Authors: Jan W. van der Scheer, MSc Sonja de Groot, PhD Riemer J.K. Vegter, MSc Johanneke Hartog, MSc Marga Tepper, MD Hans Slootman, MD ALLRISC Group DirkJan H.E.J. Veeger, PhD Lucas H.V. van der Woude, PhD

Spinal Cord Injury

ORIGINAL RESEARCH ARTICLE

Affiliations: From the Center for Human Movement Sciences, University of Groningen, University Medical Center Groningen, the Netherlands (JWvdS, SdG, RJKV, JH, LHVvdW); Amsterdam Rehabilitation Research Center|Reade, Amsterdam, the Netherlands (SdG); Department of Rehabilitation Medicine, Center for Rehabilitation, University of Groningen, University Medical Center Groningen, the Netherlands (MT, LHVvdW); Heliomare, Rehabilitation Center, Wijk aan Zee, the Netherlands (HS); Research Institute MOVE, Faculty of Human Movement Sciences, VU University Amsterdam, the Netherlands (DHEJV); and Department of Mechanical, Maritime and Materials Engineering, Delft University of Technology, Delft, the Netherlands (DHEJV). A complete list of the collaborators in the ALLRISC group is provided at the end of this article.

Correspondence: All correspondence and requests for reprints should be addressed to: Jan W. van der Scheer, MSc, Center for Human Movement Sciences, University Medical Center Groningen, Sector F; PO Box 196, 9700 AD, Groningen, the Netherlands.

Disclosures: The study was part of the research program ALLRISC (www.scionn.nl), supported financially by Fonds NutsOhra under the responsibility of ZonMw (project number 89000006). The authors certify that they have no affiliations with or involvement in any organization or entity with any interest in the subject matter or materials discussed in this article. Financial disclosure statements have been obtained, and no conflicts of interest have been reported by the authors or by any individuals in control of the content of this article.

Low-Intensity Wheelchair Training in Inactive People with Long-Term Spinal Cord Injury A Randomized Controlled Trial on Propulsion Technique ABSTRACT van der Scheer JW, de Groot S, Vegter RJK, Hartog J, Tepper M, Slootman H, Veeger DHEJ, van der Woude LHV: Low-intensity wheelchair training in inactive people with long-term spinal cord injury: a randomized controlled trial on propulsion technique. Am J Phys Med Rehabil 2015;94:975Y986.

Objective: The objective of this study was to investigate the effects of a lowintensity wheelchair training on propulsion technique in inactive people with longterm spinal cord injury.

Design: Participants in this multicenter nonblinded randomized controlled trial were inactive manual wheelchair users with spinal cord injury for at least 10 yrs (N = 29), allocated to exercise (n = 14) or no exercise. The 16-wk training consisted of wheelchair treadmill propulsion at 30%Y40% heart rate reserve or equivalent in rate of perceived exertion, twice a week, 30 mins per session. Propulsion technique was assessed at baseline as well as after 8, 16, and 42 wks during two submaximal treadmill-exercise blocks using a measurement wheel attached to a participant’s own wheelchair. Changes over time between the groups were analyzed using Mann-Whitney U tests on difference scores (P G 0.05/3).

Results: Data of 16 participants could be analyzed (exercise: n = 8). Significant differences between the exercise and control groups were only found in peak force after 8 wks (respective medians, j20 N vs. 1 N; P = 0.01; ru = 0.78).

Conclusions: Significant training effects on propulsion technique were not found in this group. Perhaps, substantial effects require a higher intensity or frequency. Investigating whether more effective and feasible interventions exist might help reduce the population’s risk of upper-body joint damage during daily wheelchair propulsion. Key Words:

Wheelchairs, Spinal Cord Injuries, Motor Skills, Upper Extremity

0894-9115/15/9411-0975 American Journal of Physical Medicine & Rehabilitation Copyright * 2015 Wolters Kluwer Health, Inc. All rights reserved. DOI: 10.1097/PHM.0000000000000289

www.ajpmr.com

Low-Intensity Wheelchair Training in Inactive SCI Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

