Efficacy and Safety of Early Neuromuscular Electrical Stimulation on Shoulder and Arm Paresis in Patients of Acute Ischaemic Stroke—A Quasi-experimental Study

Abstract

Background

Shoulder subluxation and shoulder pain are frequent after a stroke. Neuromuscular electrical stimulation (NMES) has been used widely in the sub-acute and chronic stages, but its use in the early stage is infrequent.

Aims and Objectives

The study’s objective is to see the effect of early neuromuscular electrical stimulation on hand function, shoulder subluxation, and shoulder pain after three months of stroke.

Methods

This study included 60 acute ischaemic stroke survivors. The intervention group received early NMES with standard rehabilitation, whereas the control group only received standard rehabilitation. The assessment was done at baseline and three months follow-up. The outcome measures were the presence of a sulcus sign, the Numerical Rating Scale (NRS), the Fugl-Meyer Assessment-Upper Extremity (FMA-UE, sub-scale A and A-D) and the Motor Activity Log (MAL) (sub-scale HOW WELL and AMOUNT).

Result

The intervention group had eight patients and the control group had nine patients with shoulder subluxation at follow-up. Hand functional outcomes were significantly better in the intervention group compared to the control group, with the FMA-UE-A scores of 27 (19–36) versus 18.5 (4.7–24.75), p = .007 and FMA-UE-A-D scores of 28.5 (4–48.25) versus 7.5 (4–23.75), p = .011. The HOW WELL and AMOUNT scores also showed significant differences, with 90 (30–150) versus 30 (0–90), p = .027. The intervention group did not experience any adverse events related to NMES.

Conclusion

The application of early NMES prevents shoulder subluxation and improves hand function. Randomised control trials with larger sample sizes and additional treatment sessions are needed to generalise the results.

Introduction

Stroke is the second leading cause of death and the third leading cause of disability worldwide. ¹ Less than 40% of stroke survivors regain their previous physical and mental conditions. ² Hand disability persists in 48 – 77% of stroke survivors and only 12–34% regain functional upper limb status, significantly limiting their functional and social participation.^{3, 4} Around 13–81% of stroke survivors suffer from shoulder pain and shoulder subluxation, which complicates their upper limb rehabilitation. ⁵ There are two possible mechanisms for shoulder subluxation. First, during arm abduction above 90°, the scapula must rotate upward to accommodate the humeral head. Due to flaccidity, there is a failure of this mechanism. Second, in a sitting position, the arm’s weight pulls the shoulder downward, unopposed due to flaccid shoulder muscles. ⁵ The association between shoulder subluxation and shoulder pain is unclear, but neither condition develops immediately after stroke. ⁶ Preventing and managing shoulder subluxation is crucial after a stroke. ⁷

Several rehabilitative approaches for motor control post-stroke include electrical stimulation, Rood’s technique, neurodevelopmental therapy, mirror therapy, Constrained-Induced Movement Therapy, Transcranial Direct Current Stimulation, robotics and AI-based rehabilitation. Electrical stimulation of paretic muscles is a major approach for improving motor impairment.^{5, 8} Neuromuscular electrical stimulation (NMES) is a type of electrical stimulation with physical and cortical effects. It is widely used in sub-acute and chronic stroke but is less frequently used in the acute stage. ⁹

Early rehabilitation after acute stroke can increase the favourable outcome by 13% and decrease the risk of secondary complications. ¹⁰ Specialised rehabilitation protocols are mostly applied at the sub-acute or chronic stage when the motor dysfunction starts developing; hence, using early NMES in the acute stage may be more beneficial than in the chronic stage. ⁵

Our study hypothesis is that incorporating early NMES within 10 days of stroke onset, alongside standard care, is more effective than standard care alone. Therefore, the objectives of this study were to evaluate the effectiveness of NMES applied early in the acute stage on upper extremity function, use of the hemiparetic arm, shoulder subluxation and pain, after three months of stroke.

