Comparison of the effectiveness of pulsed radiofrequency of the suprascapular nerve and intra-articular corticosteroid injection for hemiplegic shoulder pain management

Tae Hoon Kim¹, Min Cheol Chang^2,*

Show Less

¹ Department of Physical Medicine and Rehabilitation, Seoul Songdo Hospital, 04597 Seoul, Republic of Korea

² Department of Physical Medicine and Rehabilitation, College of Medicine, Yeungnam University, 705-717 Daegu, Republic of Korea

^*Correspondence: wheel633@ynu.ac.kr (Min Cheol Chang)

J. Integr. Neurosci. 2021, 20(3), 687–693; https://doi.org/10.31083/j.jin2003073

Submitted: 13 April 2021 | Revised: 30 April 2021 | Accepted: 21 May 2021 | Published: 30 September 2021

This is an open access article under the CC BY 4.0 license (https://creativecommons.org/licenses/by/4.0/).

Download PDF

Brower Figures

Cite

Abstract

Many patients complain of hemiplegic shoulder pain following stroke. Here, the effectiveness of pulsed radiofrequency stimulation of the suprascapular nerve is compared with intra-articular corticosteroid injection for chronic hemiplegic shoulder pain following stroke. This single-center, prospective, randomized controlled study included 20 patients with hemiplegic shoulder pain after stroke, randomly assigned to the pulsed radiofrequency and intra-articular corticosteroid injection treatment groups (n = 10 in each). Hemiplegic shoulder pain severity was measured by numeric rating scale and passive shoulder range motion was assessed at baseline and one and two months after each procedure. Compared to the baseline numeric rating scale scores, post-treatment scores decreased significantly in both groups (p $<$ 0.001). However, score reduction through time was significantly greater for intra-articular corticosteroid injection for pulsed radiofrequency (p $<$ 0.001). Similarly, a significant post-treatment increase was observed in almost all range of motion measurements in both groups (pulsed radiofrequency group: flexion, p = 0.015; abduction, p = 0.014; external rotation, p = 0.038; internal rotation, p = 0.063; intra-articular corticosteroid injection group: all range of motion, p $<$ 0.001). Moreover, the measurements for all ranges of motion in the intra-articular corticosteroid injection group were significantly higher than those in the pulsed radiofrequency group (p $<$ 0.001). Thus, intra-articular corticosteroid injection appears more effective than pulsed radiofrequency for control of hemiplegic shoulder pain, whereas, pulsed radiofrequency of the suprascapular nerve has minimal effect. However, in patients at risk for developing complications following corticosteroid injections, pulsed radiofrequency of the suprascapular nerve may be an option in management of hemiplegic shoulder pain.

Keywords

Pulsed radiofrequency

Corticosteroid

Hemiplegic shoulder pain

Stroke

Peripheral neurons

Neurology

1. Introduction

Following stroke many patients experience different types of pain syndrome [1, 2]. Among these, hemiplegic shoulder pain (HSP) is one of the most frequent [3, 4]. Although their etiology is unclear, adhesive capsulitis, subacromial bursitis, shoulder subluxation and spasticity are known to be associated with the development of HSP [5]. HSP resolves in most cases within six months, but approximately 20% of patients experience debilitating and persistent shoulder pain [6]. Persistent HSP may hinder functional recovery after stroke, decrease quality of life and lead to emotional problems, such as depression and anxiety [5].

Generally, clinicians employ range of motion (ROM) exercises, oral medications and other modalities of therapy to control HSP [7]. However, despite this, HSP persists in many patients. For the treatment of persistent HSP, unresponsive to conventional modes of therapy, intra-articular corticosteroid injection (ICI) into the shoulder joint is widely used, but its palliative effect has only a relatively short duration [8]. Furthermore, corticosteroids may have adverse effects, including allergic reactions, flushing, hyperglycemia, menstrual disturbances and adrenal suppression [9, 10]. Suprascapular nerve block (SSNB) is another option for relieving HSP [11, 12]. The suprascapular nerve (SSN) originates from the ventral rami of the C4, C5 and C6 spinal nerves and emerges from the upper trunk of the brachial plexus. It provides 70% of the sensory innervation of the shoulder joint [13]. Thus, blockage of pain transmission through the SSN provides effective control of HSP. However, the efficacy of SSNB varies with the study population and depends on the therapeutic modality with which it is compared [14]. Moreover, the effect of SSNB may also be limited due to the short duration of action of local anesthetic agents.

