A randomised controlled trial to evaluate surface neuromuscular electrical stimulation to the shoulder following acute stroke

 

Church C 1,2,3, Price C 1,2, Pandyan D 4, Huntley S 1, Curless R 1,2, Rodgers H 1,2,3

 

1 Department of Geriatric Medicine, Northumbria Healthcare NHS Trust

2 School of Clinical Medical Sciences, University of Newcastle upon Tyne

3 School of Population and Health Sciences, University of Newcastle upon Tyne

4 Department of Physiotherapy Studies, Keele.

 

Email: c.j.church@ncl.ac.uk ; christopher.price@northumbria-healthcare.nhs.uk

Abstract

Objective: To evaluate a 4-week programme of surface neuromuscular electrical stimulation (sNMES) to the shoulder following acute stroke. 

Design: Pragmatic randomised controlled trial.

Setting: Two district general hospitals in the North of England.

Participants: One hundred and seventy six patients within 10 days of acute stroke.

Outcome measures: The primary outcome measure was the Action Research Arm Test (ARAT) 3 months after stroke [1]. Secondary outcome measures included impairment and upper limb pain [2][3][4][5].

Results: The groups were well matched at baseline. There was no difference in arm function between groups in terms of the primary outcome measure.  The median ARAT score [1] at 3 months was 50.0 in the intervention group (n=80) and 55.5 in the control group (n=75) (p=0.068). There were however significant differences in outcomes in favour of the control group when using other measures to assess arm function (the grasp and gross subsections of the ARAT [1], and the Frenchay Arm Test [6]). There was also a significant difference in favour of the control group when assessing impairment using the arm subsection of the Motricity Score [2]. The reasons for this are currently unclear. 

Conclusion: A 4-week programme of sNMES to the shoulder does not improve functional outcome. Subgroup analysis and analysis of compliance with therapy is ongoing. 

1        Introduction

1.1               Background

Upper limb (UL) impairment affects 85% of stroke patients and over half of these still experience problems 3-6 months later [7][8]. Reduced UL function is associated with pain and is an adverse prognostic indicator for subjective well being [9][10]. Electrical stimulation has been proposed as a method of improving outcome [11].

A review of the literature has concluded that future studies should be large, randomised, sham-controlled trials early after stroke with blinded validated outcome measures [12]. We have undertaken a randomised controlled trial to evaluate a 4-week programme of surface neuromuscular electrical stimulation (sNMES) to the shoulder following acute stroke. 

1.2              Objectives

To compare the UL function and impairment of acute stroke patients who received a programme of sNMES to the shoulder with those who received placebo at the end of a 4-week intervention period and 3 months after stroke.

To compare the prevalence of post-stroke UL pain between the intervention and control group at the end of a 4-week intervention period and 3 months after stroke.

To compare disability and global health status of the intervention and control groups at 3 months after stroke.

2        Methods

Patients admitted with acute stroke were assessed against the following eligibility criteria:

·         Residence within 15 miles of the 2 participating hospitals

·         Within 10 days of stroke onset

·         New UL impairment

·         Medically stable

·         No cognitive/language impairments or previous upper limb problem likely to influence assessments

·         No other diagnosis likely to affect rehabilitation or significant previous co-morbidity

·         No contraindication to sNMES

Participants were randomised using a central independent computerised service and were stratified by severity of UL weakness (Frenchay Arm Test [6] score 0,1 vs 2-5).

Participants in the intervention group received a 4-week programme of sNMES to the shoulder (1 hour 3 times daily). Two surface electrodes were placed over supraspinatus and posterior deltoid. The basic stimulation frequency was 30Hz. The stimulator on and off time was 15 seconds with a 3 second ramp up and 3 second ramp down time. Participants received a level of stimulation that was increased until a definite ‘shrug’ movement was seen at the shoulder. Participants in the control group were given a ‘sham’ stimulator, an internal disconnection preventing any current from being delivered. All participants received usual stroke unit care.

The primary outcome measure was the Action Research Arm Test (ARAT) 3 months after stroke [1]. Secondary outcome measures included impairment (motricity score and star cancellation test) [2][3] and UL pain (5 point severity scale and 0-10 numerical rating scale) [4][5]. Blinded outcome measures were undertaken at 4 weeks and 3 months. User satisfaction and compliance were also recorded.

Using variance and effect size from previous studies, we calculated that 180 subjects were required to achieve a 90% chance (power 0.9) of detecting a clinically significant difference in ARAT scores [1] (8 points) (two tailed alpha = 0.05).  The study had ethical approval.

3        Results

This presentation describes the initial analysis of the 3 month UL outcomes.

Table 1: Main reasons for exclusion (n=1227)  

·          341(28%) - no UL deficit

·          245(20%) - residence outside the area 

·          151(12%) - not within 10 days of stroke

·          140(11%) - cognitive deficit/receptive dysphasia

·          91(8%) - other diagnosis affecting rehabilitation

·          94(8%) - medically unstable

·          55(4%) - previous UL impairment (affected side)

·          70(6%) - died or discharged prior to screening

·          40(3%) - other e.g. permanent pacemaker

 
 
 

 

 

 

 

 


Between 1st January 2002 and 29th February 2004, 1627 patients were admitted to the participating stroke units with a diagnosis of possible stroke. Of these, 206 patients had not had an acute stroke and 1227 did not meet the eligibility criteria (Table 1).

Of the 194 eligible patients, 16 refused consent. One hundred and seventy eight patients were randomised but 2 were withdrawn after randomisation as they had not had a stroke.  A total of 176 subjects participated in the study, 90 in the intervention group and 86 controls.