975

P

eople with spinal cord injury (SCI) often experience upper-body pain and pathology.1Y3 One of the causes might be joint damage during manual wheelchair propulsion, which is the primary form of mobility in many people with SCI.4Y6 Factors increasing the risk for joint damage during wheelchair propulsion include deconditioning and years as a wheelchair user.4,5 High risk for joint damage could therefore be expected in deconditioned or inactive manual wheelchair users with long-term SCI.7 The risk for joint damage during wheelchair propulsion8 may be lower when adopting a more optimal propulsion technique.1,5 This is defined as a technique in which, at a given power output, push frequency and hand-rim forces are minimized while contact angle is maximized.1,9 Absolute joint loads might then be relatively low and occurring at a minimal frequency.1,5 Only limited knowledge exists about developing a more optimal propulsion technique in long-term wheelchair users.10 One way could be exercise-induced fitness improvements when assuming that these imply physiologic adaptations, for example, reduced muscular fatigability, that allow adopting a more optimal propulsion technique.9,11Y13 Evidence is found in long-term wheelchair users: a 6-wk training resulted in improved fitness (increased peak power output and muscular strength) and a more optimal propulsion technique (higher propulsive torques at a lower push frequency).13 The training included high-resistance strength exercises and rowing exercise at 60% heart rate reserve (HRR).13 Such relatively high exercise intensities, however, have been suggested to lead to low adherence, dropout, and musculoskeletal injury in deconditioned or inactive populations.14 For these populations, low-intensity training (30%Y40% HRR) is suggested as a safer and more feasible intervention.14 A 7-wk low-intensity wheelchair training in able-bodied novices (treadmill propulsion at 30% HRR, three times a week, 70 mins per session) resulted in improved mechanical efficiency, a reduced push frequency, and longer push times.15 In inactive people with long-term SCI, it might not be feasible to perform exercise more than twice a week, 30 mins per session.16,17 The authors of this study recently investigated this exercise frequency and duration in a 16-wk low-intensity training in inactive people with long-term SCI but found no significant training effects in mechanical efficiency, peak power output, and other fitness parameters (van der Scheer et al., submitted 2014). However, it is not known how sensitive these parameters are to peripheral physiologic adaptations

976

van der Scheer et al.

that could benefit the propulsion technique. Furthermore, it is not known whether beneficial motor adaptations result from the relatively long propulsion periods in low-intensity wheelchair training.12 Such a stimulus seems absent in the daily life of most wheelchair users.18 The aim of this study was to investigate the effects of a low-intensity wheelchair training on propulsion technique in inactive manual wheelchair users with long-term SCI. A multicenter randomized controlled trial (RCT) was conducted, in which an exercise group was compared with a nonexercising control group. The exercise group performed 16 wks of wheelchair propulsion on a treadmill at 30%Y40% HRR or equivalent in rate of perceived exertion (RPE), twice a week, 30 mins per session (14). On the basis of previous training studies,13,15,19 it was hypothesized that the exercise group would show a more optimized propulsion technique than the control group.

METHODS Design A nonblinded RCT was conducted in two rehabilitation centers. The exercise group performed a 16-wk low-intensity wheelchair training, whereas no intervention was used in the control group.12 Measurements were conducted in both groups at baseline (T1), 8 wks after baseline (T2), 16 wks after baseline (T3), and 42 wks after baseline (T4). The design of the RCT is described elsewhere in more detail.12

Ethical Approval Ethical approval was obtained from committees of the VU University Medical Center (Amsterdam, the Netherlands) and the two participating rehabilitation centers (Heliomare, Wijk aan Zee and University Medical Center Groningen, Groningen, the Netherlands).12 The RCT was registered in the Dutch Trial Register (www.trialregister.nl, NTR3037).