Methodology

Study Procedure

The study was done in the inpatient department from February 2021 to January 2023. The ethical committee of the institution approved it (2021/EC/2879). Enrolled stroke participants were screened and written informed consent was obtained.

Participants

The patients were included in the study if they had a first-ever ischaemic stroke, with age above 18 years and stroke severity of 2–4 on the Modified Rankin Scale (mRS). Recurrent stroke, haemorrhagic stroke, presence of any known neurodegenerative pathology, presence of implanted electronic devices, epilepsy, respiratory insufficiency, pregnancy, peripheral neuropathies, cutaneous ulcers at the stimulation zone and inability to follow commands were excluded from the study. Participants were assigned to the two groups using a non-randomised, convenience-based allocation.

Intervention

Patient Preparation and Position

The patients were seated with arms resting on the pillow, the elbow at 90° flexion, and the wrist free and pronated. The paretic arm was free from any constraints such as jewellery, splints, monitoring devices, drip-set, clothes etc. The stimulation area was cleaned with 2% ethyl alcohol to reduce the skin impedance.

Electrode Placement and Neuromuscular Electrical Stimulation

Two carbonised rubber surface electrodes were placed on the motor points and secured with Velcro or micro-pore tapes. The active electrode size was smaller and circular, while the indifferent electrode was rectangular. The stimulating frequency was 35 Hz through the surface electrode of NMS 498 of Johari Digital, targeting type I slow-twitch fibres (small red and). Symmetrical, biphasic and rectangular impulses were used with the longest pulse time (350 µs). Impulse trains included long ‘on’ times, with progressive ramping of the impulse of 5–10 seconds and the ‘off time’ of 10 seconds between each train.

Two-channel stimulation using (four electrodes) was applied to the shoulder. An active electrode was placed on the supraspinatus, an indifferent electrode on the middle deltoid; another active electrode was placed on the posterior deltoid and an indifferent electrode on the anterior deltoid with alternate activation. For forearm and wrist stimulation, the extensor muscles of the wrist were stimulated. An active electrode was placed on the lateral epicondyle and an indifferent electrode at 10 cm above the lateral epicondyle, above the triceps insertion. An electrical session lasted 30 minutes for eight days. The NMES duration of eight days was based on the availability of patients in the acute care setting. After this treatment period, most of the patients were discharged and received conventional physiotherapy at home or at a nearby health facility. The control group did not receive early NMES. All participants received standard rehabilitation for 30 minutes a day.^{9, 11} Standard care involved active-assisted or passive mobilisation of the limb, optimal limb positioning, repetitive practice of the task and prevention of secondary complications. ⁴ Post-discharge, patients received telephonic guidance, video demonstration and printouts for the home programme. Rehabilitation was carried out by local therapists or family members, and adherence to the study protocol was monitored via telephonic call and in-person visit at one- and three-month follow-up.

Outcome Measurements

Our primary outcome measures were hand function and the use of a hemiparetic arm at three months of follow-up. Functional assessment of the hemiparetic arm was done by Fugl-Meyer Assessment-Upper Extremity (FMA-UE), ¹² (FMA sub-scale A, and A-D) and use of the hemiparetic arm was assessed by Motor Activity Log (MAL), ¹³ (sub-scale HOW WELL and AMOUNT). Assessment was done at baseline and three months of follow-up.

Our secondary outcome measures were the presence of the sulcus sign ¹⁴ for shoulder subluxation and pain intensity on the Numerical Rating Scale (NRS). ¹⁵ The sulcus sign was marked as the observable and palpable two-finger-width distance between the acromioclavicular joint and the head of the humerus at the shoulder joint in the affected arm. The NRS is a 10-point objective scale measuring pain intensity, with 10 indicating maximum pain, 0 points with no pain and 5 as medium and tolerable pain.

