A systematic review of the methods and drugs used for performing suprascapular nerve block injections for the non-surgical management of chronic shoulder pain

Abstract

Suprascapular nerve block (SSNB) injections are growing in popularity as a treatment option for people with chronic shoulder pain. The optimal method of injection and aftercare is unknown. This review describes the current methods and drugs used for performing SSNB injections in the non-surgical management of adults with chronic shoulder pain in order to inform future research in this area. Systematic searches of CINAHL, MEDLINE (OVID), AMED, Embase databases and the Cochrane Library were undertaken from inception to June 2020. Data on the method and drugs used for injection and aftercare were extracted and summarised for areas of commonality and discrepancy. We included 53 studies in this review. In total, eight different injection methods were reported within the included studies. Indirect surface land-marked methods were the most common method reported in 21 studies. Direct surface land-marked methods were reported in 12 studies. Ultrasound-guided methods used alone were reported in 16 studies. Both fluoroscopy and computed tomography methods used alone were reported in one study each. Electromyography was used in combination with other injection methods in nine studies. Wide variation in the composition of the injectate was observed between studies. Local anaesthetic was used within injectate preparations in all studies. Local anaesthetic used alone was reported in 20 studies, combined with steroid in 29 studies and combined with various other components in 5 studies. Physiotherapy following injection was reported in 26 studies. Reported details of physiotherapy varied considerably. This review identified substantial variation in the methods and drugs used to perform SSNB injection in clinical trials. Current literature demonstrates a wide range of methods used for SSNB injection administration. Consensus research defining standardised practice for SSNB injection is now needed to guide future clinical practice and research.

Introduction

Chronic shoulder pain is a major cause of disability. ¹ Its prevalence is between 2.4% and 14%. Around 20% of those affected report shoulder pain at 1 year and 14% of people report shoulder pain at 3 years.^2,3 Annually, approximately 1.5 million general practitioner (GP) consultations are related to shoulder pain in England alone with the total annual societal costs of shoulder pain in the United Kingdom estimated to be £100m. ⁴

Shoulder pain can arise from multiple structures, making a clinical diagnosis challenging. ⁵ The best treatment for people with persistent shoulder pain is uncertain. Physiotherapy is often the first-line treatment for shoulder pain. Local steroid injections can provide short-term benefit for up to 3 months ⁶ and are used for the management of shoulder pain within physiotherapy practice.^7–10 At best, local steroid injections have short-term benefits. ⁶ Repeat local steroid injections may have negative effects on tendon homeostasis, healing and future treatment outcomes for some patients.^11–14 Suprascapular nerve block (SSNB) injections may offer an effective alternative to local steroid injections for some patients, and there is growing interest in their use including within physiotherapy practice.^15,16

The suprascapular nerve is a mixed motor and sensory nerve that supplies approximately 70% of the sensory innervation to the shoulder complex.^17,18 Injections aiming to block the suprascapular nerve are used in the management of a range of chronic shoulder conditions.^15,19 The mechanisms contributing to the clinical effectiveness of SSNB injections are unclear. ²⁰ SSNB injections are recommended in the British Elbow and Shoulder Society (BESS) guidelines for the management of chronic subacromial shoulder pain ²¹ and glenohumeral osteoarthritis. ²²

A 2018 survey of BESS medical consultant and registrar members found that around 40% of respondents would use SSNB injection in the management of patients with large and massive rotator cuff tears. ¹⁶ A 2019 survey of UK physiotherapists who regularly treat patients with shoulder pain found that the most common shoulder condition that they would consider using SSNB injections for was rotator cuff tear arthropathy (glenohumeral joint osteoarthritis associated with cuff tears). ¹⁵ Some authors suggest that people with large to massive rotator cuff tears may also have an element of suprascapular neuropathy that may be more responsive to an SSNB injection. ²³

There is limited research surrounding the effectiveness of SSNB injections in these patient groups and compared to local steroid injections.^19,24 Multiple methods and drugs are used for SSNB injections. ²⁵ The optimum method and aftercare has not been defined. Defining best practice in SSNB injections, and the adjunct physiotherapy interventions used are first steps towards designing robust randomised controlled trials (RCTs). The aim of this review was to identify the methods and drugs used for performing SSNB injections for people with chronic shoulder pain and whether physiotherapy interventions were used as an adjunct treatment following the injection.

