Upper limb neuromuscular weakness occurs frequently

Transcription

1 Effects of Conventional Physical Therapy and Functional Strength Training on Upper Limb Motor Recovery After Stroke: A Randomized Phase II Study Neurorehabilitation and Neural Repair Volume 23 Number 4 May The Author(s) / Catherine Donaldson, PhD, Raymond Tallis, FmedSci, Simon Miller, DPhil, Alan Sunderland, PhD, Roger Lemon, PhD, and Valerie Pomeroy, PhD Background. Functional training and muscle strength training may improve upper limb motor recovery after stroke. Combining these as functional strength training (FST) might enhance the benefit, but it is unclear whether this is better than conventional physical therapy (CPT). Comparing FST with CPT is not straightforward. Objective. This study aimed at assessing the feasibility of conducting a phase III trial comparing CPT with FST for upper limb recovery. Methods. Randomized, observer-blind, phase II trial. Subjects had upper limb weakness within 3 months of anterior circulation infarction. Subjects were randomized to CPT (no extra therapy), CPT + CPT, and CPT + FST. Intervention lasted 6 weeks. Primary outcome measure was the Action Research Arm Test (ARAT). Measurements were taken before treatment began, after 6 weeks of intervention, and 12 weeks thereafter. Attrition rate was calculated and differences between groups were interpreted using descriptive statistics. ARAT data were used to inform a power calculation. Results. Thirty subjects were recruited (8% of people screened). Attrition rate was 6.7% at outcome and 40% at follow-up. At outcome the CPT + FST group showed the largest increase in ARAT score and this was above the clinically important level of 5.7 points. Median (interquartile range) increases were 11.5 (21.0) for CPT; 8.0 (13.3) for CPT + CPT; and 19.5 (22.0) for CPT + FST. The estimated sample size for an adequately powered subsequent phase III trial was 279 subjects at outcome. Conclusion. Further work toward a phase III clinical trial appears justifiable. Keywords: Stroke rehabilitation; Physical therapy; Clinical trial; Upper extremity hemiplegia Upper limb neuromuscular weakness occurs frequently after stroke 1 with loss of muscle strength and dexterity together considered to produce the largest impact on functional recovery. 2,3 Muscle strength may be related to functional ability 4 and may contribute more to loss of functional ability than impaired dexterity, muscle tone, sensation, or pain. 5,6 The underlying mechanisms of neuromuscular weakness after stroke possibly include atrophy of type II fibers, increased proportion of type I fibers, loss of motor units, collateral reinnervation, and altered firing of motor unit groups. 4 As substantial remodeling of motor units may occur between 2 and 6 months after stroke, 4 it might be possible to influence this process with therapies directed toward increasing muscle strength and thus functional ability. A systematic review of muscle strength training after stroke has found positive effects on both strength and functional activity. 7 However, increases in muscle strength may not translate into improvements in functional activity unless strengthening activity is provided as part of activities such as sit-to-stand 8 ; functional training after stroke might not generalize to untrained tasks 9 ; and both functional and strength training may be beneficial. 10 It might be appropriate therefore to combine functional training and muscle strength training into an intervention called functional strength training (FST). Preliminary evidence suggests that lower limb task orientated progressive resistance strength training may be effective. 11 However, those participants allocated to the control group received no therapy 11 and therefore it is not possible to infer that FST is more effective than conventional physical therapy (CPT). Comparing FST with CPT is not straightforward. Although CPT is underevaluated, it is well established in clinical practice within stroke units that are known to enhance recovery. Withholding CPT from a group of people receiving FST in a clinical trial would not be acceptable in clinical practice. Within a trial, all participants need to receive routine CPT; however, a comparison of CPT and FST + CPT is inappropriate because extra therapy (intensity) might enhance outcome after stroke. 12 The hypothesis to be tested therefore, is that FST for the paretic upper limb plus the standard amount of CPT produces greater improvements in motor impairment and functional ability than either (a) extra CPT or (b) the standard amount of CPT. Before this hypothesis can be tested in a phase III trial it is important to From Clinical Development Sciences, St George s University London, London (CD, SM); University of Manchester, Stockport (RT); School of Psychology, University of Nottingham, Nottingham (AS); Institute of Neurology, University College London, London (RL); and Faculty of Health, University of East Anglia, Norwich (VMP), United Kingdom. Address correspondence to Valerie M. Pomeroy, PhD, Faculty of Health, University of East Anglia, Norwich NR4 7TJ, UK. v.pomeroy@uea.ac.uk. 389