Eighty participants in the intervention group completed the 3 month assessment (9 died, 1 refused follow up), and 75 in the control group (9 died, 1 refused follow up, 1 lost to follow up).  Subjects were randomised at a median (IQR) of 5 (3-7) days after stroke.

The randomisation groups were well matched in terms of age, gender, stroke subtype, severity and initial UL impairment (Table 2). There was an excess of subjects with left sided impairment as many with right sided impairment were excluded due to receptive dysphasia.

Table 2: Baseline characteristics and initial affected upper limb (UL) assessment (n=176)

 

Intervention (n=90)

Control (n=86)

P value

Male – n (%)

42(46.7%)

47(54.7%)

0.158

Median age (IQR)

75.5(64-81)

73.5(65.8-79)

0.287

Previous stroke affecting same side–n (%)

 

7(4%)

 

12(14%)

 

0.282

Side of UL impairment–n(%)Left

 

59(66%)

 

53(62%)

 

0.640

Stroke subtype [13]–n (%)

Total anterior circulation

Partial ant. circulation

Lacunar

Posterior circulation

 

 

29(32%)

 

23(26%)

36(40%)

2(2%)

 

 

24(28%)

 

25(29%)

35(41%)

2(2%)

 

 

0.924

Stroke type–n (%)

Infarct

Haemorrhage

 

86(96%)

4(4%)

 

78(91%)

8(9%)

 

0.834

UL pre-stroke pain–n (%)

 

2(2%)

 

6(7%)

 

0.162

UL post-stroke pain–n (%)

 

21(23%)

 

22(26%)

 

0.861

UL impairment (Motricity Score [2])+–median (IQR)

 

61.3(36-82.4)

 

63.3(36.4-8.1)

 

0.940

UL function (Frenchay Arm Test [6])+ –median (IQR)

 

 

0.5(0-4)

 

 

0(0-4)

 

 

0.793

UL function (ARAT [1]) +  - median(IQR)

 

0(0-45.5)

 

3(0-47)

 

0.641

Visual inattention (Star Cancellation Test fail [3])–n (%)

 

 

38(42%)

 

 

31(36%)

 

 

0.494

+  Lower scores indicate poorer outcomes

The results at 3 months are shown in Table 3. The control group achieved higher ARAT [1] scores than the intervention group (medians of 55.5 and 50.0 respectively) but this did not reach statistical significance (p=0.068).

However, significant differences were seen in the grasp and gross subsections of the ARAT [1], the controls achieving higher scores than the intervention group (p values of 0.014 and 0.015 respectively).

Table 3: Three month affected upper limb (UL)

outcomes (n=155)

 

Intervention (n=80)

Control (n=75)

P value

UL function (ARAT [1])+ median(IQR)

Grasp

Grip

Pinch

Gross

 

50.0(0-57)

12(0-18)

12(0-2)

15(0-18)

9(0-9)

 

55.5(38.3-57)

18(12-18)

12(8.8-12)

18(8.5-18)

9(9-9)

 

0.068

0.014

0.071

0.155

0.015

UL function (Frenchay Arm Test [6])+-median(IQR)

 

 

4(1-5)

 

 

5(3.8-5)

 

 

0.012

UL impairment (Motricity [2])+-median(IQR)       Arm

Leg

Total

 

 

84(56-100)

92(70-100)

88(66-100)

 

 

93(77-100)

86(76-100)

89(76.5-100)

 

 

0.025

0.948

0.248

Visual inattention (Star Cancellation Test fail [3])–n(%)

 

 

25(31%)

 

 

18(23%)

 

 

0.281

UL pain–n(%)

40(50%)

35(47%)

0.678

+  Lower scores indicate poorer outcomes

The control group performed better on the Frenchay Arm Test [6] and the arm subsection of the Motricity Score [2] and this was statistically significant. When assessing visuospatial impairment using the Star Cancellation Test [3], a greater percentage of subjects failed this test in the intervention group compared to the controls but this difference was not statistically significant. The prevalence of UL pain was similar between groups. 

4        Discussion and Conclusions

There was no difference in arm function between groups in terms of the primary outcome measure. 

There were however significant differences in outcomes in favour of the control group when using other measures to assess arm function, i.e. the grasp and gross subsections of the ARAT [1], and the Frenchay Arm Test [6]. There was also a significant difference in favour of the control group when measuring impairment in terms of the arm subsection of the Motricity Score [2]. 

The reasons for this are currently unclear.  There were no significant differences between the randomisation groups at baseline. 

Our preliminary conclusion is that this surface neuromuscular electrical stimulation regime should not be part of routine clinical practice. Subgroup analysis and analysis of compliance with therapy is ongoing. 

References

[1]                 Lyle RC. A performance test for assessment of upper limb function in physical rehabilitation treatment and research. Int J Rehabil Res,4:483-92,1981.

[2]                 Demeurisse G, Demol O, Robaye E.  Motor evaluation in vascular hemiplegia. European Neurology,19:382-9,1980.

[3]                 Halligan P, Marshall J, Wade DT. Visuospatial neglect: underlying factors & test sensitivity. Lancet,2(8668):908-10,1989.

[4]                 Price DD, Bush FM, Long S, et al.  A comparison of pain measurement characteristics

[5]                 of mechanical visual analogue and simple numerical rating scales. Pain,56:217-26,1994.

[6]                 Downie WW, Leatham PA, Rhind VM, et al.  Studies with pain rating scales. Ann Rheumatic Diseases,37:378-81,1978.

[7]                 De Souza LH, Langton-Hewer RL, Miller S. Assessment of recovery of arm control in hemiplegic stroke patients. International Rehabilitation Medicine,2:3-9,1980.

[8]                 Heller A, Wade D, Wood VA, et al. Arm function after stroke: measurement and recovery over the first three months.