving saline injection alone Mean (SD) age = 56.5 (9.4) years; male:female = 9:14 Mean (SD) duration of symptoms: 20.3 (7.3) weeks Inclusion criteria Age 18 years or older Symptomatic for < 1 year (defined as the presence of shoulder pain with limitation of both active and passive movements of the glenohumeral joint of ≥ 25% in at least 2 directions (abduction, flexion, external rotation, internal rotation) as compared with the contralateral shoulder or with normal values) Total score ≥ 30 on the Shoulder Pain and Disability Index (SPADI) Exclusion criteria Adhesive capsulitis was secondary to another cause, including inflammatory, degenerative, metabolic or infectious arthritis, cerebrovascular accident or fracture Known blood coagulation disorder or allergy to radiological contrast material
Interventions	All participants were taught a 10‐minute exercise programme consisting of active and auto‐assisted ROM exercises in the planes of flexion, abduction, external rotation and internal rotation (hand behind back) to be done at home twice daily for 3 months Supervised physiotherapy plus glucocorticoid injection (N = 21) Components of physiotherapy intervention Manual therapy: mobilisation techniques (not specified) Supervised exercise: active ROM exercises (for acute adhesive capsulitis); active and auto‐assisted ROM exercises and isometric strengthening exercises (for chronic adhesive capsulitis) Electrotherapy: TENS (for acute adhesive capsulitis); therapeutic ultrasound (for chronic adhesive capsulitis) Other: ice application Dosage: 1 hour overall Frequency of administration: 3 times a week for 4 weeks (12 sessions) Provider: physiotherapist Components of glucocorticoid injection: Under fluoroscopic guidance, a 21‐gauge needle, 2.5–3" long, was directed into the shoulder joint space. Aqueous contrast material (Omnipaque; Sanofi‐Winthrop, Markham, Ontario, Canada) was injected to confirm the correct location of the needle in the joint. This was followed by injection of 40 mg triamcinolone hexacetonide (2 mL) Glucocorticoid injection alone (N = 23) The same injection method as described above was delivered Supervised physiotherapy plus placebo injection (N = 26) The same injection and supervised physiotherapy methods as described above were delivered, except that isotonic saline (2 mL) was injected into the shoulder joint space Placebo injection alone (N = 23) The same injection method as described above was delivered, except that isotonic saline (2 mL) was injected into the shoulder joint space
Outcomes	Outcomes assessed at 6 weeks, 3 months, 6 months and 1 year post randomisation Primary outcome Shoulder pain and disability index (SPADI) (0‐100 scale, where a higher score indicates worse pain and/or disability) Secondary outcomes General health status measured using the SF‐36 Active and passive range of motion in flexion, abduction and external rotation, assessed using a goniometer with the participant in a supine position
Notes	Trialists reported the following protocol violation: "Five patients (2 in the combination group and 1 in each of the other groups) received, in addition to their assigned injection, a glucocorticoid injection (triamcinolone hexacetonide, 20 mg) after randomization, and 1 patient in the saline group underwent rotator cuff repair 8 months after enrolment. All of these injections were prescribed by study investigators who were blinded to the original treatment assignment, and all were done under fluoroscopic guidance. The patient in the placebo group and the patient in the physiotherapy group each received the injection after the 6‐week visit; the 3 patients in the corticosteroid and combination group received it after the 3‐month or 6‐month visits" Unpublished data regarding study design (required for risk of bias assessment) provided by trialist on request
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The assignment scheme was generated from a table of random numbers. Random assignments to the treatment groups were stratified according to study center and balanced after every 12 assignments"  Comment: An adequate method was used to generate the allocation sequence
Allocation concealment (selection bias)	Low risk	Quote: "The opaque prenumbered envelopes containing the assignments were kept by the hospital pharmacist at each center"  Comment: An adequate method was probably used to conceal the allocation sequence
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Quote: "The syringes containing the triamcinolone hexacetonide or saline were prepared by the hospital pharmacist and covered with aluminum foil so the radiologist administering the injections and the patient were not aware of the treatment"  Comment: Participants and personnel were blind to the injection component of the intervention, but not the physiotherapy component. Participants may have had different expectations about the benefits of each intervention
Blinding of outcome assessment (detection bias)   Self‐reported outcomes	High risk	Comment: Participants self‐reported their SPADI and general health scores and were not blind to whether they had received physiotherapy. Participants may have had different expectations about the benefits of each intervention
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Quote: "Each subject was assessed by the same physiotherapist throughout the trial, with a few exceptions. The physiotherapists involved in these assessments were unaware of the treatment allocation and did not normally work in the clinics where the physiotherapy was administered"  Comment: Outcome assessors of objective outcomes were blind to treatment
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Quote: "The primary analysis was based on an intent‐to‐treat principle, and all subjects were included in the analysis. In the case of subjects lost to followup, the data from the last available assessment were imputed to all subsequent evaluations"  Quote: "Of the remaining 93 patients, 2 in the combination group, 9 in the corticosteroid group, 4 in the physiotherapy group, and 1 in the placebo group did not return for all visits"  Comment: Greater loss to follow‐up was seen in the glucocorticoid injection group compared with the other 3 groups, but it is unclear whether the reasons for loss to follow‐up were related to treatment received (or whether they were balanced across groups)
Selective reporting (reporting bias)	Unclear risk	Comment: Outcome data were fully reported for all outcomes reported in the methods section of the publication, but without a trial protocol, it is unclear whether other outcomes were measured but not reported based on the results
Other bias	Low risk	Comment: No other sources of bias were identified

OUTCOME	INTERVENTION			CONTROL			EFFECT ESTIMATE
OUTCOME	Mean	SD	n	Mean	SD	n	Mean difference (95%CI)
Function (Flexilevel Scale of Shoulder Function (FLEX‐SF, 1‐50 scale) at 4 weeks	32.7	1.7	11	31.8	1.5	12	0.90 [‐0.42, 2.22]
Function (Flexilevel Scale of Shoulder Function (FLEX‐SF, 1‐50 scale) at 8 weeks	39.9	1.8	11	32.4	1.6	12	7.50 [6.10, 8.90]
Passive internal rotation (degrees) at 4 weeks	41.3	15.59	11	39.3	14.55	12	2.00 [‐10.35, 14.35]
Passive internal rotation (degrees) at 8 weeks	45.6	20.89	11	42.9	19.75	12	2.70 [‐13.95, 19.35]
Passive external rotation (degrees) at 4 weeks	41.5	15.59	11	38	14.90	12	3.50 [‐8.99, 15.99]
Passive external rotation (degrees) at 8 weeks	60.1	16.58	11	37.7	15.59	12	22.40 [9.22, 35.58]