2 390 Neurorehabilitation and Neural Repair determine whether there is sufficient evidence of efficacy to justify a phase III study; provide an estimate of the feasibility of subject recruitment; assess whether subjects could be blinded to group allocation; and provide data to inform a power calculation together with an estimate of likely attrition rate to estimate sample size for a phase III trial. Design Method This study was an observer-blind, randomized, controlled phase II trial (Figure 1). Ethics, Randomization, and Subject Blinding Ethical approval was granted by the local ethical committee. Before the study began, the random allocation order was generated using a computer program. Allocation was stratified by baseline Action Research Arm Test (ARAT) score 13 (scores 4 to 30 and 31 to 57) in blocks of 3 within each stratum. Group allocation was concealed in sequentially numbered sealed opaque envelopes held by an independent researcher. Envelopes were opened in response to a telephone request from the research physiotherapist (blinded to measures) after the assessor (blinded to group allocation) had completed baseline measures. Subjects allocated to the control group were told that they would not be receiving extra therapy. Subjects allocated to experimental groups were advised that they were to be offered extra therapy but given no information about type of extra therapy. This procedure was used as blinding subjects to which exercise intervention they are receiving is considered to be difficult if not impossible in rehabilitation trials as subjects can see and experience the type of exercise in which they participate. It was expected that subjects would realize if they were not receiving any extra therapy, and they were therefore provided with this information. Those allocated to either of the experimental groups were aware they were receiving extra therapy but were not told which type would be provided. Underlying this procedure was the expectation that it might be possible in subsequent trials to blind subjects to type of therapy but not possible to blind them to whether or not they were receiving extra therapy. Subjects Subjects in this study gave written informed consent and met the following criteria: infarction of the anterior cerebral circulation (diagnosed through neuroimaging) between 1 week and 3 months after stroke; some voluntary muscle activity in the paretic upper limb, scoring 4+/57 on the ARAT but unable to complete the Nine Hole Peg Test (9HPT) 14 in 50 seconds or less; no obvious unilateral visuospatial neglect on clinical observation of subject s ability to orientate toward objects and people in their environment; able, prior to their stroke, to use Allocation Outcome measures Follow-up measures the paretic upper limb to lift a cup and drink from it; able to follow a one-stage command ( touch your nose with your stronger hand ); able to participate in routine therapy. Procedure Figure 1 Subject Recruitment and Attrition Flowchart Allocated to CPT Received allocated intervention measures and analyzed n = 8 Lost to outcome (n = 2) of which: new stroke n = 1 bail n = 1 measures n = 5 Lost to follow-up (n = 5) of which: unwell n = 3 abroad n = 1 bail n = 1 Assessed for eligibility (n = 371) Eligible n = 35 Baseline measures n = 30 Randomized n = 30 Allocated to CPT+CPT Received allocated intervention measures and analyzed n = 10 Lost to outcome (n = 0) measures n = 6 Lost to follow-up (n = 4) of which: unwell n = 1 died n = 2 moved house n = 1 Excluded: n = 336 Of those: Not meeting inclusion criteria: n = 331 Out of area: n = 5 Refused consent: n = 5 Allocated to CPT+FST Received allocated intervention measures and analyzed n = 10 Lost to outcome (n = 0) measures n = 8 Lost to follow-up (n = 2) of which: unwell n = 1 abroad n = 1 Abbreviations: CPT, conventional physical therapy; FST, functional strength training. Baseline measures were made by the research assessor. Intervention began on the following day and was provided for up to 1 hour, 4 days a week for 6 weeks (24 hours). Subjects repeated the measurement battery, on completion of the intervention phase (outcome), and 12 weeks thereafter (follow-up). Every effort was made to invite subjects for assessment at

3 Donaldson et al / Functional Strength Training: Motor Recovery 391 outcome and follow-up even if they had withdrawn from therapy (intention-to-treat principle). Intervention All subjects received CPT delivered by the clinical physiotherapists (see control condition below). All extra physical therapy, both CPT and FST, was delivered by the research physiotherapist (see experimental conditions below). To minimize the potential confounder of unintentional therapist bias all extra therapy was administered using standardized treatment schedules and the research physiotherapist was supervised closely throughout the course of the trial. To avoid potential bias of the clinical therapists the research physiotherapist provided extra intervention independently to the clinical team by not telling clinical therapists to which group subjects had been allocated and ensured that all extra treatment was delivered in an area where the clinical team were unable to observe which type of treatment was being provided. The control condition was CPT given by the physiotherapists in each clinical center using a standardized treatment schedule 15 (treatment recording form and descriptive booklet) generated using established methods. 