Participants Participants were inactive manual wheelchair users with SCI for longer than 10 yrs, who were included after providing written informed consent. Paramedical research assistants preselected potentially eligible people using patient databases (Fig. 1). The participants were allocated to the exercise or control group by the investigators on the basis of a blocked randomization procedure using sealed envelopes.12 Inclusion criteria comprised time since injury of longer than 10 yrs and inactivity as defined by a reference score on the Dutch Physical Activity Scale

Am. J. Phys. Med. Rehabil. & Vol. 94, No. 11, November 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

FIGURE 1 Flow through inclusion, randomization, allocation, measurements, and intervention. 1Estimated three in four eligible people declined to participate. 2Extended to a maximum of 18 wks if participants missed session(s). 3Fitness and motor abilities too limited for 30 mins per session; protocols were therefore individualized to 423 or 766 total mins of exercise. 4HRR was not used owing to impaired autonomic nervous system: n = 6 (completed: n = 5; stopped: n = 1). 5Reasons to stop training: lack of motivation due to personal problems, lack of time due to new work obligations, or kidney stones.

for Individuals with Physical Disabilities (G30 MET hr I wkj1).12,20 Other inclusion criteria of the RCT’s original design12 were adjusted to increase sample size for the present study (age, e65Ye 67 yrs; age at onset SCI, Q18YQ12 yrs). Databases were searched again with the changed criteria 1 yr after inclusion started. Exclusion criteria comprised cardiovascular contraindications for exercise, progressive disease, psychiatric problems, insufficient mastery of www.ajpmr.com

Dutch language, and plans to change lifestyle.12 Another exclusion criterion was musculoskeletal complaints that might worsen because of the training, which was assessed by a paramedic research assistant and physician.

Training The exercise group used their own wheelchair to follow a low-intensity wheelchair training.12 The Low-Intensity Wheelchair Training in Inactive SCI

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

977

protocol consisted of 16 wks of supervised treadmill propulsion in a rehabilitation center, twice a week, 30 mins per session. Relative intensity was 30%Y40% HRR or an equivalent in RPE of 1Y3 on a 10-point scale.17,21 RPE was used instead of HRR in participants with an impaired autonomic nervous system.17 A trained paramedical research assistant continuously monitored relative intensity. It was maintained during exercise by changing treadmill velocity or weight in a pulley system.12 The assistant also monitored musculoskeletal overuse symptoms during and after exercise using a measure of local perceived discomfort.22 Training would be stopped if such symptoms or other adverse events appeared. The participants could move freely over the motor-driven treadmill belt (1.2 m  5.30 or 1.2 m  5.47 m; ForceLink B.V., Culemborg, the Netherlands). The research assistant corrected steering direction if a participant moved the wheelchair too close to the edge of the belt. Protocols consisted of 18 or 24 mins of exercise in the first four sessions and 30 mins of exercise in the subsequent 28 sessions (924 mins of total exercise). Continuous 30-min periods were performed by the most fit and skilled participants, whereas others followed protocols with intermittent exercise (4  7.5 or 10  3 mins with 1Y2 mins of rest intervals).12 These intervals were further individualized in the participants with the lowest fitness levels. If a participant missed a session, it could be made up using an additional session in one of the following weeks or by extending the training period up to 18 wks.

Measurements Standardized measurement protocols12 comprised submaximal exercise blocks and a peak incremental test, similar to previous SCI cohort studies.23,24 All tests consisted of treadmill propulsion in the participants’ own wheelchairs (see BTraining[). Wheelchair configuration was standardized before each measurement occasion.12 The participants were familiarized with treadmill propulsion in a 2-min procedure. Fitness and skill levels were observed to select a treadmill velocity for the submaximal blocks and peak test (G0.56, 0.56, 0.83, or 1.11 m I sj1). The chosen velocity was expected to allow steady-state exercise in the submaximal blocks (respiratory exchange ratio [RER], e1.00). A drag test was conducted, followed by two submaximal blocks of 3 mins at different inclination angles (0% or 0.5%Y0.6%). The blocks were separated by 2 mins of rest. The participants performed both blocks using their rear

978

van der Scheer et al.

right wheel replaced with a force and torqueinstrumented measurement wheel (OptiPush; MAX Mobility, Antioch, TN), whereas the left wheel was replaced with an inertia-compensated dummy wheel. Both wheels weighed 5.7 kg, had a tire pressure of 6  105 Pa, and matched the participant’s wheel diameter (0.61, 0.64, or 0.66 m). The instrumented wheel recorded 3-dimensional forces and torques applied to the hand rim, in addition to the angle over which the wheel rotated (sample frequency, 200 Hz). After the submaximal blocks, the participants rested for 2 mins and then performed a peak incremental test.24 The inclination angle of the treadmill increased every minute by approximately 0.3 degrees until the participant could no longer maintain treadmill velocity. Oxygen uptake and RER were continuously measured during all tests using an Oxycon Delta (CareFusion, San Diego, CA).