Methods

Search methods and strategy

Electronic searches of CINAHL, MEDLINE (OVID), AMED, Embase and the Cochrane Library were conducted from inception to June 2020. The electronic search was supplemented by hand searching of reference lists of retrieved full-text articles and citation searching. The search strategy and search terms were developed with the support of a medical librarian (S.J.). Search strategies for all databases are available in Supplemental Appendix 1.

Study selection

Screening for relevant articles by title and abstract was undertaken independently by the lead author (N.S.). Full-text articles were then assessed for eligibility using the predefined inclusion and exclusion criteria (Box 1) by two researchers independently (N.S. and Z.L.). Where there were discrepancies in decisions made between the two reviewers, these articles were forwarded to a third reviewer (R.K.) who made a final decision. Full-text articles that met the inclusion criteria were included in this review.

Box 1.

Inclusion and exclusion criteria.

Inclusion criteria
Types of participants	Adults (⩾18 years) with shoulder pain managed non-surgically including frozen shoulder, subacromial pain, rotator cuff disorders, osteoarthritis, rheumatoid arthritis, non-specific shoulder pain, metastatic shoulder pain and hemiplegic shoulder pain/post-stroke shoulder pain.
Types of studies	Randomised controlled trials (RCTs), observational studies, cohort studies, audits, surveys and case series with five or more participants.
Types of intervention	All studies that report on the detail of the method used for SSNB injection, that is, land-marked or guided and the drugs used for performing SSNB injection. SSNB injection could be delivered in the treatment or controlled arms of studies.
Types of outcome measures	Not applicable.
Types of publication	Full-text peer-reviewed articles. Books and guidance manuals that are referenced within the retrieved peer-reviewed articles.
Exclusion criteria
Types of participant	Cadaveric studies, normal volunteers, participants undergoing shoulder surgery and/or thoracic surgery, patients with acute shoulder pain, that is, undergoing shoulder reduction following dislocation.
Types of studies	No specific study type excluded.
Types of intervention	Studies that only investigate SSNB denervation, pulsed radiofrequency, nerve ablation and continuous SSNB delivered by catheter. Studies that do not report on the detail of the method of SSNB injection, that is, land-marked, guided and the drugs used were excluded.
Types of outcome measures	Not applicable.
Types of publication	Conference abstracts, books or guidance manuals that are not referenced within the retrieved peer-reviewed articles were excluded.

SSNB: suprascapular nerve block.

Data extraction and synthesis

Data on the SSNB injection method used, the components making up the injectate and whether physiotherapy was delivered following the injection were extracted into predefined data extraction tables. A narrative synthesis of the findings is presented.

Results

Study selection

Our search returned 3552 results from which we removed 980 duplicates. We screened a total of 2572 records by title and abstract. We retrieved and assessed 89 full-text articles for eligibility of which 53 were included. Articles excluded were recorded at the different stages of the eligibility screening process and reported in a PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) flow diagram (Figure 1).

Figure 1.

Preferred reporting items for systematic reviews and meta-analysis (PRISMA) flow chart.

Study characteristics

We included 53 studies from 14 different countries in this review (Table 1). Of which, 13 studies came from Turkey,^26–38 7 each from Australia^{20,23,39,38,39,40–43} and India,^44–50 6 from the United Kingdom,^51–56 4 each from the United States^57–60 and South Korea,^61–64 2 each from Brazil,^65,66 Canada,^67,68 Egypt^69,70 and Italy,^71,72 and 1 each from Colombia, ⁷³ Austria, ⁷⁴ Nepal ⁷⁵ and Thailand. ⁷⁶

Table 1.