OUTCOME	INTERVENTION			CONTROL			EFFECT ESTIMATE
OUTCOME	Mean	SD	n	Mean	SD	n	Mean difference (95%CI)
Overall pain (SPADI pain scale 0‐10) at 4 weeks	4.58	1.28	28	4.01	2.1	29	0.57 [‐0.33, 1.47]
Overall pain (SPADI pain scale 0‐10) at 12 weeks	4.39	1.82	28	3.79	2.01	29	0.60 [‐0.39, 1.59]
Function (SPADI disability scale 0‐10) at 4 weeks	4.29	1.91	28	4.03	1.58	29	0.26 [‐0.65, 1.17]
Function (SPADI disability scale 0‐10) at 12 weeks	3.99	1.84	28	3.82	1.61	29	0.17 [‐0.73, 1.07]
Night pain (VAS 0‐10) at 4 weeks	4.84	1.66	28	3.91	2.61	29	0.93 [‐0.20, 2.06]
Night pain (VAS 0‐10) at 12 weeks	4.64	1.77	28	3.74	2.14	29	0.90 [‐0.12, 1.92]
Pain on motion (VAS 0‐10) at 4 weeks	4.93	1.87	28	4.06	2.13	29	0.87 [‐0.17, 1.91]
Pain on motion (VAS 0‐10) at 12 weeks	4.65	1.65	28	3.75	1.92	29	0.90 [‐0.03, 1.83]
Passive internal rotation (degrees) at 4 weeks	64.45	17.8	28	62.89	19.96	29	1.56 [‐8.25, 11.37]
Passive internal rotation (degrees) at 12 weeks	67.19	18.47	28	66.27	17.14	29	0.92 [‐8.34, 10.18]
Passive external rotation (degrees) at 4 weeks	64.9	21.52	28	65.82	17.54	29	‐0.92 [‐11.13, 9.29]
Passive external rotation (degrees) at 12 weeks	68.98	14.22	28	68.22	17.11	29	0.76 [‐7.40, 8.92]
Passive abduction (degrees) at 4 weeks	127.67	26.66	28	137.96	16.26	29	‐10.29 [‐21.80, 1.22]
Passive abduction (degrees) at 12 weeks	137.33	15.31	28	141.75	13.11	29	‐4.42 [‐11.83, 2.99]
Passive flexion (degrees) at 4 weeks	132.78	15.96	28	133.96	10.09	29	‐1.18 [‐8.14, 5.78]
Passive flexion (degrees) at 12 weeks	138.75	14.21	28	139.26	11.19	29	‐0.51 [‐7.17, 6.15]

Methods	Design: parallel‐group, 2‐arm randomised controlled trial (Turkey) Interventions: scapulothoracic exercises added or not added to glenohumeral ROM exercises, TENS, cold pack, home exercise and NSAIDs Sample size calculation: not reported Analysis: intention‐to‐treat analysis Source of funding: not reported
Participants	Number of participants: 29 (14 and 15 in each respective group) Baseline characteristics Group receiving scapulothoracic exercises plus glenohumeral ROM exercises, TENS, cold pack, home exercise and NSAIDs Mean (range) age = 49.6 (38‐62) years; male:female = 2:13 Duration of symptoms not reported Group receiving glenohumeral ROM exercises, TENS, cold pack, home exercise and NSAIDs Mean (range) age = 54.78 (42‐65) years; male:female = 5:9 Duration of symptoms not reported Inclusion criteria Range of motion in external rotation, abduction and flexion less than 50% in comparison with the other shoulder Normal radiography (anteroposterior, lateral) Secondary frozen shoulder diagnosis with MRI showing a small rotator cuff tear Secondary frozen shoulder with type II subacromial impingement syndrome on physical examination and MRI Exclusion criteria Radiculopathy Thoracic outlet syndrome Rheumatological disorders Fractures and tumours of the upper extremity Neurological disorders causing muscle weakness in the shoulder
Interventions	Scapulothoracic exercises plus glenohumeral ROM exercises, TENS, cold pack, home exercise and NSAIDs (N = 15) Components of intervention: Supervised exercise: scapulothoracic strengthening (serratus anterior, middle and lower trapezius, latissimus dorsi), upper trapezius stretching and postural exercises as follows: (1) scapular retraction with exercise band; (2) extension with exercise band; (3) scapular adduction and elevation; (4) wall, table and floor push‐ups; (5) scapular stabilisation with exercise ball in upright standing position; (6) scapular adduction in prone position; (7) extension in prone position; (8) scapular protraction in supine position; (9) push‐up in sitting position; and (10) scapular abduction in upright standing position. In addition, participants were given passive or active assistive ROM and pulley exercises led by physiotherapists at minimum pain level Home exercise: active‐assistive ROM exercises (flexion, scapular elevation and internal and external rotation exercises), posterior and inferior capsule stretching exercises and self‐stick exercises (depending on the pain status of participants) Electrotherapy: TENS Other: cold pack, NSAIDs Dosage: Supervised exercises gradually increased according to the pain and muscle strength of participants. TENS delivered for 20 minutes. During the fifth and sixth weeks of treatment, the dose of scapulothoracic and glenohumeral stretching exercises was increased (considering pain levels), TENS and NSAIDs were terminated and cold pack was applied only when the participant had pain Frequency of administration: 5 times a week for 6 weeks (30 sessions) for supervised exercises, twice a day for home exercise Provider: physiotherapist Glenohumeral ROM exercises, TENS, cold pack, home exercise and NSAIDs (N = 14) The same intervention as described above was delivered, except that scapulothoracic exercises were not undertaken
Outcomes	Outcomes assessed at the end of 6 weeks' treatment and at 12 weeks from baseline. No primary outcome was reported by trialists Pain using a visual analogue scale (scale units not reported) Function using the modified Constant score (0‐100 scale, where a higher score indicates better functional ability) Passive range of motion (elevation/flexion, external rotation and internal rotation)
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The patients were divided randomly into two groups"  Comment: No information was reported about how the allocation sequence was generated
Allocation concealment (selection bias)	Unclear risk	Comment: No information was reported about how the allocation sequence was concealed
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Comment: Given the nature of the interventions (the same multi‐component physical therapy intervention with or without additional exercises), participants were not blind to treatment and may have had different expectations about the benefits of each intervention
Blinding of outcome assessment (detection bias)   Self‐reported outcomes	High risk	Comment: Unblinded participants who may have had different expectations about the benefits of the intervention they received self‐reported pain and function
Blinding of outcome assessment (detection bias)   Objective outcomes	Unclear risk	Comment: Blinding of outcome assessors was not reported, although this could have been done for the objectively assessed outcome—range of motion
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Comment: No dropouts, losses to follow‐up or exclusions were reported, and analyses are reported as based on the number of randomly assigned participants
Selective reporting (reporting bias)	Unclear risk	Comment: Outcome data were fully reported for all outcomes reported in the methods section of the publication, but without a trial protocol, it is unclear whether other outcomes were measured but not reported based on the results
Other bias	Low risk	Comment: No other sources of bias were identified