Data Processing Forces and torques of the instrumented wheel were low-pass filtered at 20 Hz and angle data at 5 Hz (Butterworth filter).25 Custom-written MATLAB routines23,25 were used to determine power output and parameters of the propulsion technique (push frequency, peak force, contact angle, and push time) during the last 30 secs of the two submaximal blocks (Table 2). These parameters were selected on the basis of the study’s definition of a more optimal propulsion technique1,9 and their sensitivity to change in a previous study on low-intensity wheelchair training.15 Reliability of parameters is indicated by intraclass correlations of 0.87Y0.99 and the low standard error of measurements found in able-bodied people.23 Oxygen uptake and RER were determined during the last 30 secs of the submaximal blocks. Mechanical efficiency was calculated using equations based on oxygen uptake, RER, and power output derived from the drag test.24 Peak power output was determined as the highest power output maintained for at least 30 secs during the peak test (POpeak [W]).24

Statistics Required group sizes were estimated at n = 18. This sample size was based on power calculations using the primary outcome of the RCT (POpeak).12 Many group parameters were not normally distributed; hence, analyses were based on nonparametric statistics. Significance was set at P G 0.05. Fisher’s tests and Mann-Whitney U tests were used to evaluate possible group differences in

Am. J. Phys. Med. Rehabil. & Vol. 94, No. 11, November 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

characteristics and parameters at baseline. The groups recruited in the two centers were compared, as well as the exercise and control groups. The Friedman analysis of variance was used to evaluate changes within the groups over time (T1, T2, and T3). Post hoc evaluation was conducted using Wilcoxon tests. The retention period (T3 vs. T4) was analyzed separately using Wilcoxon tests owing to limitations in statistical power. All analyses included participants without missing data at T1, T2, and T3. Significance in all Wilcoxon tests was Bonferroni-corrected (p/4). Changes between the groups over time were analyzed using Mann-Whitney U tests on difference scores ($T2 j T1, $T3 j T1, and $T4 j T3), whereas significance was Bonferroni-corrected (p/3). The sample selected for the between-group analyses was identical to that in the within-group comparisons. Effect sizes were calculated using nonparametric rank order correlations: rU = 1 Y 2U / (nexercise  ncontrol).26 When rU = 0, an equal number of participants in both groups improved relative to each other. When rU G 0, more participants in the control group had a larger improvement than those in the exercise group. Effect sizes indicating large training effects were considered as rU = 0.80Y1.00, moderate as rU = 0.50Y0.80, and small as rU = 0.20Y0.50. Included in all analyses were participants who did not complete the training but continued performing measurements (intention-to-treat principle). For exploratory purposes, analyses were repeated, excluding participants who did not complete the training (auxiliary analyses).

Measurements

RESULTS

Missing data occurred, for example, owing to the inability to complete 3 mins of consecutive propulsion, technical problems, and incidence of secondary health complications (Fig. 1). Data of eight participants in both groups were available for analyses over T1, T2, and T3 (Tables 1Y3). Baseline characteristics in the group with missing data did not significantly differ from the group available for analyses (Table 1). The examined exercise group did not differ significantly from the control group in baseline characteristics or parameters (Tables 1Y3), except in wheelchair weight (respective medians, 13.7 vs. 19.8 kg; P = 0.014). The examined exercise group had motor lesions levels ranging from C5 to T12 and sensory levels ranging from C5 to L3, whereas motor and sensory levels ranged in the control group from C5 to L3. Treadmill velocity during the tests did not differ significantly between the groups (median [interquartile range] in exercise vs. control, 1.0 m I sj1 [0.8Y1.1] vs. 0.9 m I sj1 [0.8Y0.9]; P = 0.33). Power output during the submaximal blocks was also not significantly different between the groups (exercise vs. control in block 1: 14.0 W [8.5Y17.2] vs. 12.9 W [10.7Y15.6], P = 0.88; block 2: 16.3 W [13.5Y20.1] vs. 14.2 W [12.1Y17.5], P = 0.57). These power outputs did not significantly differ between the groups over time (block 1: P = 0.54Y1.00; block 2: P = 0.13Y0.65). Steady state during the submaximal blocks was indicated by an RER of 1.00 or less in both groups at each measurement occasion (block 1: 75th percentiles ranging in T1YT4 from 0.89 to 0.92; block 2: 0.88Y0.91). These RER values did not significantly differ between the groups over time (block 1: P = 0.242Y0.323; block 2: P = 0.078Y0.589).

