Suprascapular nerve block is faster and as effective as hydrodistension in relieving frozen shoulder-associated pain and disability: A prospective, single-blind observational study with a follow-up of 24 weeks

Abstract

Background and Aims:

Hydrodistension (HD) and suprascapular nerve block (SSNB) have been shown to reduce pain and improve shoulder function in frozen shoulder (FS). The aim of this study was to compare the efficacy of HD and SSNB in the treatment of idiopathic FS.

Material and Methods:

This was a prospective observational study. A total of 65 patients with FS were treated with SSNB or HD. The functional outcome was evaluated by Shoulder Pain and Disability Index (SPADI) score and active shoulder range of motion (ROM) measured at 2 weeks, 6 weeks, 12 weeks, and 24 weeks. Parametric data were analyzed using an independent sample T-test. Nonparametric data were analyzed using the Mann-Whitney test and Wilcoxon test. A P value less than 0.05 was considered significant.

Result:

At the end of 24 weeks, the two-group improved significantly from the baseline and the improvement was comparable between the two groups. ROM also improved significantly in both groups. At 2^nd week, SPADI score was significantly less in SSNB group (P < 0.05). About 43% of patients considered HD extremely painful.

Conclusion:

Both HD and SSNB are almost equally effective in reducing pain and improving shoulder function. However, SSNB leads to a faster improvement.

Introduction

Frozen shoulder (FS) is a disabling disease having a prevalence of 2–5%.^[1,2]

In most instances, the condition is self-limiting, and almost complete resolution of symptoms occurs. However, in a subset of patients, the disease may take years to resolve and may significantly affect the quality of life.^[2]

The primary complaint in FS is pain and progressive loss of shoulder range of motion (ROM). The pain is typically worst at night and disturbs sleep. Although most patients will have their symptoms resolved in the end, the long duration of the disease makes watchful neglect an impractical solution. Conservative options are still the mainstay of rehabilitation in FS. Multiple interventions have been successfully used in the early stage of disease with the optimum outcome.

Suprascapular nerve block (SSNB) has been found effective in reducing pain originating from the shoulder joint including that in FS.^[3,4] The improvement extends beyond the pharmacologic effects of block and multiple possible mechanisms have been described.^[5,6]

Hydrodistension (HD), also known as “arthrographic distension,” involves injecting normal saline in the shoulder joint with an intention to distend or ideally rupture the fibrotic capsule.^[7] HD probably decreases pain through capsular rupture/distension reducing the stretch on pain receptors in the capsule and periosteal attachments. It is usually combined with an intraarticular steroid injection. The primary objective of this study was to compare the efficacy of SSNB and HD in reducing pain and disability in FS. The secondary objective was to compare patient’s comfort during the procedure. A PUBMED search of the articles since 1990 has not produced any reference to similar previous comparisons.

Material and Methods

The study was approved by the institutional review board (IHEC-LOP/2018/IM0158). Written consent was obtained from all the patients. This study was a comparative, observational, and single-blind design. An a priori sample size calculation was done assuming a 90% power of study and 5% level of significance. Mean SPADI ± SD change at 6 weeks in SSNB was 21.2 ± 25%^[7] and in HD was 51 ± 42%;^[8] a minimum sample size of 60 patients (30 in each group) was needed. Assuming a dropout rate of 15%, a final sample size of 35 patients in each group was considered adequate.

All patients were adults and were having shoulder pain for more than 3 months. Those recruited were all in the first and second stages, as per the definition of the American Association of Orthopaedic Surgeons.^[9] Patients were recruited at anesthesia and orthopedic pain clinics between October 2018 and December 2019 [Figure 1]. The diagnosis of FS was based on criteria used by Bryant et al.^[7] and Bateman et al.^[10] Serious pathologies were excluded by measurement of erthrocyte sedimentation rate(ESR), random sugar, and rheumatoid factor. Apart from FS-associated changes, the shoulder radiographs (anteroposterior and scapular Y-views) were normal.

Figure 1.

Flow diagram showing enrolment of patients in the study

SSNB was given by an anesthesiologist, whereas HD was done by an orthopedic surgeon. As per pain clinic protocol, a repeat procedure was done only if at the completion of 3 weeks, the pain on numerical rating scale (NRS) score was 6. Those undergoing a repeat procedure were excluded from the analysis. Following each treatment, all the patients were given instructions regarding a home exercise program for rehabilitation.

For comparing the efficacy of treatment, Shoulder Pain and Disability Index (SPADI) score was used. SPADI score consists of a pain score and a disability score; the two are summated to derive a total score. We used a 13-point change in total score, to be “minimal important clinical difference” as chosen by Schmitt et al.^[11]; similarly, a three-point change in NRS was considered to be significant.^[12] SSNB was performed under ultrasound guidance in the lateral decubitus using a 5–13 Hz linear array transducer. Supraspinous fossa was scanned to identify the suprascapular notch with the pulsation of the suprascapular artery.^[13] A 22-G short bevel 8 cm Quincke needle was inserted into the suprascapular notch in the in-plane approach. After confirming the needle position, 40 mg triamcinolone acetonide mixed in 10 ml of 0.5% bupivacaine was injected.^[14] The shoulder was passively mobilized in all directions after 10 min.