Summary of studies.

Author	Year	Country	Design	Disorder	Method
Shanahan	2020	Australia	Observational	Motor neurone disease	Land-marked indirect
Malheiro	2020	Brazil	Observational	Mixed MSK	US guided
Kasapoğlu-Aksoy	2020	Turkey	RCT	Stroke	US guided
Terlemez	2020	Turkey	RCT	Stroke	US guided
Yilmaz	2020	Turkey	RCT	Impingement syndrome	Land-marked direct
Verma	2019	India	RCT	Frozen shoulder	US guided
Jung	2019	South Korea	Observational	Frozen shoulder	US guided
Aydin	2019	Turkey	Observational	Stroke	US guided
Coory	2019	Australia	RCT	Rotator cuff tears	US guided
Sencan	2019	Turkey	RCT	Stroke	EMG fluoroscopy
Kamal	2018	India	RCT	Mixed MSK and neuro	Land-marked indirect/US guided
Salt	2018	United Kingdom	Service evaluation	Mixed MSK	Land-marked direct/US guided
Fernandes	2017	Brazil	Observational	Frozen shoulder	Land-marked indirect
Kammili	2017	India	RCT	Frozen shoulder	EMG land-marked
Külcü	2017	Turkey	RCT	Stroke	Land-marked direct/EMG land-marked
Mortada	2017	Egypt	RCT	Frozen shoulder	US guided
Okur	2017	Turkey	Observational	Post-mastectomy shoulder pain	US guided
Picelli	2017	Italy	Observational	Stroke	US guided
Bradnam	2016	Australia	Observational	Rotator cuff disease	Land-marked indirect
Mitra	2016	India	RCT	Frozen shoulder	US guided
Sonune	2016	India	RCT	Frozen shoulder	US guided
Chansoria	2015	India	Observational	Non-specific shoulder pain	Land-marked direct SGN
Dorn	2015	Austria	Observational	Mixed MSK	Land-marked indirect
Bayram	2014	Turkey	RCT	Impingement syndrome	Land-marked indirect
El-Badawy	2014	Egypt	Observational	Frozen shoulder	Land-marked indirect
Jeon	2014	South Korea	Observational	Stroke	US guided
Salgia	2014	India	RCT	Mixed MSK	Land-marked indirect
Adey-Wakeling	2013	Australia	RCT	Stroke	Land-marked indirect
Arcila Lotero	2013	Colombia	Observational	Mixed MSK and neuro	EMG US
Gofeld	2013	Canada	RCT	Mixed MSK	EMG fluoroscopy
Kang	2012	South Korea	Observational	Mixed MSK	Fluoroscopy
Ozkan	2012	Turkey	Observational	Frozen shoulder	EMG fluoroscopy
Shanahan	2012	Australia	Audit	Mixed MSK and neuro	Land-marked indirect
Taskaynatan	2012	Turkey	Observational	Mixed MSK	EMG US
Yasar	2011	Turkey	RCT	Stroke	Land-marked direct
Gorthi	2010	South Korea	RCT	Non-specific shoulder pain	US guided/land-marked direct
Boonsong	2009	Thailand	RCT	Stroke	Land-marked indirect
Khan	2009	Nepal	Observational	Frozen shoulder	Land-marked indirect
Mitra	2009	United States	Observational	Frozen shoulder	Land-marked indirect
Di Lorenzo	2006	Italy	RCT	Rotator cuff tendinitis	Land-marked indirect
Taskaynatan	2005	Turkey	RCT	Non-specific shoulder pain	Land-marked direct
Shanahan	2004	Australia	RCT	Mixed MSK	Land-marked indirect/CT
Shanahan	2003	Australia	RCT	Mixed MSK	Land-marked indirect
Karataş	2002	Turkey	RCT	Frozen shoulder	Land-marked direct/EMG land-marked
Dahan	2000	Canada	RCT	Frozen shoulder	Land-marked indirect
Jones	1999	United Kingdom	RCT	Frozen shoulder	Land-marked indirect
Lewis	1999	United Kingdom	Observational	Rheumatoid/osteoarthritis	Land-marked indirect
Dangoisse	1994	United Kingdom	Observational	Mixed MSK and neuro	Land-marked indirect
Gado	1993	United Kingdom	RCT	Rheumatoid	Land-marked – indirect and direct
Vecchio	1993	United Kingdom	RCT	Rotator cuff tendinitis and tears	Land-marked direct
Wassef	1992	United States	Observational	Frozen shoulder	EMG land-marked
Long	1987	United States	Audit	Myofascial trigger point pain	Land-marked direct
Rowlingson	1986	United States	Audit	Mixed MSK	Land-marked direct