Pulsed radiofrequency (PRF) was introduced by Sluijter in 1997 [15]. Many previous studies have reported its safety and effectiveness in alleviating pain. In this procedure, an electrical field is generated around the targeted nerves or tissues, but these structures are not damaged by this stimulation. Conventional radiofrequency provides continuous electrical simulation to the targeted nerves or tissues and ablates these structures due to the high temperatures around the radiofrequency needle tip [16, 17, 18, 19]. However, PRF only provides electrical stimulation for brief durations between prolonged resting phases. Therefore, it does not produce heat sufficient to damage surrounding structures [20]. The primary mechanism of PRF is modulation of pain signals by the generated electrical field [21]. Several previous studies have reported that PRF can be effective in controlling various types of pain, such as neuralgia, muscle pain and joint pain [9, 22, 23]. Additionally, recent studies have shown that PRF of the SSN may effectively manage shoulder pain [24, 25, 26, 27]. However, the effect of PRF stimulation of the SSN in the management of HSP after stroke remains undetermined.

Therefore, this study, investigated whether PRF stimulation of the SSN could effectively reduce shoulder pain and increase the shoulder ROM in patients with HSP following chronic stroke. Furthermore, the effectiveness of PRF and ICI was compared in these patients.

2. Materials and methods

2.1 Subjects

20 consecutive stroke patients were prospectively recruited, including 13 men and 7 women, mean age 58.0 $\pm{}$ 8.7 years (range, 42–69 years). 11 patients were diagnosed with cerebral infarct and 9 with intracerebral hemorrhage. Mean stroke duration was 13.2 $\pm{}$ 2.6 months. All patients were admitted to the Department of Physical Medicine and Rehabilitation at a hospital for post-stroke rehabilitation treatment. Inclusion criteria were: (1) history of stroke, (2) stroke duration at least months, (3) age 21–70 years, (4) presence of hemiplegia caused by stroke, (5) significant shoulder pain with a minimum score of four on the numeric rating scale (NRS, where “0” indicates no pain and “10” indicates the most severe pain) persistent for at least three months, (6) no change in pain severity (NRS score) after four weeks despite taking pain medication (meloxicam, acetaminophen, and tramadol hydrochloride alone or in combination), (7) limitation in passive ROM on physical examination, (8) no history of ICI administration in the shoulder and (9) no severe aphasia or cognitive dysfunction (to ensure accurate measurement of the pain degree).

Written informed consent was obtained from all subjects. The study protocol was approved by the Institutional Review Board of Yeungnam University Hospital (2017-02-011). The sample size was based on the findings of a previous study [24]. The reduction in the NRS score after treatment (PRF of the SSN vs. lidocaine injection) was 1.50 $\pm{}$ 1.12 in that study. Applying a type I error of 0.05 for 80% power and using a two-sided test, a sample size of nine subjects per group was found to be adequate for this study. Allowing for a 10% dropout rate, 10 subjects were recruited for each group.

Subjects were allocated (by random number table) to receive either PRF stimulation of the SSN (PRF group) or ICI administration in the shoulder joint (ICI group) (Fig. 1). An experienced clinician performed the assigned procedure once for each subject under ultrasound (US) guidance. All patients underwent rehabilitation therapy (Monday to Friday: 2.5 h/day; Saturday: 1 h/day).

Fig. 1.

Study flow diagram for pulsed radiofrequency and intra-articular corticosteroid injection groups. Abbreviations: ROM, range of motion; NRS, numeric rating scale; PRF, pulsed radiofrequency; ICI, intra-articular corticosteroid injection.

2.2 Procedures

The procedure was conducted from behind the patient by a physician with the patient in an upright position. A US device (LOGIQ P6, General Electric, Republic of Korea) with an 11-MHz linear probe was used to guide injection administration. During PRF stimulation and ICI administration, the hand on the side of the injection could be placed on the contralateral shoulder. For PRF stimulation, the US probe was placed just above and parallel to the scapular spine. The supraspinatus muscle and scapular notch were visualized on the US images. The SSN was seen as a hyperechoic structure in the scapular notch, 3–4 cm below the transverse scapular ligament. A 22-gauge, 10-cm cannula with a 10-mm active tip (Cosman RF Cannula, CC10522, Cosman Medical, Burlington, MA, USA) was inserted towards the scapular notch using a medial-to-lateral approach (Fig. 2). An electrode was connected to the PRF needle. When the patient reported paresthesia during sensory stimulation (50 Hz, 1 ms pulse width, $\leq$ 0.5 V) and contraction of the supra- and infraspinatus muscles was observed during motor stimulation (2 Hz, 1 ms pulse width, $\leq$ 0.5 V), the PRF (Cosman G4 radiofrequency generator, Cosman Medical) was applied at 2 Hz and 30 ms pulse width for 360 s at 45 V. The electrode tip temperature did not exceed 42 ${}^{\circ}$ C. For ICI, the US probe was placed over the long axis of the myotendinous junction of the infraspinatus tendon, presenting the posterior glenoid labrum and the posterior part of the humeral head. The needle tip was advanced into the glenohumeral joint and a 10-mL solution containing triamcinolone actinide (20 mg), 1% lidocaine (1 mL) and normal saline (8.5 mL) was injected.