All patients undergoing HD received intravenous tramadol (50 mg), 15 min before the procedure. Using aseptic precautions and fluoroscopic guidance, a 21-gauge spinal needle was inserted into the glenohumeral joint through an anterior approach. Intraarticular position was then confirmed by the injection of a small quantity of radio-opaque contrast material (2–3 ml of Omnipaque, GE Healthcare, India). Thereafter, 5 ml, 2% lignocaine was injected into the joint. Three minutes later, saline was injected into the joint in increments of 5 ml till the capsule ruptured or the patient had intolerable pain. Once an endpoint was reached, 40 mg triamcinolone acetonide in 4-ml bupivacaine 0.5% was injected. The shoulder was passively mobilized in all directions after 10 min.

Follow-up data were collected by a blinded observer. SPADI scores were collected at baselines, 2, 6, 12, and 24 weeks after the injection. Participant’s range of active forward flexion, lateral abduction, combined extension, and internal rotation was assessed at each time point. A goniometer was used to measure abduction and forward flexion. Combined extension and internal rotation were assessed using the back scratch test. Thumb positions were reported on an eight-point Likert scale; where 1 meant thumb reaching up to the ipsilateral greater trochanter; 2 – ipsilateral buttock; 3 – ipsilateral sacroiliac joint; 4 – lumbosacral junction; 5 – lower lumbar spine (up to L3);6 – upper lumbar spine (up to L1);7 – lower thoracic spine (up to T9); and 8 – middle thoracic spine (up to T5). For statistical purposes, thumb positions 1–3 were graded as severe restriction of internal rotation; 4–6 as a moderate restriction; and 7–8 as a mild restriction.

Kolmogorov-Smirnov test revealed the data to be normally distributed. The parametric data are reported as mean ± SD (confidence interval) and was analyzed using independent sample student’s T-test and paired sample T-test. The nonparametric data were analyzed using Mann-Whitney test, Wilcoxon test, or Fisher exact test. A P value less than 0.05 was considered significant. IBM-SPSS-16 software was used for data analysis.

Results

Both the groups were comparable with respect to the baseline demographic parameters [Table 1]. All the patients had disturbed sleep and difficulty in lying on the painful side (median NRS, 8 in SSNB, and 7 in HD). In 50% of patients, pain radiated to the upper limb. Next most common symptom was difficulty in carrying heavy objects (5 kg) in the affected limb (50%).

Table 1.

Baseline characteristics of the study groups

	SSNB group (n=32)	HD (n=33)	P
Age in years (mean±SD)	51.32±13.4	50.53±6.3	0.77
Sex (male: female)	11:21	15:18	0.36
Duration of pain (in months) Mean±SD, [95%CI]	4.46±1.97 [3.7-5.2]	4.7±1.88 [3.985-5.4]	0.64
Diabetic:nondiabetic	16:16	14:19	0.54
Baseline active lateral abduction in degree (mean±SD [95% CI]	66.4±12 [61.6-71.2]	65.6±11.8 [61.2, 70.2]	0.82
Baseline passive lateral abduction in degree [95% CI]	100.5±18.4 [93.5-107.5]	100±15.5 [94.1-105.9]	0.9
Severity of limitation in internal rotation (moderate:severe)	23:9	28:5	0.2

P<0.05 is significant. “CI” stands for confidence interval. Parametric data compared using Independent sample T-test. Nonparametric data compared using Wilcoxon test and Fischer exact test

About 43% of patients experienced excruciating pain during HD and hence the capsule could not be ruptured. The mean volume of saline injected in HD was 27 ± 5.6 ml. All the patients in SSNB group reported it to be mildly painful.

All patients completed the follow-up at 2^nd week. At 2 weeks, pain, disability, and total score were significantly lower from the baseline in both the groups. (P value < 0.0001). In the 6^th week, three and two patients in the SSNB and HD, respectively, required a repeat procedure, hence they were excluded from the analysis. In the SSNB group, the pain, disability, and total scores were significantly higher in comparison to those at the 2^nd week (P value 0.00078, <0.00001, and 0.0148, respectively). In the HD group, the pain and total scores were significantly lower than the scores at 2^nd weeks (P value < 0.0001, 0.0098, respectively); the disability score at 2 and 6 weeks was comparable (P value 0.186). The pain, disability, and total scores in both the groups did not show the further change between the 6^th and 12^th weeks (P value 0.83 and 0.37 for total score in the SSNB and HD groups, respectively). In both the groups, disability score and total score at 24^th week was significantly lower than that at 12^th week (P value 0.002, 0.01, respectively); pain score was comparable (P value 0.09) [Figure 2].

Figure 2.

The trend of the SPADI score over the follow-up duration

A between-group analysis showed a significantly lower pain, disability, and total score in SSNB group at 2^nd week (P value 0.0001, 0.0003, and 0.0001, respectively). Beyond the 2^nd week, the three scores were comparable in the two groups.