Methods	Design: parallel‐group, 3‐arm, single‐blind randomised controlled trial (Hong Kong) Interventions: home exercise plus electroacupuncture or home exercise plus interferential electrotherapy or no treatment Sample size calculation: not reported Analysis: per‐protocol analysis Source of funding: not reported
Participants	Number of participants: 74 (25, 24 and 25 in each respective group) Baseline characteristics: Sex of participants was reported as 22 males and 48 females. Age range for all participants was reported as 33‐90 years Group receiving electroacupuncture plus home exercise Mean (SD) duration of treatment = 6.71 (6.5) months Group receiving interferential electrotherapy plus home exercise Mean (SD) duration of treatment = 6.7 (6.05) months Group receiving no treatment Mean (SD) duration of treatment = 8.26 (7.94) months Inclusion criteria Patients who reported localised pain over 1 shoulder, experienced night pain and had restricted active and passive shoulder motions Exclusion criteria History of trauma, fractures, previous shoulder surgery, cervical or thoracic pain syndrome, complex regional pain syndrome, malignancies or anticoagulant therapy Had received acupuncture treatment to the painful shoulder in the past 6 months
Interventions	Home exercise plus electroacupuncture (N = 25) Components of intervention Home exercise: Participants were instructed to follow a chart and perform a standard set of shoulder mobilisation exercises 5 times a day, which included 4 directions: (1) forward flexion—with the help of an overhead pulley system; (2) external rotation—keeping the arm close to the trunk, using a small bamboo to externally rotate the shoulder through pushing against the palm; (3) horizontal adduction—pressing a horizontally adducted arm against the chest with the other arm to achieve horizontal adduction and (4) internal rotation—placing the affected arm behind the back and grasping 1 end of a towel, the other hand then pulling the opposite end of the towel to achieve maximum internal rotation Electrotheracupuncture: Sterile stainless steel acupuncture needles were inserted 15–25 mm intramuscularly into 3 acupoints including 1 trigger point, 1 local point (LI 15: Jianyu) and 1 distal point (ST38: Tiaokou) (14). Trigger points were identified by areas of greatest tenderness around the painful shoulder that were determined on an individual basis. The 2 needles in the shoulder region (trigger point and LI 15) were connected to an electroacupuncture device (Model: ES‐160, ITO Co. Ltd, 3‐3‐3 Tpupta, al‐Minami, Nerima‐ku, Tokyo 176‐86 05, Japan) and were stimulated with an alternating frequency of 2–100 Hz at a pulse duration of 100–400 μs for 20 minutes. The intensity of the stimulation was adjusted to a tolerance level just below the pain threshold. The needle that was applied at the distal point S T38 (Tiaokou) was retained for 20 minutes and was manually lifted and thrusted every 10 minutes Dosage Home exercise: not reported Electroacupuncture: 40 minutes Frequency of administration Home exercise:5 times a day for 6 months Electroacupuncture: 2 to 3 times a week for 4 weeks (10 sessions in total) Provider: physiotherapist Home exercise plus interferential electrotherapy (N = 25) Components of intervention Home exercise: See above Interferential electrotherapy: An interferential electrotherapy machine (a Phyaction Guidance E unit) delivered a current swept from 80‐120 Hz, and 4 suction‐type electrodes were placed around the shoulder region in a co‐planar arrangement. The intensity of the stimulation was adjusted to just below the pain threshold, and the stimulation lasted for 20 minutes Dosage Home exercise: not reported Interferential electroacupuncture: 20 minutes Frequency of administration Home exercises:5 times a day for 6 months Interferential electroacupuncture: 10 sessions over 4 weeks Provider: physiotherapist No treatment (N = 25)
Outcomes	Outcomes assessed at the end of 4 weeks' treatment and at 1, 3 and 6 months' follow‐up for Groups 1 and 2, but only at the end of 4 weeks' treatment for Group 3. No primary outcome was reported by the trialists Constant score (0‐100 scale, where a higher score indicates better functional ability) Pain severity at the moment of assessment, measured using a 10‐cm visual analogue scale, with "No pain" anchored at the left and "Pain as bad as it could be" anchored at the right
Notes	No outcome data from this trial could be analysed, as no outcome data were reported for the comparison group of interest to this review (no treatment group)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The subjects were randomly allocated into: (i) the EA group (n = 24); (ii) IFE group (n = 23); or (iii) control group (n = 23)"  Comment: No information was reported on how the allocation sequence was generated
Allocation concealment (selection bias)	Unclear risk	Comment: No information was reported on how the allocation sequence was concealed
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Quote: "The study was a double‐blind, randomized, controlled clinical trial. An independent assessor was blind to the group allocation"  Comment: Despite reporting this trial as "double‐blind," given the nature of the interventions, participants were not blind to treatment and may have had different expectations about the benefits of each intervention
Blinding of outcome assessment (detection bias)   Self‐reported outcomes	High risk	Comment: Unblinded participants, who may have had different expectations about the benefits of the intervention they received, self‐reported pain and some components of the Constant score
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Quote: The study was a double‐blind, randomized, controlled clinical trial. An independent assessor was blind to the group allocation"  Comment: Outcome assessors of some components of the Constant score were probably blind to treatment
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "One participant dropped out of each of the electroacupuncture group and interferential electrotherapy group, both because of time conflict, and two participants dropped out of the no treatment group because they experienced no improvement" Comment: Although dropout is related to treatment in the no treatment group, the number of dropouts is small and is unlikely to affect function and pain outcomes
Selective reporting (reporting bias)	High risk	Comment: The trialists reported mean (SD) scores for the Constant Murley Assessment scale and VAS pain at the end of 4 weeks' treatment for the electroacupuncture and interferential electrotherapy groups, but not for the no treatment group, because the no treatment group did not have a statistically significant improvement from baseline. Also, without a trial protocol, it is unclear whether other outcomes were measured but not reported based on the results
Other bias	Low risk	Comment: No other sources of bias were identified

Methods	Design: parallel‐group, 3‐arm randomised controlled trial (India) Interventions: active and passive mobilisation exercises plus shoulder wheel and pulley exercises plus ultrasound or manipulation under anaesthesia or glucocorticoid injection (all received home exercise) Sample size calculation: not reported Analysis: per‐protocol analysis Source of funding: not reported
Participants	Number of participants: 72 (24 per group) Baseline characteristics: Baseline characteristics by group were not reported. Sex was not reported Age range: 40‐73 years Duration of symptoms: 0‐2 months (n = 33), 2‐4 months (n = 23), 4‐6 months (n = 16) Inclusion criteria Pain and stiffness of shoulder for 6 months or less Mild osteoporosis Exclusion criteria Diabetes mellitus, rheumatoid arthritis, hyperthyroidism, locked posterior and anterior dislocation, subacromial impingement syndrome or rotator cuff lesion Disease duration longer than 6 months
Interventions	All participants were advised to perform active shoulder mobilisation exercises at home Mobilisation exercises plus shoulder wheel and pulley exercises plus ultrasound (N = 24) Components of intervention Supervised exercises: active and passive shoulder mobilisation exercises plus shoulder wheel and pulley exercises Electrotherapy: ultrasound Dosage: not reported Frequency of administration: for 6 months (number of sessions per week not reported) Provider: physiotherapist Manipulation under anaesthesia (N = 24) Components of intervention: After general anaesthesia, manipulations were done in the sequence of flexion, extension, abduction, adduction, external rotation and internal rotation. Analgesics were given post manipulation period for 2 to 3 days, and shoulder mobilisation exercises started 3 to 4 days after manipulation, which was taught previously Dosage: not reported Frequency of administration: once Provider: not reported Glucocorticoid injection (N = 24) Components of intervention: An injection of methylprednisolone in 40‐mg dosage was given intra‐articularly by the anterior approach under strict aseptic preparation Dosage: See above Frequency of administration: Average of 3 doses at 3‐week intervals Provider: not reported
Outcomes	Outcome assessed at the end of 6 months' treatment Clinical improvement rated as "Good" (no pain, no tenderness present, ROM is equal or comparable with normal limb and no muscle wasting is present), "Fair" (mild pain and tenderness may or may not be present, mild restriction of ROM still present even after 6 months and muscle wasting may or may not be present) or "Poor" (gross restriction of movement is still present, with or without pain)
Notes	To analyse the "treatment success" outcome, we dichotomised participants into those who had a clinical improvement rating of "Good" versus those who had a rating of "Fair" or "Poor." Trialists reported that participants in the study had "almost equal right and left sided affection with one having bilateral affection." However, the group that the bilaterally affected participant was allocated to was not reported, nor was any mention made of controlling for the correlation between shoulders (but this is unlikely to have affected the results substantially, given the dichotomous 'clinical improvement' outcome used)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "These patients were randomly allocated in 3 groups"  Comment: No information was reported about how the allocation sequence was generated
Allocation concealment (selection bias)	Unclear risk	Comment: No information was reported about how the allocation sequence was concealed
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Comment: Given the nature of the interventions, participants were not blind to treatment and may have had different expectations about the benefits of each intervention
Blinding of outcome assessment (detection bias)   Self‐reported outcomes	High risk	Comment: Unblinded participants, who may have had different expectations about the benefits of the intervention they received, self‐reported pain and tenderness
Blinding of outcome assessment (detection bias)   Objective outcomes	Unclear risk	Comment: No information was reported about whether assessors of muscle wasting and range of motion were blind to treatment
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Comment: Only 1 participant (in the glucocorticoid injection group) was lost to follow‐up. This is unlikely to have biased the results
Selective reporting (reporting bias)	Unclear risk	Comment: Results of the single outcome reported in the methods section of the publication (treatment success) were fully reported, but without a trial protocol, it is unclear whether other outcomes were measured but not reported based on the results
Other bias	Low risk	Comment: No other sources of bias were identified