Fig. 2.

Ultrasound image of suprascapular nerve with linear ultrasound probe. Abbreviations: TM, trapezius muscle; SM, supraspinatus muscle; SN (arrow head), suprascapular nerve; arrow, pulsed radiofrequency needle pathway.

2.3 Outcome measurements

Patients were assessed prior to the study (baseline) and at one and two months after the procedure. The same investigator, blinded to the therapeutic intervention, assessed the clinical outcomes before and during follow-up. The intensity of pain in the affected shoulder was assessed using the NRS. Passive shoulder joint ROM was investigated using a goniometer. Shoulder flexion, abduction and external and internal rotations were assessed with the patients in a supine position. Degree of shoulder flexion and abduction was measured with the elbow in extension and ROM of external and internal rotations evaluated with the elbow at 90 ${}^{\circ}{}$ flexion and the arm at 90 ${}^{\circ}{}$ abduction. For precise measurement of the ROM, the shoulder joint was positioned with the proximal segment stabilized to exclude articular movement. To measure passive ROM, the examiner moved the patient’s arm until the movement was mechanically limited or the patient felt pain.

Change in NRS score was measured as the difference between the pre-treatment score and the score at each time point during post-treatment follow-up. Change in NRS (%) was calculated as “((pre-treatment score - score after treatment)/pre-treatment score) $\times{}$ 100”. Successful treatment was defined as $\geq$ 50% reduction in the NRS score at the two month follow-up when compared to the pre-treatment NRS score.

2.4 Statistical analysis

Data were analyzed using Statistical Product and Service Solutions software (SPSS, version 22.0, IBM Corporation, Armonk, NY, USA). Intergroup comparison of the demographic data and the rate of successful pain relief employed the Mann–Whitney U-test and Fisher’s exact test, respectively. A normality test employed the Kolmogorov–Smirnov test. Repeated measure one-factor analysis was used to evaluate changes in the NRS scores in the PRF and ICI groups. A repeated measure two-factor analysis was used to compare the changes between the groups over time. Multiple comparisons were obtained using the Bonferroni correction. The level of statistical significance was set at p $<$ 0.05.

3. Results

All enrolled subjects completed the study and no adverse events were observed in either group. There were no significant differences in the demographic characteristics between the two groups (p $>$ 0.05) (Table 1).

Table 1.Patient demographic characteristics and baseline clinical data.

		PRF group	ICI group	p-value
		(n = 10)	(n = 10)	p-value
Age, years		56.5 $\pm$ 9.3	59.4 $\pm$ 8.3	0.579
Sex (Male : Female), n		6:4	7:3	0.648
Affected hemisphere (Right : Left), n		6:4	4:6	0.656
Months from onset		13.1 $\pm$ 2.5	13.3 $\pm$ 2.9	0.912
Stroke type (infarct : hemorrhage), n		7:3	4:6	0.648
Initial NRS		5.9 $\pm$ 1.4	6.2 $\pm$ 1.1	0.684
Initial passive ROM of shoulder, degree
	Flexion	123.5 $\pm$ 29.8	126.0 $\pm$ 24.1	0.796
	Abduction	121.5 $\pm$ 28.3	122.0 $\pm$ 26.6	0.971
	External rotation	56.0 $\pm$ 11.7	61.0 $\pm$ 9.9	0.353
	Internal rotation	67.0 $\pm$ 8.6	69.5 $\pm$ 6.0	0.393
MBC		2.3 $\pm$ 1.1	2.1 $\pm$ 0.9	$<$ 0.999
FAC		2.4 $\pm$ 0.5	2.4 $\pm$ 0.5	0.912
MMSE		27.2 $\pm$ 2.7	27.7 $\pm$ 2.7	0.684
Values are given as mean $\pm$ standard deviation. Abbreviations: PRF, pulsed radiofrequency; ICI, intra-articular corticosteroid injection; NRS, numeric rating scale; ROM, range of motion; MBC, modified Brunnstrom classification (scores range from 1 to 6; higher scores indicate better hand function); FAC, Functional Ambulation Category (scores range from 0 to 5; higher scores indicate better walking ability); MMSE, Mini-Mental State Examination.