MSK: Musculoskeletal; RCT: randomised controlled trial; EMG: electromyography; CT: computed tomography; US: ultrasound; SGN: spinoglenoid notch.

We included 28 RCTs,^{23,26,27,29–31,33,35,36,38,41–44,46,48–52,55,57,61,67,68,70,72,76} 21 observational studies,^{20,28,32,34,37,39,45,53,54,57,58,62–66,69,71,73–75} 3 audits^40,59,60 and 1 service evaluation. ⁵⁶ Participant numbers were generally small within the included RCTs ranging from 10 participants ⁷⁶ to 100 participants.^47,49 Of the included observational studies, again participant numbers were generally small ranging from 9 participants ⁵⁸ to 102 participants. ⁶⁴ The largest number of SSNB injections reported within a single study was from a retrospective audit of patient records that included 1005 SSNB injections undertaken within a Community Rheumatology service in Australia. ⁴⁰

Population and shoulder conditions

A wide range of shoulder disorders were included in this review. Of the 53 studies included, 15 were undertaken with participants with a diagnosis of frozen shoulder,^{28,30,46–49,55,57,58,64,65,68–70,75} 10 with participants with post-stroke shoulder pain^{26,27,29,32,33,35,43,63,71,76} and 10 with participants with a range of different musculoskeletal shoulder disorders.^{37,41,42,44,56,60,62,66,67,74} Six studies were undertaken with participants with rotator cuff–related shoulder pain, of which two were with participants with a diagnosis of impingement syndrome^31,36 and one each with participants with rotator cuff tears, ²³ rotator cuff disease, ²⁰ rotator cuff tendinitis and/or tears ⁵¹ and rotator cuff tendinitis. ⁷² Four studies were undertaken with participants who had either a musculoskeletal shoulder disorder or a neurological condition causing their shoulder pain.^40,50,53,73 Three studies were undertaken with participants with non-specific shoulder pain, where no diagnosis was provided.^38,45,61 One study was undertaken in participants with rheumatoid arthritis, ⁵² either rheumatoid arthritis and/or osteoarthritis, ⁵⁴ myofascial trigger point shoulder pain, ⁵⁹ motor neurone disease ³⁹ and with participants who developed shoulder pain following mastectomy for breast cancer. ³⁴

Injection methods

A variety of different surface land-marked and guided SSNB injection methods were reported within the included studies (Figure 2).

Figure 2.

Summary of injection methods used in studies included. CT: Computed Tomography; US: Ultrasound; SGN: Spinoglenoid notch; EMG: Electromyography.