Methods	Design: parallel‐group, 2‐arm quasi‐randomised controlled trial (USA) Interventions: high‐velocity, low‐amplitude chiropractic manipulative therapy to the cervical and thoracic spine and glenohumeral joint plus home exercise therapy, or grade 4 mobilisation of the glenohumeral joint plus home exercise therapy Sample size calculation: not reported Analysis: intention‐to‐treat analysis Source of funding: funded and completed as part of a Master’s in Science (chiropractic) degree for Dr Daniel Rainbow and supervised by Dr Paul Weston at the Division of Health and Social Care (formerly the European Institute of Health and Medical Sciences), University of Surrey, Guilford, United Kingdom
Participants	Number of participants: 8 (4 per group) Baseline characteristics: Baseline characteristics were not reported by group Age range: 35‐60 years Male:female = 2:6 Duration of symptoms not reported Inclusion criteria Male or female 30‐65 years old Diagnosis of primary (idiopathic) phase 2‐3 frozen shoulder, confirmed according to modified guidelines developed by Hannafin, Brotzman and Lundberg, and by history and physical examination Read and signed informed consent document with the understanding that this study received prior institutional review board approval Note that phase 2 frozen shoulder was diagnosed as duration of symptoms 3‐9 months with: decreased glenohumeral motion and loss of synchronous shoulder‐girdle motion; active and passive restriction in elevation (about 90 degrees); external rotation 50%‐60% of normal; and chronic pain with active and passive range of motion Note that phase 3 frozen shoulder was diagnosed as duration of symptoms 9‐15 months with: minimum pain except at the end of range of motion; and significant limitation in range of motion with "rigid end feel" Exclusion criteria Past or current history of stroke, cardiac disease, diabetes, thyroid disorders, intracranial and/or intrathoracic disorders Relative or absolute contraindications to chiropractic manipulative therapy, particularly dislocation, fracture, instability, autoimmune inflammatory disorders and acute severe capsulitis Case history and/or physical and orthopaedic examination findings that disconfirmed a diagnosis of primary phase 2‐3 frozen shoulder Unwillingness to agree or comply with the allocated interventions for 12 treatments (2 treatments per week for 6 weeks) Unwillingness to agree to refrain from or comply with engaging in other forms of treatment during the course of the trial Refusal to read, sign and give informed consent
Interventions	Both groups performed home exercises including Codman's pendulum exercise and simple wall walking. Participants were provided with verbal and written home exercise (with technical instructions and fully illustrated for later reference) Chiropractic manipulative therapy (N = 4) Components of intervention: high‐velocity, low‐amplitude chiropractic manipulative therapy to the cervical and thoracic spine and glenohumeral joint. No more than 2 spinal adjustive procedures and 2 glenohumeral joint adjustive procedures were performed per region (a maximum of 4) per visit. Selected methods for motion palpation and adjustive techniques were left to the discretion of the treating doctor Dosage: not reported Frequency of administration: twice a week for 6 weeks (12 sessions) Provider: chiropractor Mobilisation (N = 4) Components of intervention: grade 4 mobilisation of the glenohumeral joint for 3 minutes according to the supine glenohumeral mobilisation technique. Participant is supine with the arm outstretched and in slight abduction. The chiropractor grasps either side of the proximal humerus with both hands, and the participant's extended forearm is held against the chiropractor's thoracic cage. The chiropractor lightly distracts the shoulder inferiorly and produces mild distraction, circumduction and movement of the shoulder in all directions Dosage: 3 minutes Frequency of administration: twice a week for 6 weeks (12 sessions) Provider: chiropractor
Outcomes	Outcome assessed at the second treatment session each week for 6 weeks Primary outcome Shoulder Pain and Disability Index (SPADI) (0‐100 scale, where a higher score indicates worse pain and/or disability) Secondary outcome Adverse events
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Quote: "Once selected, subjects were allocated to 1 of 2 treatment groups by systematic assignment. Group allocation was determined by the order in which patients qualified and entered the trial. For example, the 1st subject accepted into the trial was allocated randomly to Group 1, accomplished by folding 2 sheets of completely obscured paper (one marked Group 1, the other marked Group 2) and placed in an envelope. One sheet was picked from the envelope and revealed assignment to Group 1. Further quasi‐randomization occurred as the next patient was automatically allocated to Group 2, the 3rd to Group 1, etc. Subjects were unaware of this allocation process prior to presentation"  Comment: A non‐random (predictable) sequence was used to allocate participants
Allocation concealment (selection bias)	High risk	Comment: A non‐random (predictable) sequence was used to allocate participants (see above quote)
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Comment: Given the nature of the interventions, participants were not blind to treatment and may have had different expectations about the benefits of each intervention
Blinding of outcome assessment (detection bias)   Self‐reported outcomes	High risk	Quote: "The SPADI was selected as the primary and singular outcome measure"  Quote: "At each treatment session, patients' subjective symptoms were briefly noted by the primary researcher. The treating doctor (a registered/licensed doctor of chiropractic) was then taken into the room by the primary researcher, and the assigned treatment was administered. This procedure was followed for 12 treatments, with the SPADI administered on the 2nd treatment visit of each week immediately prior to the application of any treatment. All treating practitioners remained blinded to the results generated from the outcome measures throughout the duration of the trial"  Quote: "These findings must be interpreted cautiously, however, secondary to...the absence of blinded assessors"  Comment: Unblinded participants, who may have had different expectations about the benefits of the intervention they received, self‐reported the SPADI
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "All 8 subjects who entered the trial completed the trial with no reported complications, side effects, or adverse reactions"  Quote: "Because no subjects dropped out of the study, intention‐to‐treat analysis was not necessary to calculate"  Comment: No losses to follow‐up, drop‐outs or exclusions occurred
Selective reporting (reporting bias)	Unclear risk	Comment: Outcome data were fully reported for all outcomes reported in the methods section of the publication, but without a trial protocol, it is unclear whether other outcomes were measured but not reported based on the results
Other bias	Low risk	Comment: No other sources of bias were identified