The intragroup analysis showed that the NRS scores differed significantly over time in both the groups (PRF group: p = 0.009; ICI group: p $<$ 0.001). In the PRF group, the NRS scores at baseline and at one and two months post-treatment were 5.9 $\pm{}$ 0.8, 5.0 $\pm{}$ 2.2, and 5.2 $\pm{}$ 2.0, respectively. In the ICI group, the mean NRS score decreased from 6.2 $\pm{}$ 1.1 at baseline to 2.8 $\pm{}$ 0.9 at one month and 3.2 $\pm{}$ 1.1 at two months post-treatment. In both groups, the post-treatment scores at one and two months were significantly lower than the pre-treatment scores (PRF group: one month, p = 0.010, two months, p = 0.025; ICI group: p $<$ 0.001 for both one and two months) (Fig. 3). Except for internal rotation, all measurements of the passive shoulder ROM differed significantly over time in the PRF group (flexion ROM, p = 0.015; abduction ROM, p = 0.014; external rotation ROM, p = 0.038; and internal rotation ROM, p= 0.063). A significant increase in the passive shoulder ROM was observed in the flexion, abduction, and external rotation at one and two months after PRF stimulation (flexion ROM at one month, p = 0.022; flexion ROM at two months, p = 0.025; abduction ROM at one month, p = 0.005; abduction ROM at two months, p = 0.025; external rotation ROM at one month, p = 0.017; external rotation ROM at two months, p = 0.048; internal rotation ROM at one month, p = 0.052; and internal rotation ROM at two months, p = 0.081) (Fig. 3). In the ICI group, all ROM measurements increased significantly over time (p $<$ 0.001), and a significant increase was noted at one and two months after ICI (p $<$ 0.001) (Fig. 4).

Fig. 3.

Changes in the numeric rating scale scores. Compared to the pre-treatment NRS scores, both groups showed a significant decrease in the scores at one and two months after treatment. However, the NRS scores were significantly lower in the ICI group than in the PRF group at one and two months after the procedure. Abbreviations: NRS, numeric rating scale; PRF, pulsed radiofrequency; ICI, intra-articular corticosteroid injection. *p $<$ 0.05: Intragroup comparison of NRS values at pre- and post-treatment (repeated measure one-factor analysis). †p $<$ 0.05: Intergroup comparison at each time point (repeated measure two-factor analysis).

Fig. 4.

Changes in the passive shoulder joint range of motion measurements. In the PRF group, a significant increase in the shoulder ROM was observed in flexion, abduction, and external rotation at one and two months after PRF. In the ICI group, a significant increase was found in all ROM measurements at one and two months after ICI administration. In the intergroup comparison, all ROM measurements were significantly higher in the ICI group than in the PRF group. Abbreviations: ROM, range of motion; PRF, pulsed radiofrequency; ICI, intra-articular corticosteroid injection. *p $<$ 0.05: Intragroup comparison between the shoulder ROM at pre- and post-treatment (repeated measure one-factor analysis). †p $<$ 0.05: Intergroup comparison at each time point (repeated measure two-factor analysis).

In the intergroup analysis, changes in the NRS scores and all ROM measurements differed significantly over time between the two groups (p $<$ 0.001). The NRS score was significantly lower and all ROM measurements were significantly higher in the ICI group as compared to the PRF group at one and two months post-treatment (p $<$ 0.001) (Figs. 3,4). The individual data of all subjects are presented in Supplementary 1.

Pain relief (pain improvement $\geq$ 50%) was achieved in two out of ten subjects in the PRF group (20%) and in seven out of ten subjects in the ICI group (70%). There was a trend towards more successful pain relief in the ICI group than in the PRF group (p = 0.070).

4. Discussion

This study investigated the effectiveness of PRF of the SSN in alleviating pain and compared it with ICI treatment in patients with HSP after chronic stroke. The severity of pain measured by NRS was significantly reduced at one and two months after both interventions. However, the NRS scores were significantly lower in the ICI group than in the PRF group at one and two months post-treatment. Furthermore, only two out of ten patients (20%) who underwent PRF had successful pain relief ( $\geq$ 50% pain reduction at two months after treatment). In contrast, ICI showed a trend towards more successful pain relief (70%) than PRF, although the difference was not statistically significant. Moreover, it was found that both interventions improved nearly all the shoulder ROM measurements and the effect was sustained for at least two months. However, similar to the improvement in pain, all ROM measurements were significantly higher in the ICI group as compared to the PRF group.