Surface land-marked methods without any use of guidance were used in 32 studies. Of these, the indirect surface land-marked approach was reported in 20 studies^{20,36,39–44,50,53–55,57,65,68,69,72,74–76} and the direct suprascapular notch surface land-marked approach was reported in 10 studies.^{27,29,30,35,38,51,56,59–61} In one study, the direct spinoglenoid notch method was used ⁴⁵ and in one study, a combination of both the direct suprascapular notch and indirect method was used. ⁵² Ultrasound guidance used alone was reported in 16 studies.^{23,26,27,32,34,46,47,49,50,56,61,63,64,66,70,71} Fluoroscopy used alone and computed tomography (CT) used alone was each reported in one study.^62,41 The additional use of electromyography (EMG) was reported in nine studies, of which four combined EMG with a surface-landmarked method,^30,35,48,58 three combined EMG with fluoroscopy^28,33,67 and two combined EMG with ultrasound.^37,73 From the studies included in our review, the earliest injection method reported was the land-marked direct suprascapular notch method in 1986. ⁶⁰ The land-marked indirect method was first reported in 1994. ⁵³ The earliest guided SSNB injection methods reported were EMG guided in 1992, ⁵⁸ CT guided in 2004, ⁴¹ ultrasound guided in 2010 ⁶¹ and fluoroscopy guided in 2012. ⁶² The most commonly reported injection method within the last decade was ultrasound-guided SSNB injections in 16 studies (Figure 3).

Figure 3.

Trend in method used for SSNB injection since 2010. SGN: Spinoglenoid notch; EMG: Electromyography; US: Ultrasound.

Of the 53 studies included in this review, six studies compared different injection methods. One study compared ultrasound-guided injection with the land-marked indirect method. ⁵⁰ Two studies compared ultrasound-guided injection with the land-marked direct method.^56,61 Two studies compared EMG land-marked with the land-marked direct method^30,35 and one study compared CT-guided injection to the land-marked indirect method. ⁴¹

Medication used

In total, 21 different combinations of medication were used within the injectates for SSNB injection for the studies included in this review with bupivacaine combined with methylprednisolone being the most common in 12 studies (Figure 4).

Figure 4.

Summary of medications used within studies included.

Local anaesthetic was used in all 53 studies. Local anaesthetics were used alone in 20 studies,^{27,29,58–60,62,63,67,68,70,74,76,30,33,37,38,46,49,54,57} used with steroid in 29 studies^{20,23,26,28,32,34–36,39–45,47,48,50–53,55,56,64,66,69,71,73,75} and used with a combination of steroid and other injectate components such as saline,^27,72 serum, ³² adrenaline ⁶⁵ and dextrose ⁶¹ in a further 5 studies. No study used steroid alone (Figure 5).

Figure 5.

Summary of injectates.

Bupivacaine was the most common local anaesthetic used at various concentrations and volumes in 30 studies.^{20,31,46,47,49–53,55,58,59,33,60,64–66,68–71,73,75,34,39–44} Of which, 24 studies reported using 0.5% bupivacaine^{20,31,46,49,51–53,55,64,65,68,69,33,70,71,73,75,34,39–44} and 6 studies reported using 0.25% bupivacaine.^{47,50,58–60,66} Bupivacaine was used at different volumes ranging from 10 mL of 0.5% bupivacaine,^{20,39–44,55,64,68,70,71} 3 mL of 0.25% bupivacaine ⁶⁶ to 1 mL of 0.5% bupivacaine. ⁵¹

Lidocaine was the choice of local anaesthetic used in 16 studies.^{26,27,61,63,67,72,75,76,28,30,32,38,45,48,56,57} Of which, 10 studies reported using 1% lidocaine,^{28,30,38,45,56,57,63,67,75,76} 4 studies used 2% lidocaine^26,27,61,72 and 2 studies used 10% lidocaine.^32,48 Lidocaine was used at various volumes ranging from 2 mL of 10%, ³² 2 mL of 2% ⁶¹ and 2 mL of 1% ⁶⁷ lidocaine to 10 mL of 1% lidocaine.^{30,38,45,56,76} A combination of lidocaine (5 mL/1%) mixed with bupivacaine (4 mL/0.5%) was used in one study. ⁷⁵

Prilocaine was used in five studies.^{29,35–37,54} Of which, prilocaine 2% was used in three studies^29,36,37 and prilocaine 1% used in two studies.^35,54 The volume of prilocaine ranged from 10 mL of 2% prilocaine ²⁹ to 4 mL of 1% prilocaine. ⁵⁴