16,17 The format of the CPT treatment schedule was similar to that produced for the lower limb 16 and the content included soft tissue mobilization, facilitation of muscle activity/movement, positioning, and education for patient/carer. 17 The CPT treatment schedule emphasizes interventions provided by a therapist facilitating and guiding movement (therapist hands-on) to provide sensory input to optimize joint alignment in preparation for voluntary movement. Key differences between CPT and FST are outlined in the Appendix. Experimental condition 1 was CPT plus extra CPT given by the research physiotherapist (CPT + CPT). The additional CPT was recorded using the treatment schedule. 15 Experimental condition 2 was CPT plus FST given by the research physiotherapist (CPT + FST). Delivery of FST gave prominence to: directing the subject s attention to the exercise/ activity being performed, 18 appropriate verbal feedback on performance, 18 repetition, and goal-directed functional activity (therapist hands-off). FST was based on the key elements of normal upper limb function, that is, on positioning the hand and then using it to manipulate objects. The focus was on improving the power of shoulder/elbow muscles to enable appropriate placing of the hand; improving the production of appropriate force in different muscles to achieve the specific task; and on specific interventions for the wrist and finger muscles to maximize ability to manipulate objects. The initial level of resistance was the maximum load that still permitted 5 repetitions of movement/action through the available range of muscle length. Treatment was progressed using repetition, altering the size and weight of items, and using heavier weights. The content of FST was divided into muscle group specific movements, for example, elbow flexion/extension; upper limb gross movement patterns underlying functional activity (eg, shoulder flexion/abduction/external rotation + elbow flexion); hand reaching/retrieval activity (eg, reaching to a shelf while seated); hand grip activities; hand manipulation involving entire everyday activities and using objects such as screw top canisters, pegs, food items (eg, bag of dried pasta, tin of baked beans), shoe laces, mugs, and pens. Activities included using the paretic upper limb to place different food items into a shopping bag and then lift the bag onto a shelf; tighten/loosen nuts/bolts; and taking a pen cap on and off. Measurement Battery Outcome measures were undertaken with subjects seated in an adjustable upright chair ( Handicare Cirrus Jubilee ), which allowed for a posture in which knees and hips were maintained at 90. A D-shaped height-adjustable table (width 100 cm, depth 76 cm) was placed in front of the subject so arms were supported on the table with elbows directly below glenohumeral joint. Choice of outcome measures was informed by the need to measure both muscle strength and functional ability; to use measures employed in published trials of upper limb stroke rehabilitation; and to consider potential fatigue of subjects that could result from undertaking a large battery of tests. We were therefore concerned to compile a measurement battery that would provide sufficient clinically relevant data allowing the hypothesis to be tested on patients studied early after stroke. The primary outcome measure was 1. Upper limb function Action Research Arm Test (ARAT). 13 The secondary outcome measures were 2. Dexterity Nine Hole Peg Test (9HPT). 14 The time limit was set at 50 seconds and 1 trial was conducted. 3. Upper limb strength with a myometer (MIE Medical Research Ltd, Leeds, UK) held in a secure position with a purpose-designed clamp. Force values were obtained during 3 trials, with the greatest value obtained used for data analysis. Tasks were a. Hand grip force and pinch grip force the upper limb position for both pinch and grip force was standardized 19 and the myometer was set to zero after the subject was positioned with his or her hand/digits around the bars, at rest. The assessor instructed the subject to squeeze as hard as you can. b. Isometric elbow flexion and extension force measured with the elbow positioned at 90. The cuff of the myometer was positioned around the wrist, proximal to the ulnar styloid process and with angle of pull at 90 to the forearm. The assessor instructed the subject to pull/ push as hard as you can. 4. Blinding of subjects to group allocation (outcome only) using a questionnaire requiring subjects to indicate whether they: had been in the group who received FST as extra therapy; CPT as extra therapy; no extra therapy; or were unsure which treatment group they had been in. Sample Size A total of 30 subjects (10 per group) were chosen as this sample size allowed estimation of standard deviations to inform a power calculation for a subsequent phase III clinical trial.