The mechanism of action of PRF of the SSN has not yet been clearly identified. However, based on previous studies, some mechanisms can be suggested. Cosman et al. [28] reported that pain reduction after PRF stimulation was correlated with the long-term depression of synaptic transmission following low-frequency electrical stimulation of neurons by PRF. Therefore, the PRF of the SSN seems to inhibit the transfer of nociceptive stimuli from the shoulder joint. Additionally, Hagiwara [29] reported that when the electromagnetic field produced by PRF stimulates the peripheral nerves, excitatory C-fibers are inhibited and activity of the descending serotonergic and noradrenergic pain inhibitory pathways is enhanced. The effectiveness of ICI in this study is consistent with the findings of several previous studies on HSP [5, 30, 31]. Here, the potent anti-inflammatory effect of corticosteroids seemed to have reduced the shoulder pain and increased the extensibility of the shoulder joint capsule in the ICI group [5]. It is believed that the outcome of ICI was better than that of PRF because the corticosteroid injection was administered at the origin of pain. In contrast, PRF controlled the shoulder pain indirectly by acting on the SSN. Despite the superior effect of the ICI, its potential adverse effects should be considered [9, 10]. Therefore, PRF treatment of the SSN may be a helpful therapeutic modality in the management of HSP, especially in patients contraindicated for or who experience significant side effects with corticosteroid injections. Furthermore, US was used to guide the PRF needle when approaching the SSN. US not only increases the safety and accuracy of the procedure, but also diagnoses shoulder disorders accurately [32, 33].

Several previous studies have demonstrated the efficacy of PRF treatment of the SSN for shoulder pain due to rotator cuff pathology or adhesive capsulitis. However, to the author’s knowledge, only one study has evaluated the effectiveness of PRF stimulation of the SSN in HSP after stroke [24, 25, 26, 27]. In 2018, Picelli et al. [34] retrospectively evaluated the effect of PRF of the SSN in six chronic stroke patients with HSP. They reported that the intensity of shoulder pain evaluated using the visual analog scale was significantly reduced and the shoulder ROM significantly improved with PRF, with a sustained effect for at least four months. The present study is the first prospective randomized controlled study to compare the effectiveness of PRF and ICI in HSP after chronic stroke. However, its limitations should be considered. First, there was no investigation of the long-term effects of the therapy beyond two months. Second, patients were not blinded to the treatment they received. Third, the study did not include a placebo group. Finally, the mechanisms causing HSP, such as soft tissue problems, motor impairment and neural factors, were neither considered nor evaluated.

5. Conclusions

Results indicate that the effectiveness of PRF of the SSN in pain reduction, as evaluated by NRS score, was minimal in patients with HSP after stroke. Although NRS scores were significantly reduced following PRF, only 20% of patients experienced successful pain reduction. In contrast, there was a greater decrease in the severity of pain after ICI, and 70% of patients experienced successful pain reduction. Furthermore, improvement in the shoulder ROM was greater with ICI than with PRF. Therefore, it is suggested that ICI may more effectively control HSP in patients with stroke than PRF of the SSN. However, in patients at a risk of complications with corticosteroids, PRF of the SSN may be a reasonable clinical option.

Abbreviations

HSP, hemiplegic shoulder pain; ROM, range of motion; SSNB, suprascapular nerve block; SSN, suprascapular nerve; PRF, pulsed radiofrequency; ICI, intra-articular corticosteroid injection; NRS, numeric rating scale; US, ultrasound.

Author contributions

THK and MCC conceived and designed the experiments; THK and MCC performed the experiments; THK and MCC analyzed the data; THK and MCC contributed reagents and materials; THK and MCC wrote the paper.

Ethics approval and consent to participate

Informed consent was obtained from all the participants. The Institutional Review Board of Yeungnam University Hospital approved this study (code 2017-02-011).

Acknowledgment

Not applicable.

Funding

This study was supported by a National Research Foundation of Korea Grant funded by the Korean government (Grant no. NRF-2021R1A2C1013073).

Conflict of interest

The authors declare no conflict of interest.

Supplementary Material

Supplementary Table 1.docx

References

[1]

Choi G, Chang MC. Effects of high-frequency repetitive transcranial magnetic stimulation on reducing hemiplegic shoulder pain in patients with chronic stoke: a randomized controlled trial. International Journal of Neuroscience. 2018; 128: 110–116.