Ropivacaine was used in two studies.^23,74 Of which, one study used 1 mL of 1% ropivacaine ²³ and one study used 5 mL of 0.5% ropivacaine. ⁷⁴ Mepivacaine (2 mL/1%) was used in one study. ⁶²

The lowest volume of local anaesthetic used was 1 mL of 1% ropivacaine combined with 1 mL of betamethasone delivered by ultrasound. ²³ The largest volume of local anaesthetic used was 10 mL of 0.5% bupivacaine,^{20,39–44,55,64,68,70,71} 10 mL of 1% lidocaine,^{30,38,45,56,76} 10 mL of 2% lidocaine ⁷² and 10 mL of 2% prilocaine ²⁹ by both land-marked and ultrasound-guided techniques. The largest dosage of local anaesthetics reported was 400 mg of lidocaine (4 mL/10%) injected using EMG guidance, ⁴⁸ 200 mg of lidocaine (2 mL/10%) injected using ultrasound guidance ³² and 200 mg of lidocaine (10 mL/2%) delivered land-marked. ⁷² Conversely, the smallest dosage of local anaesthetic used was 5 mg of bupivacaine (1 mL/0.5%) injected using the surface land-marked direct suprascapular notch method. ⁵¹ The mean dosage of bupivacaine used was 36.25 mg for land-marked and 28.18 mg for guided methods. The mean dosage of lidocaine used was 107 mg for land-marked and 130 mg for guided methods (Figure 6).

Figure 6.

Mean dosage of local anaesthetic used in mg for guided and landmarked Bupivacaine and Lidocaine medications.

Methylprednisolone was the most common steroid used in combination with local anaesthetic in 16 studies.^{20,28,39–45,47,50,51,53,56,71,75} Of which, 40 mg was used in 15 studies^{20,28,39–45,47,50,51,56,71,75} and 80 mg was used in 1 study. ⁵³ Triamcinolone was used in seven studies.^{26,34,36,48,55,64,66} Of which, 40 mg was used in four studies^26,36,48,66 and 20 mg in three studies.^34,55,64 Betamethasone at a volume of 1 mL was used in four studies.^23,27,32,35 Dexamethasone was used in two studies.^69,73 Of which, one study reported using 1 mL of dexamethasone ⁶⁹ and the other study reported using 4 mg of dexamethasone. ⁷³ Prednisolone 40 mg was used in one study. ⁵² Only one study compared the effectiveness of SSNB injections given with or without steroid. ⁶⁶

Additional treatment following SSNB injections

The use of physiotherapy as an adjunct intervention following SSNB injection was reported in 26 studies (Table 2).^{23,26–28,31–34,36,43,44,46,48,49,55–58,63,64,68–70,72,75,76} Most physiotherapy interventions included some aspect of shoulder exercises, either physiotherapy led or in the form of home exercises with written and verbal advice after SSNB injection. The specific details of the reported physiotherapy intervention and exercises varied considerably between studies. No study evaluated if physiotherapy maximised any potential benefit or improved long-term outcomes in participants after SSNB injection.

Table 2.

Summary of physiotherapy interventions reported following SSNB injection.