Statistical Methods

All statistical analyses were performed using IBM SPSS (version 22) for Windows. The chi-square test was used for categorical variables and the independent t-test was used for continuous variables. The Signed-rank test was used for within-group analysis. Mann-Whitney U test and Kruskal-Wallis test were used accordingly for inter-group analysis.

Results

A total of 750 patients were screened for participation, and finally, 60 acute stroke patients were included, 30 patients in group 1 (early NMES along with standard care) and 30 patients in group 2 (Standard care alone) (The flow diagram, Figure 1).

Figure 1. Study Flow Diagram.

Note: N = Number of participants, ICU = Intensive care unit, Group 1 = Intervention group, NMES = Neuromuscular electrical stimulation, Group 2 = Control group.

Baseline characteristics, such as age, sex, onset of stroke, hand dominance and type of infarct, were identical in both groups. Table 1 explains the baseline and clinical characteristics of the included patients.

Table 1. Baseline Characteristics and Baseline Scores of Participants of Intervention and Control Groups.

Demographic Characteristics
Variables		Group 1 (N = 30)	Group 2 (N = 30)	p Value
Age		62 (51–71.25)	55 (47.25–65)	.164
Gender: Male/Female		18/12	10/20	N/A
Hand dominance: Right/Left		30/0	30/0	N/A
Onset of stroke (days)		6 (5–7.25)	6 (6–8)	.177
Clinical Characteristics
mRS		4 (4–4)	4 (4–4)	.445
Side affected: Left/Right		16/14	12/18	N/A
Radiological Features
Anterior circulation		21	26	.21
Posterior circulation		06	03	.47
Others		3	1	.605
Outcome Variables
FMA-UE-A		6.5 (4–19)	6 (4–18.25)	.787
FMA-UE-A-D		7.5 (4–23.75)	4 (4–24)	.879
MAL	HOW WELL	0 (0–30)	0 (0–15)	.515
	AMOUNT	0 (0–30)	0 (0–15)	.515
NRS		4 (3–4.25)	4 (4–4)	.617

Note: Group 1 = Intervention group, Group 2 = Control group, NMES = Neuromuscular electrical stimulation, mRS = Modified Rankin Scale, FMA-UE-A = Fugl-Meyer Assessment, sub-scale A, FMA-UE-A-D = Fugl-Meyer Assessment, sub-scale A to D, MAL-HOW WELL = Motor Activity Log, sub-scale HOW WELL score, MAL-AMOUNT = Motor Activity Log, NRS = Numerical Rating Scale.

Intervention

No session was omitted or altered due to safety concerns and no adverse effects were observed during or after the NMES sessions. Visible muscle contractions were ensured in group 1. Baseline characteristics and baseline scores in both groups showed no significant difference between included patients across all variables. Both groups were comparable at baseline.

Outcomes

The outcomes were assessed after three months. Shoulder subluxation was absent at baseline in both groups but was observed at follow-up; eight participants in the intervention group and nine in the control group developed shoulder subluxation. No significant differences were observed between the groups (Table 2).

Table 2. Shoulder Subluxation Analysis.

Shoulder Subluxation	Group 1 FU (N = 30)	Group 2 FU (N = 30)	Fisher’s Exact Test, p Value
Present	8 (26.7%)	9 (30%)	1
Absent	22 (73%)	21 (70%)

Note: Group 1 = Intervention group, Group 2 = Control group, FU = Follow-up.

After three months of follow-up, all the variables except the NRS, in both groups, showed statistically significant differences from their baseline values. Although the NRS did not show statistically significant changes in either group, the intervention group maintained a lower median score with a p value of .487, indicating minimal change from baseline and suggesting a potential reduction in pain perception. In contrast, the control group exhibited a higher median score, with a p value of .06, bordering on significance, implying a possible increase in pain perception (Table 3).

Table 3. Within-group Analysis.