Methods	Design: parallel‐group, 2‐arm, double‐blind randomised controlled trial (India) Interventions: Maitland’s or Mulligan’s mobilisation technique for 2 weeks followed by home exercise programme for 2 weeks Sample size calculation: not reported Analysis: intention‐to‐treat analysis Source of funding: not reported
Participants	Number of participants: 40 (20 per group) Baseline characteristics: duration of symptoms not reported Group receiving Maitland's mobilisation technique plus home exercise programme Mean (SD) age = 59.2 (7.18) years; male:female = 7:13 Group receiving Mulligan's mobilisation technique plus home exercise programme Mean (SD) age = 51.15 (8.53) years; male:female = 12:8 Inclusion criteria Diagnosis of second‐stage adhesive capsulitis with both primary and secondary causes who were showing capsular pattern Exclusion criteria Previous shoulder surgery, shoulder arthritis, painful shoulder, severe systemic illness and use of regular analgesics
Interventions	Both groups received hot fomentation for 10 minutes, along with a 2‐week home exercise programme consisting of Codman's pendulum exercises, scapular setting exercises, finger ladder, wand exercises and stretching of the tightened muscles of the shoulder girdle. The 2‐week home exercise programme was delivered after the 2‐week mobilisation programme Maitland's mobilisation technique (N = 20) Components of intervention: Maitland's graded oscillations technique. The grade of glide was decided during treatment depending on the participant's symptoms. Individual glides delivered were reported as follows: (1) posterior glide: With participant in supine position, the therapist holds his/her arm proximally, applies a distraction force and glides the humeral head posteriorly; (2) inferior glide: Participant lies supine, while the therapist stands at the head end of the participant facing his/her feet. Holding the proximal arm of the participant, the therapist gives a distraction force to the glenohumeral joint and glides the humeral head inferiorly; and (3) the anterior glide: Participant is positioned prone, and the therapist holds the distal arm above the epicondyles with 1 hand for distraction; with the other hand, the therapist applies anterior glide to the humeral head Dosage: dependent on participants' symptoms Frequency of administration: 6 times a week for 2 weeks (12 sessions) Provider: physiotherapist Mulligan's mobilisation technique (N = 20) Components of intervention: Mulligan's mobilisation with movement technique. Passive overpressure was applied in the end of range, and 3 sets of 10 repetitions were given for each mobilisation. Individual glides delivered were reported as follows: (1) flexion and abduction: Participant sits with the therapist posterolateral to him/her. Therapist places the Mulligan belt across the humeral head and to his waist. Leaning backward, he applies a posterolateral glide to the shoulder joint and then asks the participant to perform the painful/restricted movement of shoulder flexion or abduction, which would be pain free; (2) internal rotation: Participant sits or stands with the therapist by his/her side. The therapist applies an inferior glide to the humerus head with the participant’s shoulder in available degrees of abduction. With the glide maintained, the participant actively rotates the shoulder internally without pain; (3) external rotation: Participant lies supine with his/her shoulder horizontally flexed till 90 degrees. The therapist places the belt at the humeral head, applying a lateral distraction to the joint the participant was asked to rotate the shoulder externally Dosage: dependent on participants' symptoms Frequency of administration: 6 times a week for 2 weeks (12 sessions) Provider: physiotherapist
Outcomes	Outcomes assessed at the end of 4 weeks' treatment. No primary outcome was reported by trialists Pain using a 100‐mm visual analogue scale (with a total possible score of 10) Range of motion in flexion, extension, abduction, internal rotation and external rotation using a goniometer (not reported whether passive or active) Shoulder Pain and Disability Index (SPADI) (0‐100 scale, where a higher score indicates worse pain and/or disability) Adverse events
Notes	No measures of variation were reported for any of the outcomes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Subjects were randomly allotted to the two groups, Maitland group and Mulligan group by computerised random sequence generator"  Comment: An adequate method was used to generate the allocation sequence
Allocation concealment (selection bias)	Unclear risk	Comment: No information was reported on how the allocation sequence was concealed
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "Double blinding was done with the assessment therapist and the patient both being blinded with respect to treatment protocol followed"  Comment: Participants received different types of mobilisation but were not provided information that would make them perceive the intervention they received as superior or inferior to the alternative intervention
Blinding of outcome assessment (detection bias)   Self‐reported outcomes	Low risk	Quote: "Double blinding was done with the assessment therapist and the patient both being blinded with respect to treatment protocol followed"  Comment: Blinded participants self‐reported pain, function and adverse events
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Quote: "Double blinding was done with the assessment therapist and the patient both being blinded with respect to treatment protocol followed"  Comment: Assessors of range of motion were probably blind to treatment
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "...no patient dropped out of the study due to problems in therapy"  Comment: No dropouts, losses to follow‐up or exclusions were reported
Selective reporting (reporting bias)	Unclear risk	Comment: Only means with no measures of variation were reported for all outcomes specified in the methods section of the publication. However, it is not clear whether data were incompletely reported based on the statistical significance or magnitude of the results. Also, without a trial protocol, it is unclear whether other outcomes were assessed but not reported based on the results
Other bias	Low risk	Comment: No other sources of bias were identified

Methods	Design: parallel‐group, 3‐arm, double‐blind randomised controlled trial (Japan) Interventions: high‐frequency (more than 2 times a week), moderate‐frequency (once a week) or low‐frequency (less than once a week) joint mobilisation and home self‐exercises Sample size calculation: not reported Analysis: per‐protocol analysis Source of funding: not reported
Participants	Number of participants: 120 (40 per group) Baseline characteristics: Age and duration of symptoms were not reported by group Mean (SD) age: 63.7 (9.1) years Duration of symptoms: 37 had condition for less than 1 month, 39 for less than 3 months, 21 for less than 6 months and 13 for longer than 7 months Group receiving high‐frequency joint mobilisation Male:female = 18:21 Group receiving moderate‐frequency joint mobilisation Male:female = 17:18 Group receiving low‐frequency joint mobilisation Male:female = 17:19 Inclusion criteria Diagnosed with limited glenohumeral joint mobility arising from adhesive capsulitis Unremarkable medical history and no clinical or radiological findings that could explain the decrease in shoulder motion Exclusion criterion Received corticosteroid injections into the joint during the intervention
Interventions	All groups received home exercise performed 2 to 3 times a day, including Codman's or pendulum exercises (circumduction) and passive stretching exercises such as "climbing the wall exercise" (i.e. facing a wall about three quarters of an arm's length away and raising the affected arm up to the shoulder level using only one's fingers without using shoulder muscles). Mean (SD) duration of treatment was 4.6 (1.2) months High‐frequency joint mobilisation (N = 40) Components of intervention: Mobilisation techniques used were those described by Vermeulen 2000, which are performed in the end ranges of limited joint mobility Dosage: 40 minutes Frequency of administration: More than twice a week for a mean (SD) duration of treatment of 4.6 (1.2) months Provider: physical therapist Moderate‐frequency joint mobilisation (N = 40) Components of intervention: Mobilisation techniques used were those described by Vermeulen 2000, which are performed in the end ranges of limited joint mobility Dosage: 40 minutes Frequency of administration: once a week for a mean (SD) duration of treatment of 4.6 (1.2) months Provider: physical therapist Low‐frequency joint mobilisation (N = 40) Components of intervention: Mobilisation techniques used were those described by Vermeulen 2000, which are performed in the end ranges of limited joint mobility Dosage: 40 minutes Frequency of administration: less than once a week for a mean (SD) duration of treatment of 4.6 (1.2) months Provider: physical therapist
Outcomes	Outcomes assessed until the time required to reach the range of motion plateau point (see below). Average length of therapy was 4.6 ± 1.2 months, and follow‐up time was 5.9 ± 1.3 months. No primary outcome was reported by the trialists Improved angle of the shoulder joint, defined as the value obtained by subtracting the affected shoulder joint angle at baseline from the angle of the ROM plateau point (the point in time at which improvement in range of motion (ROM) of the shoulder joint had plateaued for longer than 1 month). Unit of measurement was degrees. Angle of joints was evaluated every week using a large handheld goniometer in terms of active abduction angle at which the capsular pattern of frozen shoulder characteristically caused the greatest restriction Time required to reach the range of motion plateau point (point of time at which improvement in range of motion of the shoulder joint had plateaued for longer than 1 month). Unit of measurement was months
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "From the frequency of sessions for joint mobilization by physical therapists in the hospital setting, the patients were divided randomly into the high‐frequency session group (HF group, more than two times a week), moderate‐frequency session group (MF group, once a week), and low‐frequency session group (LF group, less than once a week)"  Comment: No information was reported on how the allocation sequence was generated
Allocation concealment (selection bias)	Unclear risk	Comment: No information was reported on how the allocation sequence was concealed
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "We performed a randomized, placebo‐controlled, participant and single assessor‐blinded trial"  Comment: Participants, but not personnel, were probably blind to treatment (i.e. probably did not know how the length of their intervention compared with that of participants in other groups)
Blinding of outcome assessment (detection bias)   Self‐reported outcomes	Unclear risk	–
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Quote: "ROM was measured every week by one examiner who was not informed of the group designation of patients"  Comment: Objectively assessed outcomes were probably assessed by an assessor who was blind to treatment
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "The results of ten patients (men, 7; women, 3) were excluded from the data analysis because they received corticosteroidal injections during this study"  Comment: One participant in the high‐frequency group, 5 participants in the moderate‐frequency group and 4 participants in the low‐frequency group were excluded from the analysis because they received corticosteroid injections during the study. As the reason for exclusions was the same for each group, the results are unlikely to be biased because of these exclusions
Selective reporting (reporting bias)	Unclear risk	Comment: Outcome data were fully reported for all outcomes reported in the methods section of the publication, but without a trial protocol, it is unclear whether other outcomes were measured but not reported based on the results
Other bias	Low risk	Comment: No other sources of bias were identified