Author (year)	Country	Disorder	Physiotherapy
Kasapoğlu-Aksoy (2020)	Turkey	Stroke	Hemiplegic shoulder rehabilitation passive and active–assisted ROM, stretching, neurophysiologic exercises performed by same experienced physiotherapist.
Terlemez (2020)	Turkey	Stroke	Routine home exercise programme ROM and stretching.
Yilmaz (2020)	Turkey	Impingement syndrome	Codman exercises, ROM and strengthening HEP.
Verma (2019)	India	Frozen shoulder	HEP, pendulum, crossover arm stretch, passive internal and external rotation, sleeper stretch.
Jung (2019)	South Korea	Frozen shoulder	Shoulder rehabilitation programme, active and passive ROM ex’s, strengthening rotator cuff, scapular stabilising in outpatients on day of first intervention.
Aydin (2019)	Turkey	Stroke	HEP. Passive and active ROM exercises. Three sets daily. 20 in each set.
Coory (2019)	Australia	Rotator cuff tears	Physiotherapy exercises Murdoch protocol.
Sencan (2019)	Turkey	Stroke	Physiotherapy exercises 24 hours after injection. 10 sessions over 2 weeks. Active and passive ROM, neurophysiological exercises guided by a physiotherapist.
Salt (2018)	United Kingdom	Mixed MSK	Individualised HEP.
Kammili (2017)	India	Frozen shoulder	Mobility ex’s six times a day. Assisted flexion, abduction, circumduction, rotations slightly within painful range. Control group only – ex’s mobility ex’s six times a day. Assisted flexion, abduction, circumduction, rotations slightly within painful range.
Mortada (2017)	Egypt	Frozen shoulder	Physiotherapy led exercises programme, joint stretching and mobilisation, rotator cuff strengthening under supervision of physiotherapist three times a week for 3 weeks.
Okur (2017)	Turkey	Post-mastectomy shoulder pain	Prescribed shoulder home exercises by physiatrist.
Sonune (2016)	India	Frozen shoulder	Advised on regular therapeutic exercises.
Bayram (2014)	Turkey	Impingement syndrome	HEP. Passive and active ROM exercises. Three sets daily. 20 in each set.
El-Badawy (2014)	Egypt	Frozen shoulder	Codman exercises started 15 minutes after SSNB. Home exercises 10 reps three times a day, 12 weeks.
Jeon (2014)	South Korea	Stroke	Continuous neurodevelopmental therapy.
Salgia (2014)	India	Mixed MSK	Static and dynamic physiotherapy exercises as an outpatient.
Adey-Wakeling (2013)	Australia	Stroke	Routine therapy and standard ward care (Allen 2010 protocol).
Ozkan (2012)	Turkey	Frozen shoulder	Verbal and written instructions on shoulder HEP.
Boonsong (2009)	Thailand	Stroke	Passive ROM stretching exercises, occupational physiotherapist, ambulation training, heat pad.
Khan (2009)	Nepal	Frozen shoulder	Shoulder manipulation and home exercises.
Mitra (2009)	United States	Frozen shoulder	Manipulation then physiotherapy sessions × 8.
Di Lorenzo (2006)	Italy	Rotator cuff tendinitis	Physiotherapy.
Dahan (2000)	Canada	Frozen shoulder	Shoulder exercises to complete at home.
Jones (1999)	United Kingdom	Frozen shoulder	Verbal and written instructions on shoulder exercises at home self-mobilisation, stretching and rotator cuff strengthening.
Wassef (1992)	United States	Frozen shoulder	Physiotherapy programme after SSNB.

ROM: range of movement; HEP: home exercise program, reps: repetitions; SSNB: suprascapular nerve block.

Discussion

The aim of this systematic review was to identify the methods and drugs used for performing SSNB injections in the management of chronic shoulder pain and what adjunct physiotherapy interventions are used following injection. We have not included studies of modulation techniques targeting the suprascapular nerve. This is an important piece of work but is beyond the scope of this review which we have done to inform future research on injections used for SSNBs. Similarly, we are not seeking to draw any conclusions here on the clinical effectiveness of SSNBs, which is outside the scope of this review.

