Rheumatoid and osteoarthritis of the glenohumeral joint may cause severe disability as a result of pain and limitation of movement. Suprascapular nerve block has been used successfully to alleviate such pain and improve movement [1,2] on the basis that this nerve provides sensory fibres to the anterior and posterosuperior regions of the shoulder joint . However, Buchanan  described an articular branch to the shoulder joint that arises from the circumflex nerve inferior to the glenoid and provides a sensory supply to the anterior and inferior areas of the joint capsule . Therefore, on anatomical grounds, it seems possible that neural blockade of the suprascapular nerve (SSNB) combined with a block of the articular branches of the circumflex nerve (ACNB) may be required to relieve severe pain and reduced mobility of the shoulder joint by producing a widespread sensory blockade of the glenohumeral joint. Clinical impressions gained by the author after carrying out combined SSNB and ACNB supported this hypothesis, which was investigated further in the present small clinical study. A search of the literature failed to reveal any previous reports of combined SSNB and ACNB having been used for this purpose.
Patients and methods
Sixteen patients referred to the author at the Blackpool Pain Relief Clinic were studied. Thirteen were female and three male; ages ranged between 53 and 86 years; mean 69 years (10.5). Thirteen had been diagnosed on clinical findings as suffering from rheumatoid arthritis and three from osteoarthritis of the shoulder joint. Clinical diagnosis was supported by radiological evidence, which was available in 10 cases investigated.
The intensity of pain was assessed by using a visual analogue scale (VASP 0-10 cm), where 0=no pain and 10=worst pain ever. Movements of the shoulder joint were determined by measuring abduction, adduction and flexion using a Baseline 360 Goniometer (Nottingham Rehab, UK) or by clinical estimation. Functional external and internal rotation of the shoulder joint were assessed by the ability of the patient to touch the back of the neck (TBN) and middle of the back (TMB), respectively, with the ipsilateral hand. Observations were made immediately before combined SSNB and ACNB, 20-40 min afterwards and again at out patient review.
After having obtained written informed consent for the study, which has been approved by the Institution Ethics Committee, i.v. access was established using a No. 16 SG Venflon. Patients were positioned at the end of the theatre table with arms folded across the chest in order to make bony landmarks of the scapula more prominent. Eight patients who were on long-term steroid therapy were given hydrocortisone 100 mg i.v. and one who was unduly anxious 2 mg midazolam i.v. before neural blockade was carried out.
This was based on the method described by Gordh  using an Abbott spinal pack and 22 SG spinal needle (Abbott Laboratories, Kent, UK). However, when anatomical deformities were present, the needle was inserted close to the spine of the scapula until the bony floor of the supraspinous fossa was reached. This procedure was adopted to reduce the risk of the needle passing superior and deep to the scapula causing pneumothorax.
Following the infiltration of surrounding soft tissue with 5 mL of 1% prilocaine, a 22 SG spinal needle was introduced about 4 cm medial to the lateral border of the acromium to impinge on the spine of the scapula (Fig. 1) and then redirected antero-inferiorly to the posterior margin of the inferior one-third of the glenoid (Fig. 2).
Four millilitres of 1% prilocaine were injected slowly into the supraspinous notch and also the post-glenoid area after aspiration tests for blood, air and synovial fluid were found to be negative. Following confirmation of pain relief by the patient (diagnostic nerve block), 4 mL of 6% aqueous phenol were injected slowly at each of these sites.
Radiographic control was used to facilitate neural blockade in two patients studied. However, this was abandoned as the body of the C arm of the image intensifier used had to be positioned posterior and in close proximity to the scapula. This created unacceptable obstruction to the practical procedure.
The systolic blood pressure was measured (mmHg) during nerve block procedures using the Datex Cardioscope (Helsinki, Finland), and clinical examination of the chest was carried out 20-40 min afterwards. Patients were then observed in the ward for a few hours before being discharged home.
Statistical analysis of data was carried out using the Student's t-test and sign test (for TBN and TMB values). P<0.05 was considered to be significant.
The patients who were studied suffered from chronic pain of mean duration 4.4 (3.2) years. The mean value of VASP determined before combined SSNB and ACNB was 8.5 in the 16 patients studied. All patients admitted to some pain relief following neural blockade using 1% prilocaine, although no attempt was made to measure this on the VASP in view of the risk of displacing the needle tip before the injection of phenol.
A substantial relief of shoulder pain was observed shortly after combined SSNB and ACNB had been carried out and also at review after a mean time of 13 weeks. Thus, the mean value of VASP determined before neural blockade showed a statistically significant fall of between 67% and 69% (Table 1).