Outcome Measures		Group 1 (N = 30), Median (IQR)			Group 2 (N = 30), Median (IQR)
		BL	FU	p Value	BL	FU	p Value
FMA-UE-A		6.5 (4–19)	27 (19–26)	<.001	6 (4–18.25)	18.5 (4.75–24.75)	<.001
FMA-UE-A-D		7.5 (4–23.75)	47 (25.75–56.5)	<.001	4 (4–24)	28.5 (4–48.25)	<.001
MAL	HOW WELL	0 (0–30)	90 (30–150)	<.001	0 (0–15)	30 (0–90)	<.001
	AMOUNT	0 (0–30)	90 (30–150)	<.001	0 (0–15)	30 (0–90)	<.001
NRS		4 (3–4.25)	4.5 (2–6)	.487	4 (4–4)	5.5 (2–7)	.06

Note: Group 1 = intervention group, Group 2 = control group, mRS = Modified Rankin Scale, FMA-UE-A = Fugl-Meyer Assessment, sub-scale A, FMA A-D = Fugl-Meyer Assessment, sub-scale A to D, MAL-HOW WELL = Motor Activity Log, sub-scale HOW WELL score, MAL-AMOUNT = Motor Activity Log, sub-scale AMOUNT score, NRS = Numerical Rating Scale.

A comparison of outcome scores between the two groups showed the superiority of the intervention over the control group. The hand functional outcomes measure FMA-UE-A-score 27 (19–36) versus 18.5 (4.7–24.75), p = .007 and FMA-UE-A-D scores 28.5 (4–48.25) versus 7.5 (4–23.75), p = .011 were significantly better in the intervention group than the control group, respectively. The HOW WELL and AMOUNT scores 90 (30–150) versus 30 (00–90), p = .027 and 90 (30–150) versus 30 (00–90), p = .027 also showed a statistically significant difference between the two groups, which shows more use of the hemiparetic arm in the intervention group (Table 4).

Table 4. Between-group Analysis.

Variables		Group 1 FU (N = 30), Median (IQR)	Group 2 FU (N = 30), Median (IQR)	p Value
FMA-UE-A		27 (19–36)	18.5 (4.7–24.75)	.007
FMA-UE-A-D		28.5 (4–48.25)	7.5 (4–23.75)	.011
MAL	HOW WELL	90 (30–150)	30 (00–90)	.027
	AMOUNT	90 (30–150)	30 (00–90)	.027
NRS		4.5 (2–6)	5.5 (2–7)	.366

Note: Group 1 = intervention group, Group 2 = control group, FU = follow-up, IQR = inter-quartile range, mRS = Modified Rankin Scale, FMA-UE-A = Fugl-Meyer Assessment, sub-scale A, FMA-A-D = Fugl-Meyer Assessment, sub-scale A to D, MAL-HOW WELL = Motor Activity Log, sub-scale HOW WELL score, MAL-AMOUNT = Motor Activity Log, sub-scale AMOUNT score, NRS = Numerical Rating Scale

The inter-group comparison between left and right hemiparesis showed no significant difference. Hemiparetic sides when compared across the total sample, irrespective of group allocation, the right hemiparetic side showed greater improvement in FMA-UE-A-D (p = .02), FMA-UE-A (p = .01), HOW WELL (p = .00) and AMOUNT (p = .00), scores were significant, but there was a significant difference in pain perception (p = .334) (Table 5).

Table 5. Comparison of Right Versus Left Hemiparesis.