In our systematic review, we found that a variety of injection methods and medications can be used for SSNB injections. The direct surface land-marked method first described by Wertheim and Rovenstine ⁷⁷ was reported in 12 studies in our review. This direct approach aims to deliver the injectate directly into the suprascapular notch where the suprascapular nerve enters the scapular. The direct approach is reported to pose a greater risk of pneumothorax, nerve damage and veno-puncture, due to potential needle tip injury to the suprascapular nerve, suprascapular blood vessels and the apex of the lung if the needle tip penetrates too deeply through the suprascapular notch, compared to the indirect approach. ⁴⁰

Dangoisse et al. ⁵³ demonstrated that it was not essential to use the direct suprascapular notch approach for SSNB injection. An indirect approach where a deposit of 5–10 mL of injectate delivered deep into the supraspinatus fossa, away from the suprascapular notch, was sufficient to fill the supraspinatus fossa and bathe the suprascapular nerve, thereby reducing the risks associated with the direct approach. ⁵³ The indirect approach was a common method, reported in 20 studies in our review. The indirect approach appears to be a safe and well-tolerated method of SSNB injection. A retrospective case audit of over 1005 SSNB injections, administered utilising the indirect approach, revealed no serious complications with this method. ⁴⁰

In total, 16 studies utilised ultrasound guidance for SSNB injections in our review. Ultrasound-guided approaches were the most common method for SSNB injection within our review for studies published within the last decade, presumably due to technological advances, training opportunities and equipment availability. The earliest study in our review using ultrasound guidance for SSNB injection was in 2010. ⁶¹ Ultrasound guidance removes the radiation risks associated with other guided approaches such has fluoroscopy and CT-guided injections which appear to have become less common over time. Although the use of ultrasound guidance may be needed for accurate needle placement for denervation procedures at the suprascapular nerve, at present there is insufficient evidence regarding the efficacy of ultrasound-guided SSNB injections compared to surface land-marked SSNB injection approaches. ²⁴

Some authors suggest that using guided nerve block methods facilitates the use of lower drug dosages to achieve effective blockade; however, this was not always consistent with the trends and patterns of practice reported within studies in our review. In fact, the largest dosage of local anaesthetic used was 400 mg of lidocaine (4 mL/10%) delivered utilising EMG ⁴⁸ and the smallest dosage of local anaesthetic used was 5 mg of bupivacaine (1 mL/0.5%) injected using the surface land-marked direct suprascapular notch method. ⁵¹

Within the 53 studies in this review, 21 different variations of drugs were used for SSNB injections, with further variation in the volumes and concentrations of drugs. Local anaesthetic was used alone in 20 studies, combined with steroid in 29 studies and combined with various other components in five studies. Robust evidence supporting the addition of corticosteroid to local anaesthetic in the hope it prolongs the duration of nerve blockade or the effectiveness of nerve block injection is lacking. ⁷⁸ Only one small study undertaken with 34 stroke patients compared the effectiveness of SSNB injection with and without the addition of steroid. ⁶⁶ Although a trend in favour of the addition of steroid was seen no statistical differences were observed in pain scores post-injection compared to placebo injection and SSNB given with local anaesthetic alone. At present, it remains unclear if the addition of corticosteroid to local anaesthetic for SSNB injection improves outcomes.

This review indicates that injection method, choice of drug, drug dosage, volume and concentration used for SSNB injection may be based upon clinician preference rather than evidence based. Although physiotherapy interventions were commonly employed following SSNB injection, in 26 of the studies included in this review, there was wide variation in the interventions and methods of delivery. Many of the studies included in our review had heterogeneous cohorts of patients with mixed shoulder conditions. Although patients with large to massive rotator cuff tears and patients with rotator cuff arthropathy appear to be common patients groups selected to receive SSNB injection in clinical practice in the United Kingdom,^15,16 interestingly our review found no clinical trials specifically focusing on these patient groups to support this trend.

Conclusion

This review identified substantial variation in the methods and drugs used to perform SSNB injection in clinical trials. Standardisation of the delivery method and drug used for SSNB injection is required in order to evaluate its effectiveness in a multi-centre clinical trial. The current literature demonstrates a large range of SSNB injection administration. Consensus research on defining standardised practice in SSNB injection is therefore now needed to guide future clinical and research practice. This will ensure delivery of the intervention is standardised, can be delivered consistently across different trial sites and is reproducible in clinical practice.