Active lateral abduction at 24^th week, although significantly better than the baseline was comparable between the two groups [Table 2a]. Active forward flexion at 24^th week was significantly better in the HD group (P value 0.004) [Table 2b]. At 24^th week, all the patients had a moderate degree of restriction while performing combined extension and internal rotation; 23 improved by 2 grades and 9 improved by 3 grades in SSNB group; 25 and 8 patients improved by 2 and 3 grades, respectively, in the HD group (P value 0.8).

Table 2a.

Comparison of active lateral abduction

	Baseline	24 weeks	P
SSNB group in degree (mean±SD) [95%CI]	66.4±12 [61.6-71.2]	95.3±20.5 [87.9-102.7]	<0.0001*
Hydrodistension group in degree (mean±SD) [95% CI]	65.6±11.8 [61.2-70.2]	90.2±13.3 [85.5-94.9]	<0.0001*
P	0.82	0.13

P<0.05 is significant. Means compared using independent sample t-test

Table 2b.

Comparison of active forward flexion

	Baseline	24 weeks	P
SSNB group	71±12.7	87.8±20.5	<0.0001*
Hydrodistension group	71.2±12.2	100.3±13.3	<0.0001*
P	0.69	0.004*

P<0.05 is significant. Means compared using independent sample t-test

Discussion

In the current study, we found that till the 24^th week, both SSNB and HD are equally efficacious in reducing FS-associated pain and disability. Improvement in ROM is also similar in both groups. However, SSNB has an earlier onset of pain relief as is evident from the lower pain and disability scores at 2^nd week. The sample size is small, but the statistical methods are well proven to give significant results.

In both groups, the duration of symptoms was approximately 4–5 months. At the baseline, passive ROM was approximately 50% more than the active ROM suggesting a stage 1 FS. Because pain is the most distressing complaint in FS, the patients are most likely to present in stage 1, hence the results of this study are clinically relevant.^[15]

HD works by stretching and rupturing the capsule. The higher pain and disability score in the HD group at 2^nd week can be explained by the more traumatic nature of the procedure.^[16,17] About 43% of the patients described HD to be extremely painful. Bryant et al.,^[7] compared HD with physiotherapy alone and found HD leads to significant improvement in pain and ROM. However, they studied the outcome parameters only beyond the 6^th week. Otherwise, their results are in accord with the findings in this study. Buchbinder et al.,^[16] also compared HD with placebo and concluded that HD leads to a significantly better outcome at 3 weeks follow-up. The duration of symptoms in their study was approximately 3–4 months that is less than in the current study.

Jones et al.^[18] also demonstrated that SSNB leads to a significantly faster improvement of pain score; improvement in pain score occurred as early as 1 week and persisted well beyond the 12^th week.^[15] In their study, the FS seemed to be of a lesser grade as the baseline active lateral abduction was 100 degree, which is better than in our study. In our study, the pain and disability score in the SSNB group showed a slight rise in 6^th week; a higher grade of FS can explain the shorter duration of relief.

The suprascapular nerve supplies up to 70% of sensory fibers to the glenohumeral joint.^[17,18] Therefore, SSNB provides an immediate significant reduction in pain, which enables better participation of the patient in the rehabilitation. The exact mechanism of action of SSNB in reducing pain is rather obscure. Local anesthetic has a duration of action of only a few hours, addition of steroid to the injectate can have a systemic effect. A study arm without steroids in the injectate, could have answered the question, but we could not incorporate it because of the ethics committee’s objection. Apart from systemic effect, the steroid may have a locoregional effect by the exertion of membrane stabilizing effect.^[19]

Although both these modalities have been studied individually in many studies,^{[3,4,18,20-23]} no study has compared HD with SSNB. Because FS may have many confounding variables, the observations of other studies may not help in establishing the superiority of one technique over the other. In this study, we found that SSNB and HD are equally effective in relieving FS pain. HD is more painful than SSNB and needs to be done in the operation theater. On the contrary, SSNB is relatively painless and may be done in an OPD setting. HD may produce a transient increase in pain that may persist for a few days. SSNB being atraumatic gives faster rehabilitation.

However, in view of the authors, the results of the study may not be generalizable. The results of intervention in FS are likely to get affected by the stage of FS. In our study, the baseline passive ROM was more than the active ROM implying a less severe adhesive capsulitis (FS stage 1). The results of the study may not be reproducible in other stages FS. In FS stage 1, the ROM is reduced mainly because of pain; in higher stages, pain progressively reduces while stiffness increases.^[2] We speculate that in FS stage 2, unlike stage 1, HD may be more difficult to perform because of a thickened capsule. Similarly, the natural course of FS in stage 2 may outlast the effective duration of SSNB.

We acknowledge some shortcomings in our study. First, being a prospective observational study, allocation bias is a concern. A randomized controlled trial would have been more appropriate. Because the assessor was blinded, the observer bias was minimized. Moreover, the post-treatment protocol was similar in either group to reduce the impact of confounding variables. Second, although some studies have shown that patients with diabetes do worse after treatment,^[22,24] we were unable to do subgroup analysis because of the small sample size.

Conclusion

HD and SSNB are effective and safe and have comparable efficacy in reducing pain and disability of stage 1 FS. However, SSNB provides earlier pain relief.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.