Variables		Right (Median, IQR)	Left (Median, IQR)	p Value
FMA-UE-A-FU		19 (8.75–33.5)	21.5 (19–35)	.015
FMA-UE-A-D		40 (9.25–52)	36.5 (22–50.5)	.022
MAL-FU	HOW WELL	75 (0.0–120)	60 (2.5–120)	.008
	AMOUNT	75 (0.0–120)	60 (2.5–120)	.008
NRS-FU		5 (2–7)	4.5 (2–6.75)	.334

Note: IQR = inter-quartile range, mRS = Modified Rankin Scale, FMA-UE-A = Fugl-Meyer Assessment, sub-scale A, FU = follow-up, FMA A-D = Fugl-Meyer Assessment, sub-scale A to D, MAL-HOW WELL = Motor Activity Log, sub-scale HOW WELL score, MAL-AMOUNT = Motor Activity Log, NRS = Numerical Rating Scale

Regression analysis was done to see the effect of comorbidities on stroke outcomes. Hypertension and pre-stroke exercise status were significant predictors of poor outcomes in the intervention group. Pre-exercise status was associated with high pain levels in the control group. Risk factors such as smoking and dyslipidaemia were not significant in both groups (Supplementary Table).

Discussion

Our study aimed to determine the effect of early NMES, applied within 10 days of stroke onset, on functional outcome, pain levels and shoulder subluxation after three months. The application of NMES within this early acute period differentiates our study from the previous research, where electrical stimulation was typically applied in sub-acute and chronic stages. A detailed assessment of baseline characteristics, within-group and between-group analyses, as well as regression analysis, was conducted to evaluate the impact of early NMES and other variables on upper limb functional recovery.

Our findings demonstrated the safety and feasibility of early NMES in acute care settings, with no adverse effects reported during or after the intervention. Although electrical stimulation has been used for motor relearning after stroke, ¹⁶ the feasibility of NMES in early acute care settings was a concern. Fletcher et al addressed this in a study involving patients within 72 hours of stroke onset, showing that early NMES was feasible and effective than standard care alone. ¹⁷ A similar study on traumatic brain injury patients in acute care showed no adverse effect on patients. ¹⁸ Additionally, research on critically ill patients with heterogeneous aetiology, including stroke, confirmed the feasibility and efficacy of early electrical stimulation in acute care. ¹⁹

A study in 2011 highlighted that anterior and inferior shoulder subluxation often develops after a stroke and electrical stimulation effectively prevents subluxation. The effect was more profound during the treatment, although it may diminish during follow-up. Preventing subluxation in the acute stage can subsequently improve motor control of the paretic arm during rehabilitation. If untreated, subluxation may lead to shoulder and arm pain, requiring more intensive and prolonged rehabilitation. ²⁰ Similarly, a meta-analysis in 2017 confirmed that electrical stimulation in the acute stage helps prevent shoulder subluxation. ⁵ However, contrary to these findings, patients in our study did not develop shoulder subluxation during the initial stage but had shoulder subluxation at follow-up.

In our study, despite the use of NMES and standard care, shoulder subluxation occurred in both groups. Possible contributing factors might be postural control, flaccid hemiparesis, mishandling, poor positioning and lack of adequate support for the affected limb during early mobilisation. ⁵

Our results suggest that early NMES has substantial effects on hand function after three months. Both the intervention and control groups improved significantly across all parameters, but the magnitude of improvement was greater in the intervention group. However, there was no significant difference in shoulder subluxation and shoulder pain across both groups, but the magnitude of pain in the control group during follow-up was higher compared to the intervention group. Our study is the first to see the effect of early NMES after acute ischaemic stroke and the application of electrical stimulation parameters following previous research. ⁹

In contrast to our results, a 2018 study found no significant effect of NMES on wrist function in acute ischaemic stroke. This study had several limitations, including a lack of information on stroke severity. Inclusion was done within six weeks and follow-up was done within three months, which may be too early to detect improvement. Also, the status of shoulder subluxation and shoulder pain was not mentioned, which may affect the patient’s motor ability. ²¹ Another study in 2019 compared robot-assisted therapy combined with functional electrical stimulation to intensive conventional therapy, showing no significant difference between the groups. However, when the patients with motor-evoked potential were analysed, robot-assisted therapy and functional electrical stimulation showed better results than intensive conventional therapy with similar stroke severity. ²² In contrast to these studies, our study applied electrical stimulation within 10 days of stroke onset, which may decrease the incidence of developing shoulder subluxation and shoulder pain.