Methods	Design: parallel, 2‐arm randomised controlled trial (China) Interventions: functional exercises plus shoulder traction plus intermediate frequency electrical stimulation or shoulder traction plus intermediate frequency electrical stimulation Sample size calculation: not reported Analysis: intention‐to‐treat analysis Source of funding: not reported
Participants	Number of participants: 72 (36 per group) Baseline characteristics: Baseline characteristics were not reported by group Mean (range) age: 52 (44‐63) years Male:female = 19:53 Mean (range) duration of symptoms: 4 (2‐7) months Inclusion criterion Unilateral shoulder pain, recurrent and limited shoulder function Exclusion criteria Not reported
Interventions	Both groups received shoulder traction plus intermediate frequency electrical stimulation once a day for 15 days. Shoulder traction is translated as follows: Using a traction grid, the participant is in a sitting or lying position. A strap, attached to a rope pulley suspension and weight, is placed on the participant's wrist. The angle and weight are selected appropriately according to the degree of pain and functional limitation. The participant performs shoulder flexion, abduction, extension, adduction, etc, at different traction angles. Each angle is continued for 20‐30 minutes, with a rest period in the middle. An appropriate session has no more than 2 body positions. If the weight of the suspension does not cause obvious pain to the participant, the traction angle and weight are gradually increased according to the participant's improvement progress, to the point where recovery is that of normal range of motion. Intermediate‐frequency electrical stimulation is translated as follows: Using the J48A model computer with diathermy instrument, the electrode plate is placed on the shoulder at the prominent pain site, with a frequency of 4‐5 KHz, modulating the sine and square waveform, and is adjusted to the participant's tolerance and at a comfortable heat. Participants received 1 session per day (lasting 20 minutes) for 15 days Functional exercises plus shoulder traction plus intermediate‐frequency electrical stimulation (N = 36) Components of intervention: supervised functional exercises. Different exercises were performed depending on the duration of the participant's symptoms. If duration of symptoms was 1‐2 months, the following exercises were performed: hand exercises, Codman's hanging and swinging movement, flexion in front of the body at 90 degrees, hanging the limb and swinging it backward and forward, inside and out and in circles and wall climbing the 'ladder' exercises. If duration of symptoms was 2‐3 months, the following exercises were performed: passive range of motion exercises, active exercises with gym equipment, active exercises with a wooden stick and shoulder wheel exercises. If duration of symptoms was 3‐6 months, pulley exercises were performed Dosage: not reported Frequency of administration: once a day for 15 days (15 sessions) Provider: physical therapist Shoulder traction plus intermediate‐frequency electrical stimulation (N = 36) See above
Outcomes	Outcomes assessed at the end of 15 days' treatment. No primary outcome was reported by the trialists Pain rated on a 30‐point scale (direction of scale unclear) Function rated on a 30‐point scale (direction of scale unclear) Range of motion (not reported whether active or passive) rated on a 25‐point scale (direction of scale unclear)
Notes	This article was written in Chinese and was translated by MC. As the direction of the scale was unclear for all outcomes, and no information about how range of motion was assessed on a 25‐point scale was reported, no outcome data were extracted from this trial
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: [Translated] "Patients were randomly divided into treatment and control group, with each group having 36 patients" Comment: No information was reported about how the allocation sequence was generated
Allocation concealment (selection bias)	Unclear risk	Comment: No information was reported about how the allocation sequence was concealed
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Comment: Given the nature of the interventions (the same multi‐component physical therapy intervention with or without additional exercises), participants were not blind to treatment and may have had different expectations about the benefits of each intervention
Blinding of outcome assessment (detection bias)   Self‐reported outcomes	High risk	Comment: Unblinded participants, who may have had different expectations about the benefits of the intervention they received, self‐reported pain and function
Blinding of outcome assessment (detection bias)   Objective outcomes	Unclear risk	Comment: No information was reported about whether assessors of range of motion were blind to treatment
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Comment: No dropouts, losses to follow‐up or exclusions were reported, and the analyses are reported as based on the number of randomly assigned participants
Selective reporting (reporting bias)	Unclear risk	Comment: Outcome data were fully reported for all outcomes reported in the methods section of the publication, but without a trial protocol, it is unclear whether other outcomes were measured but not reported based on the results
Other bias	Low risk	Comment: No other sources of bias were identified