4 392 Neurorehabilitation and Neural Repair Analysis and Interpretation Change scores, baseline to outcome, and baseline to follow-up, were compared between groups. Where subjects withdrew before either outcome or follow-up they were not included in the analysis (ie, imputation of last observed response was not used). Data, for this exploratory phase II trial, were interpreted with reference to clinically important differences. These were set as 5.7 points for ARAT 20 ; 1 peg in 50 seconds (0.02 pegs per second) for 9HPT (arbitrary threshold as lack of published data); and 10 N in muscle force (equivalent of 1 kg; arbitrary threshold as lack of published data). Although statistical analysis is not essential for this pilot study we did examine differences between groups with the Kruskal Wallis test (data were not normally distributed). A power calculation to estimate sample size for a subsequent phase III trial used ARAT data, clinically important difference of 5.7 point, 20 was undertaken based on 2 independent samples in a 2-way test using the standard deviation found for the CPT group with 90% power and 5% significance. Blinding procedures were considered adequate for a subsequent double-blind randomized controlled trial if 25% or less of subjects correctly identified their group allocation. Results A total of 371 stroke patients were screened of whom 35 (9%) met the study criteria. Five potential subjects did not give informed consent; 30 subjects (8%) were recruited (Figure 1). Subjects baseline characteristics are given in Table 1. In summary, their mean age was 72.8 years (standard deviation, SD = 11.9) and 17/30 were women. The mean time after stroke was 20.2 days (SD = 14.0). At baseline, the CPT + FST group tended to have lower values than the other 2 groups for all measures, but these differences were not statistically significant (Kruskal Wallis test): ARAT P =.47; 9HPT P =.61; hand grip force P =.49; pinch force P =.40; elbow flexion force P =.10; and elbow extension force P =.61. Two subjects, both from the control group, were unable to complete outcome measures. One subject was bailed and another had suffered a second stroke. The attrition rate at outcome was therefore 7% (Figure 1). At follow-up, an additional 9 subjects were unable to attend for measurement. Two had died, 4 were medically unwell, 2 had moved house, and 2 were abroad. The attrition rate at follow-up was 40.0% (Figure 1). Table 2 shows the amount of upper limb physical therapy received by each subject who completed outcome measures. There was no obvious difference in (a) the median (interquartile range, IQR) hours of routine CPT provided or (b) the median (IQR) hours of experimental therapy provided. There was a difference, however, in the total hours of therapy received by the 2 experimental groups compared with the control group with median (IQR) values of 2.81 (3.7) hours for CPT; 13.8 (7.1) hours for CPT + CPT; and 17.7 (7.5) hours for CPT + FST. Blinding of Subjects to Allocated Intervention The majority of subjects (68%) who completed outcome measures were unsure as to which group they had been allocated (CPT 75%, CPT + CPT 60%, CPT + FST 70%; Table 3). Only 4 of the 28 subjects (14%) correctly identified the treatment they received. Even in the CPT group who had been told that they would receive no extra therapy, only 1 person correctly identified their grouping. Efficacy Results are presented in the text in terms of descriptive statistics and clinically important differences only as the present exploratory trial is designed as a precursor to a definitive phase III trial. Results of a Kruskal Wallis test are, however, included in Tables 4 and 5 but need to be interpreted cautiously because this phase II study was not sufficiently powered to test the hypothesis. Primary Outcome Measure The Action Research Arm Test. The CPT + FST group showed the largest change in ARAT score, median 19.5 (IQR 22) with the smallest median change, 8.0 (IQR 13.25), in the CPT + CPT group (Table 4). The CPT + FST group increased median score by 8.0 points in excess of the CPT group, which is above the clinically important difference of 5.7 points (Table 4). Secondary Outcome Measures The Nine Hole Peg Test. The median values found for the 9HPT suggest a trend toward a greater improvement in the CPT + FST group with a median change of 0.11 (IQR 0.27) pegs/second, which is 0.3 pegs/second higher than the median change in the CPT group of 0.08 (IQR 0.17) and above the level of clinical importance (Table 4). Hand grip force. The median change was found to be highest for the CPT group, 28.0 N (IQR 47.0), and lowest for the CPT + CPT group, 10.5 N (IQR 40.25; Table 5). Pinch grip force. The largest median change was found for the CPT+ FST group, 19.0 N (IQR 19.75), and the smallest for the CPT + CPT group, 9.0 N (IQR 13.75; Table 5). The difference between the 2 higher scoring groups was below the set clinically important level. Isometric elbow flexion force. The median change values shown in Table 5 suggest a trend toward a greater change for the CPT + FST group (32.5, IQR N). The smallest change was found for the CPT group (5.5, IQR N). The difference between the 2 experimental groups was 17.4 N (Table 5), which is 174% of the set clinically important change.