Hand function is a critical issue following acute stroke, which can only be achieved with a stable shoulder. A stable shoulder provides the foundation for improved motor control and coordination in the arm and wrist.^{23, 24} Our study showed significantly improved shoulder and elbow motor control in the intervention group than in the control group. A similar study on acute stroke found that NMES and transcutaneous electrical nerve stimulation combined with standard rehabilitation led to significant improvement in posture, balance and functional mobility, further supporting the use of early transcutaneous electrical nerve stimulation for motor recovery. ²⁰

The initiation of treatment after the stroke is very crucial. A study on lower-limb recovery demonstrated significant improvement after two weeks of stroke in patients receiving early intensive rehabilitation, compared to standard care. Although the mRS score at three months was similar, the early rehabilitation group showed faster and better recovery. ²⁵ A comparison between early and delayed rehabilitation indicates that the timing of intervention plays a major role in motor recovery. ²⁰ Early electrical stimulation promotes motor recovery and has positive feedback, developed by active muscle contraction.

The study comparing the effects of NMES, transcranial magnetic stimulation and standard rehabilitation during acute and sub-acute stages showed a significant improvement in hand function in the transcranial magnetic stimulation and NMES groups compared to the standard rehabilitation alone. These findings underscore the importance of adopting approaches in the early phase post-stroke to maximise motor recovery. ²⁶

Comparison of right versus left hemiparesis showed that right hemiparetic patients are more aware of motor deficit and are motivated for rehabilitation; they will be more depressed with high demand and slow progress. In contrast, left hemiparetic patients have poor memory and impulsive behaviour, which may hinder their recovery. ²⁷ Our results are similar to the study, as the right hemiparesis has shown better improvement in arm functions compared to the left hemiparesis. The result also demonstrates the effect of the intervention on the dominant versus non-dominant hand, as all the patients were right-hand dominant.

Our results did not show any difference between right and left hemiparesis when comparing between groups. When all the right hemiparetic patients were compared to left hemiparetic patients, irrespective of group allocation, the right hemiparetic patients showed greater improvement. This is likely because the right-handed dominant patients tend to use their ipsilateral side more and the right hemiplegic patients often complete the task more successfully than left hemiplegic patients.^28–30 An earlier study shows higher activation levels of brain areas in left hemisphere stroke, in right-hand dominant patients. ³¹

Analysing the heterogeneous nature of stroke, the severity of impairment and the complex timeline of recovery, improvement in wrist function and fine motor control within three months after stroke may not be justifiable. Therefore, early rehabilitation should focus on preventing shoulder subluxation and pain, maintaining joint and muscle integrity and maximising the available functional capacity of the arm in the initial stages.

Limitation

The study has certain limitations. First, the relatively small sample size may have reduced the power to detect between-group differences, particularly in NRS scores. Second, important prognostic factors such as infarct location and size were not included. Third, during the initial days of hospitalization, patients may not have been fully motivated for rehabilitation, which may have influenced outcomes. Finally, the number of NMES sessions was limited, and there was no standardized control of the rehabilitation process after discharge.

Conclusion

Early initiation of NMES, within 10 days of stroke onset, may contribute to the prevention of shoulder subluxation and improvement in hand function at three months of stroke. Despite the limited number of NMES sessions, utilising an early rehabilitation window may enhance functional outcomes. NMES did not significantly affect shoulder pain. Randomised control trials with larger sample sizes and extended sessions of early NMES throughout follow-up period are needed to validate these findings.

Data Availability Statement

Data are available with a reasonable request from the corresponding author.

Statement of Ethics

This study was approved by the ethical committee of the Institute of Medical Sciences, Banaras Hindu University (2021/EC/2879). The study was conducted following the Declaration of Helsinki. All participants or their first-degree relatives gave written informed consent before participating in the study and to use the data for research and publication.