Methods	Design: multiple‐treatment, 2‐group, 3‐arm randomised controlled trial (Taiwan) Interventions: end range mobilisation or midrange mobilisation or mobilisation with movement Sample size calculation: 15 participants per group were estimated to be needed based upon detection of a difference of 5 degrees of ROM at the 5% level of statistical significance with 80% power Analysis: per‐protocol analysed data reported (but intention‐to‐treat analysis was also performed, and results were similar to those of the per‐protocol analysis) Source of funding: National Science Council, Taiwan (non‐industry)
Participants	Number of participants: 30 (15 per group) Baseline characteristics Group receiving midrange mobilisation, then end range mobilisation, then midrange mobilisation, then mobilisation with movement Mean (SD) age = 53.3 (6.5) years; male:female = 1:13 Mean (SD) duration of symptoms: 18.8 (8) weeks Group receiving midrange mobilisation, then mobilisation with movement, then midrange mobilisation, then end range mobilisation Mean (SD) age = 58 (10.1) years; male:female = 3:11 Mean (SD) duration of symptoms: 22 (10) weeks Inclusion criteria Painful stiff shoulder for at least 3 months Limited range of motion of a shoulder joint (range of motion losses of 25% or greater compared with the non‐involved shoulder in at least 2 of the following shoulder motions: glenohumeral flexion, abduction or medial and lateral rotation) Consent of the participant's physician to participate in the study Exclusion criteria Diabetes mellitus History of surgery on the particular shoulder Rheumatoid arthritis Painful stiff shoulder after a severe trauma Fracture of the shoulder complex Rotator cuff rupture Tendon calcification
Interventions	Midrange mobilisation, then end range mobilisation, then midrange mobilisation, then mobilisation with movement (N = 15) Components of intervention End range mobilisation: At the start of each intervention session, the physical therapist examined the participant's range of motion to obtain information about the end range position and the end‐feel of the glenohumeral joint. Then, the therapist's hands were placed close to the glenohumeral joint, and the humerus was brought into a position of maximal range in different directions. 10‐15 repetitions of intensive mobilisation techniques, varying the plane of elevation or varying the degree of rotation in the end range position, were applied Midrange mobilisation: With the participant in a relaxed supine position, the humerus was moved to the resting position (40° of abduction). While the humerus was held in this position, 10‐15 repetitions of the mobilisation techniques were applied Mobilisation with movement: With the participant in a relaxed sitting position, a belt was placed around the head of the humerus to glide the humerus head appropriately, as the therapist's hand was used over the appropriate aspect of the head of the humerus. A counterpressure also was applied to the scapula with the therapist's other hand. The glide was sustained during slow active shoulder movements to the end of the pain‐free range and was released after return to the starting position. 3 sets of 10 repetitions were applied, with 1 minute between sets Dosage: 30 minutes Frequency of administration: Each treatment phase lasted twice a week for 3 weeks Provider: physical therapist Midrange mobilisation, then mobilisation with movement, then midrange mobilisation, then end range mobilisation (N = 15) See above (the only difference is the order in which each phase of mobilisation was provided)
Outcomes	Outcomes assessed at 3, 6, 9 and 12 weeks (i.e. end of each 3‐week treatment phase): No primary outcome was reported by the trialists Function using the Flexi‐Level Scale of Shoulder Function (FLEX‐SF), a self‐administered, shoulder‐specific, fixed‐item index consisting of 3 levels of function. In this scale, respondents answer a single item that grossly classifies their level of function as low, medium or high. They then respond only to the items that targeted their level of function. Scores are recorded from 1, indicating the most limited function, to 50, indicating the absence of limited function in the subject Active range of motion (arm elevation, scapular tipping, scapulohumeral rhythm, humeral lateral rotation, humeral medial rotation) using the FASTRAK motion analysis system
Notes	Unpublished outcome data provided by trialist on request Trial was registered at ClinicalTrials.gov (http://clinicaltrials.gov/show/NCT00172601)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Consenting subjects were randomly assigned by computer‐generated permuted block randomization of 5 by sequentially numbered, sealed, opaque envelopes to receive different mobilization treatments. In group 1, an A‐B‐A‐C (A=MRM,B=ERM, and C=MWM) multiple treatment design was used. In group 2, an A‐C‐A‐B multiple‐treatment design was used. The 2 groups used here were intended to counterbalance the order effects of treatments"  Comment: An adequate method was used to generate the allocation sequence
Allocation concealment (selection bias)	Low risk	Quote: "Consenting subjects were randomly assigned by computer‐generated permuted block randomization of 5 by sequentially numbered, sealed, opaque envelopes to receive different mobilization treatments. In group 1, an A‐B‐A‐C (A=MRM,B=ERM, and C=MWM) multiple treatment design was used. In group 2, an A‐C‐A‐B multiple‐treatment design was used. The 2 groups used here were intended to counterbalance the order effects of treatments"  Comment: An adequate method was used to conceal the allocation sequence
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Comment: Participants received slightly different types of mobilisation, but it is unclear whether they were provided any information that would make them perceive the type of mobilisation they received as superior or inferior to the alternative type of mobilisation
Blinding of outcome assessment (detection bias)   Self‐reported outcomes	Unclear risk	Comment: Participants self‐reported function, but it is unclear whether they were provided any information that would make them perceive the type of mobilisation they received as superior or inferior to the alternative type of mobilisation
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Quote: "To minimize bias, an independent trained outcome assessor, masked to treatment allocation, evaluated the participants at baseline and at 3‐week intervals for 12 weeks"  Comment: Assessors of shoulder complex kinematics were probably blind to treatment
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "Thirty subjects were recruited and randomly assigned to 2 groups (Tab. 1). Two subjects failed to attend the treatment. In addition, 3 subjects in the A‐B‐A‐C group were lost to follow‐up because there was no improvement during MRM treatment at 9 weeks. In the A‐C‐A‐B group, 2 subjects were lost to follow‐up because there was no improvement during MRM treatments at 3 weeks and 9 weeks"  Quote: "For the analysis, dropout data were excluded. Additionally, intention‐to‐treat analysis was performed by including the dropout data (carrying the last data point forward into analysis)"  Quote: "Similar results were found between exclusion of dropout data and intention‐to‐treat analysis (inclusion of dropout data)"  Quote: "No benefit was shown during MRM treatment, but different missing data due to subjects dropping out due to lack of improvement at 3 and 9 weeks between the 2 groups makes interpretation difficult. We addressed this by secondary analysis (i.e. analysis of dropping out between 2 groups and survival analysis). There were no differences in numbers of subjects dropping out and no significant differences in the survival experiences of the 2 groups. These findings suggest that the multiple treatment trial on our 2 groups was balanced. It may be, however, that subjects continued in the treatment for reasons other than treatment effectiveness"  Comment: The numbers of dropouts (and reasons for this) were similar between groups, and results of intention‐to‐treat and per‐protocol analysis were similar
Selective reporting (reporting bias)	High risk	Comment: The following outcomes were specified in the trial registry entry but were not reported in the publication: Disabilities of the Arm, Shoulder and Hand (DASH), disability assessment by self‐reports, pain perception and SF‐36 health survey
Other bias	Low risk	Comment: No other sources of bias were identified

Study	Reason for exclusion
Abdelshafi 2011	All 3 groups received the same manual therapy and exercise intervention with or without continuous suprascapular nerve block under ultrasound guidance or intra‐articular steroid injection
Alicicco 2000	Mobilisation was provided to all groups (with or without an electrotherapy modality)
Amir‐us‐Saqlain 2007	Both groups received the same manual therapy intervention (manipulation) with or without steroid injection under anaesthesia
Arslan 2001	Ineligible intervention: randomised controlled trial of glucocorticoid injection versus physical therapy plus non‐steroidal anti‐inflammatory drug. Not able to separate out the effects of physical therapy. Included in Cochrane review of corticosteroid injection for shoulder disorders
Bumin 2001	Exercises were provided to all groups (with or without an electrotherapy modality)
Calis 2006	Stretching and Codman exercises were provided to all groups (along with an electrotherapy modality)
Diercks 2004	Not a randomised controlled trial
Khan 2005	Participants received physical therapies alone versus physical therapies and shoulder arthrography with intra‐articular steroid
Kivimaki 2007	The same manual therapy intervention was delivered to both treatment groups
Koh 2013	Manual therapy and transcutaneous electrical nerve stimulation (TENS) were provided to all groups (with or without bee venom acupuncture)
Leclaire 1991	Manual stretching and pulley exercises were provided to all groups (with or without an electrotherapy modality)
Lee 1973	Exercises were provided to all groups (with or without an electrotherapy modality)
Leung 2008	Stretching exercises were provided to all groups (with or without an electrotherapy modality)
Longbottom 2009	Not a randomised controlled trial
Ma 2013	Passive joint mobilisation was provided to all groups (with or without cryotherapy)
Rizk 1991	The same manual therapy intervention was delivered to all 4 treatment groups
Srour 2008	Conventional physiotherapy comprising passive mobilisation, specific mobilisations, stretches and muscle tension raises was provided to all groups; this RCT primarily tested the effect of cold pack versus cold air ventilation versus no cold intervention
Zhu 2004	Ineligible intervention: trial comparing exercises plus Chinese medicine iontophoresis versus pain block therapy. A manual therapist/physical therapist/physiotherapist would be unable to deliver the Chinese medicine components