5 Donaldson et al / Functional Strength Training: Motor Recovery 393 Table 1 Subject Characteristics at Baseline Days Elbow Elbow Subject Age Since Lesion Grip Pinch Flexion Extension Number (Years) Gender Stroke Area ARAT 9HPT (N) (N) (N) (N) CPT 2 a 63 Female 18 R MCA Female 11 L MCA b 44 Male 14 L MCA Male 16 L MCA Male 8 L MCA Male 9 L MCA b 56 Female 11 R MCA Male 10 R MCA a 90 Female 22 R MCA a 84 Female 15 R MCA Mean/ratio : CPT + CPT 1 81 Female 20 L MCA Female 12 L MCA Female 36 R MCA a 70 Male 15 R MCA a 74 Male 21 L MCA a 71 Male 21 L MCA a 64 Female 22 R MCA Male 7 R MCA Female 47 R MCA Male 55 R MCA Mean : CPT + FST 5 89 Female 12 L MCA Female 7 L MCA Female 8 L MCA a 73 Male 24 R MCA a 85 Female 25 R MCA Female 7 R MCA Male 12 R MCA Female 28 L MCA Female 61 R ACA Male 33 L MCA Mean : Abbreviations: ARAT, Action Research Arm Test; 9HPT, Nine Hole Peg Test; CPT, conventional physical therapy; FST, functional strength training; MCA, middle cerebral artery; ACA, anterior cerebral artery; R, right; L, left. a Subject withdrew between outcome and follow-up. b Subject withdrew before outcome. Isometric elbow extension force. The smallest median change was found for the CPT + CPT group, 9.0 N (IQR 27.75), and the largest for the CPT + FST group, 13.5 N (IQR 36.5). However, this difference of 4.5 N is below the set level of clinical important change. Estimate of Sample Size for Subsequent Definitive Phase III Trial The sample size is estimated as 93 per group (279 for 3-group trial) to determine a difference at outcome. With this estimated sample size a subsequent phase III trial would be able to detect a difference of 0.4 pegs/second for 9HPT, 7.0 N for pinch force, 13.8 N for hand grip force, 22.2 N for elbow flexion force, and 10.4 N for elbow extension force. Discussion This exploratory phase II trial found a trend for enhanced motor recovery for the CPT + FST group for all measures except hand grip force. The improvements found achieved set clinical importance for ARAT, 9HPT, and isometric elbow flexion force. It is possible that the clinically important advantage seen in the CPT + FST group was due to their receiving 3.9 hours more therapy than the CPT + CPT group (Table 2). Thus the CPT + FST group received close to the 16 hours of extra therapy that has been suggested as being required to enhance motor recovery. 13 This could have been influenced by the research physiotherapist having an unintentional bias toward providing more FST despite the close supervision provided. Arguing against this

6 394 Neurorehabilitation and Neural Repair Table 2 Duration of Upper-Limb Treatment Recorded for Each Subject by Groups Group Subject ID CPT (Hours) Extra CPT (Hours) FST (Hours) Total Treatment (Hours) CPT a a Mean (SD) = 11.3 Mean (SD) = (3.7) (3.7) 18 a a 22 0 b b CPT + CPT 1 0 b b Mean (SD) = 10.2 Mean (SD) = 10.2 Mean (SD) = (3.0) (4.9) (7.1) CPT + FST Mean (SD) = 11.4 Mean (SD) = 12.1 Mean (SD) = (4.4) (4.2) (7.5) Abbreviations: SD, standard deviation; CPT, conventional physical therapy; FST, functional strength training. a Subject withdrew before outcome. b Subject received no upper limb CPT from physiotherapists in clinical setting. Table 3 Subject Awareness of Treatment Group Allocation Within Groups CPT (%); CPT + CPT CPT + FST Overall (%); Response n = 8 a (%); n = 10 (%); n = 10 n = 28 Correct Incorrect Unsure Data missing Abbreviations: CPT, conventional physical therapy; FST, functional strength training. a Two subjects withdrew before outcome measurement. possibility is the finding that for all measures except isometric elbow extension force, the CPT group showed greater improvement than the CPT + CPT group despite receiving 10.9 hours less therapy. It is possible therefore that intensity of therapy alone does not enhance motor recovery but that the type of therapy is important. This possibility is supported by findings of earlier studies of intensity, which showed no improvement with increased intensity of conventional, Bobath based therapy 21,22 and enhanced recovery with task-specific exercises. 23 The subsequent phase III trial will therefore also need to include 3 groups (2 experimental and 1 control). Another possible reason for the clinically important advantage seen in the CPT + FST group was that they started at a lower baseline. The subsequent phase III trial will consider this possibility by using minimization in the randomization order to minimize imbalance between experimental groups. For example, the Pocock and Simon s range method could be used to determine the allocation of each individual by the ARAT categorized as score 39 to 57 (57 = able to complete all items normally); score 20 to 38 (38 = able to complete all items albeit slowly/abnormally); and score 0 to 19 (4 = able to complete 4 items partially and 19 = able to complete all items partially). Exact details of the minimization procedure will be devised in partnership with a statistician working in a Clinical Trials Unit and will be subject to peer review and possible amendment through a funding application process. The design of a subsequent phase III trial needs to consider the recruitment and attrition rate found in the current exploratory trial together with the results of the power calculation.