All patients studied were found to have severe limitation of active movement of the shoulder joint before neural blockade. Thus, prenerve block mean values for abduction, adduction and flexion were less than 30% of corresponding normal levels (abduction 90°, adduction 90°, flexion 180°). Furthermore, only 25% of the cases investigated were able to touch the back of the neck (TBN) and 12.5% the middle of the back (TMB) before neural blockade. Following this procedure mean values for abduction, adduction and flexion increased by 36% to 67% above corresponding prenerve block levels. In addition, the overall number of patients studied who were able to touch the back of the neck or middle of the back increased to 75% after neural blockade. These findings were statistically significant (Table 1).
Values for the systolic blood pressure determined during neural blockade ranged from 120 to 200 mmHg (mean 150, SD 27). No episodes of hypotension were observed during this procedure, and no evidence of pneumothorax was found on clinical examination of the chest after nerve blocks.
Ten of the 13 patients reviewed volunteered information regarding the quality of life after neural blockade. Five were more cheerful, three were pleased with the result, two felt relieved and one declined to attend for review on the grounds that his symptoms had resolved. Nine of these cases also said that their ability to carry out daily self-care routines had increased.
In the present study, SSNB and ACNB produced a substantial relief of shoulder pain, which occurred shortly after neural blockade and was maintained at review 13 weeks later. These findings were statistically significant (Table 1). The pain relief observed may have been caused by neural blockade of afferent C and A δ fibres by prilocaine and phenol and subsequent depletion of the neurotransmitter substance P  and nerve growth factor  in the synovium and afferent C fibres of the glenohumeral joint. A further factor contributing to pain relief could have been a decrease in central sensitization of dorsal horn nociceptive neurones  or 'wind-down' consequent upon a reduction of peripheral nociceptive input following neural blockade.
Stretching of the glenohumeral capsule during movement may cause an increase in nociceptive and mechanoreceptor input. This alone or in association with central sensitization could have enhanced existing muscle spasm and also impaired synergism between muscle groups , thus limiting active movement of the glenohumeral joint. Therefore, it is possible that a reduction in sensory input by SSNB and ACNB was responsible for the statistically significant increase in active movement of the shoulder joint that was found in the present study after this procedure (Table 1). ACNB could possibly have been particularly important in producing observed increases in abduction, flexion and TBN (Table 1), as such movements could stretch the inferior area of the glenohumeral capsule, which is innervated by the circumflex nerve .
Although no complications were observed during the present study, the procedure of SSNB and ACNB is not without potential hazards. The importance of i.v. access, identification of anatomical landmarks, monitoring of blood pressure and clinical examination of the chest cannot be overemphasized.
The findings in the present study support previous reports that neural blockade of the arthritic shoulder joint reduces pain, increases movement of the glenohumeral joint and improves the quality of life. This may be of particular value when shoulder arthroplasty is contraindicated or not readily available. However, whether or not combined SSNB and ACNB offers advantages over SSNB alone [1,2] requires clinical evaluation in the form of a formal, randomized, controlled, comparative study.
Using the data obtained in the present investigation, conventional power calculations (80% power, 5% test) indicate that 17 patients per group would be necessary to detect a change of one standard deviation (about 2 VASP) or 64 per group to detect a change of 0.5 standard deviations.
I am grateful to the theatre staff in the NHS Community Trust, Blackpool and Fylde and also the clerical staff of the Medical Records Department, Victoria Hospital, Blackpool for their assistance in carrying out this work. I am also indebted to Dr R. Henderson, Department of Mathematics and Statistics, Lancaster University, for his assistance in the statistical analysis of data.
1 Emery P, Wedderburn L, Grahame R. Suprascapular nerve block for shoulder pain in rheumatoid arthritis. Br Med J
2 Brown DE, James DC, Roy S. Pain relief by suprascapular nerve block in glenohumeral arthritis. Scand J Rheumatol
3 Soames RW. The skeletal system. In: Gray's Anatomy,
38th edn. New York: Churchill Livingstone, 1995: 631.
4 Wood Jones F ed. Buchanan's Manual of Anatomy,
7th edn. London: Bailliere Tindall and Cox, 1946: 412.
5 Gardner ED. The innervation of the shoulder joint. Anat Rec
6 Gordh T. Suprascapular nerve block. In: Eriksson E, ed. Illustrated Handbook in Local Anaesthesia
1st edn. London: Lloyd-Luke (Medical Books), 1969: 81-82.
7 Deal CL, Schnitzer T, Lipstein E, Seibold JR, Randal M, Stevens MD, Levy MD, Albert D, Reynolds F. Treatment of arthritis with topical capsaicin. A double blind trial. Clin Therap
8 Aloe L, Tuveri MA, Carcassi E, Levi-Moltalcini R. Nerve growth factor in the synovial fluid of patients with chronic arthritis. Arth Rheum
9 Woolf CJ. Somatic pain-pathogenesis and prevention. Br J Anaesth
10 Jerosch J, Castro WH, Halm H, Drescher H. Does the glenohumeral capsule have proprioceptive capability. Knee Surg Sports Traumatol Arthrosc