Methods	Not yet assessed
Participants	Not yet assessed
Interventions	Not yet assessed
Outcomes	Not yet assessed
Notes

PERMALINK

Manual therapy and exercise for adhesive capsulitis (frozen shoulder)

Matthew J Page

Sally Green

Sharon Kramer

Renea V Johnston

Brodwen McBain

Marisa Chau

Rachelle Buchbinder

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Summary of findings for the main comparison. Combination of manual therapy and exercise compared with glucocorticoid injection for adhesive capsulitis (frozen shoulder).

Summary of findings 2. Combination of manual therapy, exercise, electrotherapy and placebo injection compared with glucocorticoid injection for adhesive capsulitis (frozen shoulder).

Summary of findings 3. Combination of manual therapy, exercise, electrotherapy and placebo injection compared with placebo injection for adhesive capsulitis (frozen shoulder).

Summary of findings 4. Combination of manual therapy, exercise, electrotherapy and glucocorticoid injection compared with glucocorticoid injection for adhesive capsulitis (frozen shoulder).

Summary of findings 5. Combination of manual therapy and exercise following joint distension compared with sham ultrasound following joint distension for adhesive capsulitis (frozen shoulder).

Summary of findings 6. Combination of manual therapy, exercise, electrotherapy and oral NSAID compared with oral NSAID for adhesive capsulitis (frozen shoulder).

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Main outcomes

Other outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Summary of findings tables

Results

Description of studies

Results of the search

1.

Included studies

1. Characteristics of manual therapy and exercise interventions.

2. Outcome Reporting Bias In Trials (ORBIT) matrix.

Excluded studies

Risk of bias in included studies

2.

3.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Is the combination of manual therapy and exercise (with or without electrotherapy) more effective than placebo, no intervention or another active intervention (e.g. glucocorticoid injection, oral non‐steroidal anti‐inflammatory drug (NSAID), arthrographic joint distension)?

Manual therapy plus exercise versus glucocorticoid injection

3. van der Windt 1998: Passive mobilisation and supervised exercise (intervention) versus glucocorticoid injection (control).

Main outcomes

Other outcomes

Manual therapy plus exercise plus electrotherapy plus placebo injection versus glucocorticoid injection

Main outcomes

1.1. Analysis.

Trial name or title	Outcomes Following Dynamic Splinting and/or Physical Therapy for Patients With Adhesive Capsulitis
Methods	Parallel‐group, 2‐arm randomised controlled trial (United States)
Participants	Inclusion criteria Diagnosis of idiopathic adhesive capsulitis Greater than 50% limitation in active flexion, internal rotation or external rotation Tight GHJ capsule in at least 2 directions Capsular end‐feel primarily limiting ROM Age 40‐64 years Normal x‐rays
Interventions	Dynasplint Along with standard manual physical therapy, participants will use a stretching device (Dynasplint) in rehabilitation to regain ROM in stiff joints. Participants will use this device 20 to 30 minutes 2 times per day at home. Dynamic splinting utilises the protocols of low‐load prolonged stretch (LLPS) with calibrated adjustable tension to increase total end range time (TERT) to reduce contracture. The Dynasplint or "experimental" group will add this therapy to its standard of care regimen Physical therapy Participants in the physical therapy group will receive standard manual treatments during their usual physical therapy visits with no additional intervention
Outcomes	Primary outcome measures Number of physical therapy treatments required  Weeks of Dynasplint treatment  Secondary outcome measures Shoulder range of motion  Sharp FAS neck and shoulder  Disabilities of Arm, Hand and Shoulder Questionnaire
Starting date	January 2006
Contact information	Paul Gaspar, DPT, gasparPT@hotmail.com
Notes	http://clinicaltrials.gov/show/NCT00873158

Trial name or title	The Study of Different Treatment Programs for Patients With Frozen Shoulder
Methods	Parallel‐group, 2‐arm randomised controlled trial (Taiwan)
Participants	Inclusion criteria Patients with shoulder pain and range of motion limitation Exclusion criterion Rotator tear
Interventions	Group 1: Physical modalities + therapeutic exercise Group 2: Physical modalities + therapeutic exercise + intra‐articular injection of steroid Group 3: Therapeutic exercise
Outcomes	Primary outcome measures Pain scale and range of motion within 6 weeks
Starting date	December 2010
Contact information	Sui‐Foon Lo, MD, d4659@www.cmuh.org.tw
Notes	http://clinicaltrials.gov/show/NCT01249040

Trial name or title	AA4500 for the Treatment of Adhesive Capsulitis of the Shoulder
Methods	Parallel group, 2‐arm randomised controlled trial (United States)
Participants	Inclusion criteria Male or female 18 years of age or older Female of childbearing potential, with negative urine pregnancy test and using an effective contraception method (i.e. abstinence, intrauterine device (IUD), hormonal (oestrogen/progestin) contraceptives, or barrier control) for at least 1 menstrual cycle before study enrolment and for 1 menstrual cycle following end of study, or surgically sterile Unilateral idiopathic adhesive capsulitis of 1 shoulder for at least 3 months but not longer than 12 months before the screening visit, with stage 2 (frozen or adhesive stage) disease, as determined by investigator Normal range of motion in the contralateral shoulder, as determined by the investigator Restricted active range of motion (AROM) in the affected shoulder, defined as a deficit of at least 60 degrees in total AROM in the affected shoulder as compared with total AROM in the contralateral shoulder, and a deficit of at least 30 degrees in AROM in at least 1 of the following planes as compared with the contralateral shoulder: forward flexion, abduction, external rotation with elbow up to 90 degrees of abduction, internal rotation with elbow up to 90 degrees of abduction Voluntary signing and dating of an informed consent agreement approved by the institutional review board/independent ethics committee (IRB/IEC). Participant must also sign an authorisation form to allow disclosure of protected health information (PHI). The PHI authorisation form and the informed consent form may be an provided as an integrated form or as separate forms, depending on the institution Completion and understanding of the various rating instruments available in English
Interventions	Group 1: Collagenase Clostridium histolyticum(AA4500) Group 2: Shoulder exercises
Outcomes	Primary outcome measure Change (degrees) from baseline in active forward flexion in affected shoulder  Secondary outcome measures Change (degrees) from baseline in passive forward flexion in affected shoulder  Change (degrees) from baseline in abduction (active and passive)  Change (degrees) from baseline in external rotation at 90° abduction (active and passive)  Change from baseline in internal rotation with the elbow at 90° abduction (active and passive)
Starting date	November 2011
Contact information	Diane McCaul, dmccaul@auxilium.com
Notes	http://clinicaltrials.gov/show/NCT01483963