7 Donaldson et al / Functional Strength Training: Motor Recovery 395 Table 4 Change Baseline to Outcome and Baseline to Follow-up for Functional Ability: ARAT and 9HPT Mean (SD) and [Median (IQR)] Median (IQR) Change Baseline Outcome Baseline to Outcome KW P Value ARAT CPT (n = 8) (13.07) [28.0 (17.0)] (13.93) [50.5 (26.0)] (21.00) CPT + CPT (15.52) [34.5 (26.0)] (17.83) [52.5 (30.0)] 8.00 (13.25).232 CPT + FST (9.80) [27.0 (11.0)] (18.90) [51.5 (24.0)] (22.00) 9HPT CPT (n = 8) 0.53 (0.09) [0.00 (0.10)] 0.15 (0.14) [0.10 (0.29)] 0.08 (0.17) CPT + CPT 0.57 (0.08) [0.10 (0.13)] 0.16 (0.15) [0.13 (0.29)] 0.05 (0.22).928 CPT + FST 0.03 (0.06) [0.00 (0.07)] 0.17 (0.15) [0.15 (0.34)] 0.11 (0.27) Abbreviations: ARAT, Action Research Arm Test; 9HPT, Nine Hole Peg Test; SD, standard deviation; IQR, interquartile range; CPT, conventional physical therapy; FST, functional strength training; KW, Kruskal Wallis test. Table 5 Change Baseline to Outcome and Baseline to Follow-up for Grip Force (N), Pinch Force (N), Elbow Flexion (N), and Elbow Extension (N) Mean (SD) and [Median (IQR)] Median (IQR) Change Baseline Outcome Baseline to Outcome KW P Value Grip force (N) CPT (n = 8) (25.90) [23.0 (52.0)] (39.25) [66.0 (66.0)] (47.00) CPT + CPT (45.218) [46.5 (91.0)] (49.45) [83.5 (99.0)] (40.25).522 CPT + FST (20.42) [21.5 (31.0)] (60.18) [43.5 (69.0)] (44.00) Pinch force (N) CPT (n = 8) 9.25 (13.45) [4.5 (12.0)] (19.70) [22.5 (40.0)] (29.00) CPT + CPT (18.81) [25.0 (34.0)] (23.11) [26.5 (37.0)] 9.00 (13.75).599 CPT + FST (n =10) 9.40 (13.17) [1.0 (26.0)] (21.26) [23.0 (45.0)] (19.75) Elbow flexion force (N) CPT (n = 8) (33.72) [44.0 (67.0)] (38.67) [78.0 (69.0)] 5.50 (91.25) CPT + CPT (52.47) [54.0 (75.0)] (58.73) [64.5 (64.0)] (25.25).702 CPT + FST (25.56) [29.0 (49.0)] (44.69) [60.0 (60.0)] (55.75) Elbow extension force (N) CPT (n = 8) (31.33) [58.5 (54.0)] (39.61) [71.5 (74.0)] (36.50) CPT + CPT (61.16) [36.0 (73.0)] (44.56) [46.0 (43.0)] 9.00 (27.75).604 CPT + FST (23.11) [34.0 (34.0)] (34.19) [46.0 (47.0)] (36.50) Abbreviations: SD, standard deviation; IQR, interquartile range; CPT, conventional physical therapy; FST, functional strength training. The recruitment rate was 8% of stroke patients admitted to the study center, which is comparable with other published stroke rehabilitation trials focused on the upper limb, for example, 3%, 23 9%, 24 11%, 10 and 22%. 21 The attrition rate at outcome was reasonable (7%) but rose to 40% by follow-up, which was higher than that seen in other studies, for example, 12% 23 and 25%. 10 The reasons for these differences are unclear, but withdrawals in the present trial were not apparently linked to the interventions provided as attrition was similar across groups (Figure 1). Such an attrition rate therefore needs to be considered in the design of subsequent trials. If sufficient time was built into the timing of follow-up, for example 12 ± 2 weeks then it is expected that subsequent trials should be able to undertake follow-up measures on some of the subjects who were unwell, had moved house, or were abroad at the time of measurement. In this way, it is possible that attrition at follow-up could be reduced to 15% to 20% based on the findings of the present trial and considering that found in other trials. 10,23 Using a 20% attrition rate suggests that if a subsequent phase III trial is to be powered to detect real differences between groups at follow-up, then 335 subjects would be required. Based on recruitment data from the present pilot trial, approximately 4188 patients would need to be screened to recruit a homogenous group of 335 well-characterized subjects. Hence, a multicenter trial would be essential. Estimating an admission rate of 400 to 500 patients per year at each clinical center would therefore require 4 or 5 clinical centers to ensure that data collection was completed within a 3-year period. The recruitment rate might be increased if the study criteria were made less stringent. For example, people with stroke in the anterior circulation resulting from hemorrhage could be included in

8 396 Neurorehabilitation and Neural Repair subsequent trials and it might also be possible that extra therapy, either CPT or FST, is beneficial for people with obvious unilateral spatial neglect or stroke in the territory of the posterior circulation. Therefore, before a subsequent phase III trial is undertaken, further preliminary research should investigate whether more stroke survivors may benefit from FST. It was interesting that the majority (63%) of subjects were unsure which group they were allocated to and only 13.3% of subjects correctly identified their group. This was not found solely in the 2 experimental groups. Control subjects were also unsure of their allocated group even though at the time of allocation they were told that they would not receive any extra therapy. These findings indicate that in subsequent trials it may be possible for subjects to be blinded as to which of the 3 interventions they have been allocated, not just the experimental interventions. Although the findings of the present trial indicate a trend in favor of clinically important improvements in the CPT + FST group, the sample size was small and variation around the mean was relatively large. Consequently, it is difficult to judge whether the optimal dose of extra CPT or FST was provided in the present trial. Of interest is that neither experimental group received the 16 hours of extra therapy that have been suggested as being needed to enhance motor recovery. 13 It may be advisable therefore to undertake a dose-finding investigation before embarking on a definitive phase III trial. This is particularly pertinent considering the lack of such dose-finding studies in rehabilitation research. 25 In summary, this exploratory trial justifies future trials comparing CPT with CPT + CPT and CPT + FST. Research is needed to continue to build toward a robust multicenter, double-blind randomized phase III trial that tests optimal dose of both experimental interventions, considers the potential problems of using the same therapist to deliver both forms of experimental therapy, and uses minimization procedures to address potential imbalance between groups. Such a trial would be based on a clear rationale for treatment; would provide explicit definition of interventions to enable replication in further trials and transferability of results into clinical practice; would include outcome directly relevant to the aims of the intervention; would be double-blind; would require sufficient funding to employ sufficient physiotherapists to avoid the potential problems that could arise from the same physiotherapist delivering both experimental interventions; and would use an intentionto-treat analysis. In this way, the methodological flaws observed in many stroke rehabilitation trials would be avoided. 26 Acknowledgments We gratefully acknowledge funding provided by The Wellcome Trust for this trial and the technical assistance provided by Ashraf Sameja. We are also grateful for the assistance with undertaking blinded measures from Andy Hiscock and Catherine Coleman (Clinical Research Training Fellows funded by the Dowager Countess Eleanor Peel Trust). Appendix Outline of Conventional and Functional Strength Training Treatment Approaches Type of Therapy Conventional Functional Strength Training Description Therapist hands-on therapy with an emphasis on Therapy incorporating specific functional tasks or specific of therapy preparation and joint alignment via sensory input. The movements, in preparation for functional tasks using a patient is encouraged to practice grasp and release and therapist hands-off approach while maintaining patient reaching for objects or spatially, but not as part of a safety. Verbal prompting is used rather than using sensory functional task, that is, reaching for a cup to drink. feedback by the therapist. Activities are progressed using Instead, for example, they would reach for a cone repetition, size, and resistance and to move it from one table to another Progression of therapy Repetition 5 repetitions of the same specific task One or more sets of 10 repetitions of the same specific task with up to 5 sets of 10 repetitions Resistance No systematic progression in treatment activities Systematic progression in treatment activities from light progression although different weights of objects may be used for to heavy objects/resistance, starting with a resistance treatment activities that allows a 5-repetition maximum load throughout the whole available range of the muscle length. Progressed when the patient can achieve 50 repetitions (5 sets of 10) against the initial load Size Different sizes of objects may be used in treatment Systematic progression from medium to very small, or progression activities but with no systematic progression to make medium to very large objects, so that tasks are the activity more challenging progressively more challenging

9 Donaldson et al / Functional Strength Training: Motor Recovery 397 References 1. Nakayama H, Jorgensen S, Raaschou HO, Olsen TS. Recovery of upper extremity function in stroke patients: the Copenhagen Stroke Study. Arch Phys Med Rehabil. 1994;75: Burke D. Spasticity as an adaptation to pyramidal tract injury. Adv Neurol. 1988;47: Landau WM. Spasticity: the fable of a neurological demon and the emperor s new therapy. Arch Neurol. 1974;31: Patten C, Lexell J, Brown HE. Weakness and strength training in persons with poststroke hemiplegia: rationale, method and efficacy. J Rehabil Res Dev. 2004;3A: Kamper DG, Fischer HC, Cruz EG, Rymer WZ. Weakness is the primary contributor for finger impairment in chronic stroke. Arch Phys Med Rehabil. 2006;87: Harris JE, Eng JJ. Paretic upper-limb strength best explains arm activity in people with stroke. Phys Ther. 2007;87: Ada L, Dorsch S, Canning CG. Strengthening interventions increase strength and improve activity after stroke: a systematic review. Aust J Physiother. 2006;52: Bohannon RW. Muscle strength and muscle training after stroke. J Rehabil Med. 2007;39: Dean CM, Shepherd RB. Task-related training improves performance of seated reaching tasks after stroke. A randomised controlled trial. Stroke. 1997;28: Winstein CJ, Rose DK, Lewthwaite R, Chui HC, Azen SP. A randomized controlled comparison of upper-extremity rehabilitation strategies in acute stroke: a pilot study of immediate and long-term outcomes. Arch Phys Med Rehabil. 2004;85: Yang Y-R, Wang R-Y, Lin K-H, Chu M-Y, Chan R-C. Task-orientated progressive resistance strength training improves muscle strength and functional performance in individuals with stroke. Clin Rehabil. 2006;20: Kwakkel G, van Peppen R, Wagenaar RC, et al. Effects of augmented exercise therapy time after stroke. A meta-analysis. Stroke. 2004;35: Yozbatiran N, Der-Yeghiaian L, Cramer SC. A standardized approach to performing the Action Research Arm Test. Neurorehabil Neural Repair. 2008;22: Mathiowetz V, Kashman N, Volland G, Weber K, Dowe M, Rogers S. Grip and pinch strength: normative data for adults. Arch Phys Med Rehabil. 1985;66: Donaldson C, Pomeroy V, Tallis RC. A treatment schedule to record current physical therapy interventions for the upper limb after stroke: validity and reliability. Cerebrovasc Dis. 2007;23(suppl 2): Pomeroy VM, Cooke E, Hamilton S, Whittet A, Tallis RC. Development of a schedule of current physiotherapy treatment used to improve movement control and functional use of the lower limb after stroke: a precursor to a clinical trial. Neurorehabil Neural Repair. 2005;19: Hunter S, Crome P, Sim J, Donaldson C, Pomeroy VM. Development of treatment schedules for research: a structured review to identify methodologies used and a worked example of mobilisation and tactile stimulation for stroke patients. Physiotherapy. 2006;92: Robertson IH, Ridgeway V, Greenfield E, Parr A. Motor recovery after stroke depends on intact sustained attention: a 2-year follow-up study. Neuropsychology. 1997;11: Pryce JC. Wrist position between neutral and ulnar deviation that facilitates maximum power grip strength. J Biomech. 1980;13: Van der Lee JH, De Groot V, Beckerman H, Wagenaar RC, Lankhorst GJ, Bouter LM. The intra- and interrater reliability of the Action Research Arm Test: a practical test of upper extremity function in patients with stroke. Arch Phys Med Rehabil. 2001;82: Lincoln NB, Parry RH, Bass CD. Randomized, controlled trial to evaluate increased intensity of physiotherapy treatment of arm function after stroke. Stroke. 1999;30: Rodgers H, Mackintosh J, Price CI, et al. Does an early increased-intensity interdisciplinary upper limb therapy programme following acute stroke improve outcome? Clin Rehabil. 2003;17: Kwakkel G, Wagenaar RC, Twisk JWR, Lankhorst GJ, Koestier JC. Intensity of leg and arm training after primary middle-cerebral-artery stroke: a randomized trial. Lancet. 1999;354: Higgins J, Salbach NM, Wood-Dauphinee S, Richards CL, Cote R. Mayo NE. The effect of a task-orientated intervention on arm function in people with stroke: a randomised controlled trial. Clin Rehabil. 2006;20: Dobkin BH. Confounders in rehabilitation trials of task-orientated training: lessons from the designs of the EXCITE and SCILT multicenter trials. Neurorehabil Neural Repair. 2007;21: Pomeroy VM, Tallis RC. Neurological rehabilitation: a science struggling to come of age. Physiother Res Int. 2002;7: For reprints and permission queries, please visit SAGE s Web site at