Participants

Analyses of Training Effects

Approximately 200 potentially eligible people were found in the patient databases (Fig. 1). Twentynine could be included (Table 1). All participants had individually fitted wheelchairs without mechanical problems weighing approximately 15Y20 kg (Table 1). The participants were randomly allocated to the exercise group (n = 14) or the control group (n = 15).

No significant training effects were found on the propulsion technique (Tables 2 and 3). The propulsion technique in the exercise group did not significantly change during the 16-wk period, except peak force of block 2 (T1 G T2 at P = 0.008; T2 9 T3 at P = 0.008). In addition, no significant changes were found between the groups over time, except in peak force of block 2 in $T2 j T1 (medians of exercise vs. control: j20 N vs. 1 N; P = 0.01; ru = 0.78). The exercise group’s variance in changes over time was relatively large (Tables 2 and 3), as illustrated by the large interindividual differences in push frequency (Fig. 2). Significant training effects were also not found in mechanical efficiency and POpeak. Mechanical efficiency in submaximal block 1 did not differ significantly during the 16-wk period in the exercise

Training Eleven participants adhered to the training protocol. They performed a median of 894 mins of exercise for 32 sessions in 17 wks while maintaining a median intensity of 38% HRR and an RPE of 2 (Fig. 1). Three participants stopped training after 7Y9 sessions because of lack of motivation, lack of time, or kidney stones (Fig. 1). Adverse events due to the training did not appear. www.ajpmr.com

Low-Intensity Wheelchair Training in Inactive SCI Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

979

TABLE 1 Baseline characteristics in the included group (N = 29) and the group available for analyses during T1, T2, and T3 (n = 16) Included Group

Group size Men/women Paraplegia/tetraplegiab Complete/incompleteb AIS A/B/C/D C4Y6/C7Y8/ Th1Y9/Th10YL5

Analyzed vs. Not Analyzeda

Analyzed Group

Ex vs. Con Exercise Control

Total

Exercise

Control

N

n

n

P

Pa

Pa

29 22/7 20/9 20/9 17/3/7/2 5/4/13/7

8 7/1 6/2 6/2 6/0/0/2 2/0/3/3

8 4/4 7/1 5/3 4/1/1/2 1/0/5/2

0.282 1.000 1.000 NA NA

1.000 0.580 1.000 NA NA

0.282 0.282 1.000 NA NA

1.000 0.382 0.721 0.645 0.878 0.959 0.142 0.014d 0.645

0.852 0.345 0.852 1.000 0.108 0.662 0.394 0.138 0.142

0.867 0.694 0.094 0.121 0.536 0.613 0.955 0.271 1.000

Mdn (25thY75th) Mdn (25thY75th) Mdn (25thY75th) Age, yrs 57 (45Y63) 54 (44Y65) Height, m 1.80 (1.69Y1.86) 182 (171Y187) Body mass, kg 88 (78Y100) 88 (81Y115) BMI, kg/m2 28 (25Y32) 29 (24Y34) TSI, yrs 17 (14Y29) 17 (14Y30) Age at onset SCI, yrs 30 (23Y44) 30 (22Y47) PASIPD (MET hr I wkj1) 8.0 (4.2Y14.6)c 7.7 (4.3Y9.9) Wheelchair weight, kg 17.9 (14.7Y20.8)c 13.7 (12.4Y15.5) Rear wheel diameter, m 0.61 (0.61Y0.64) 0.62 (0.61Y0.64)

54 (49Y62) 176 (166Y185) 95 (84Y101) 30 (25Y34) 19 (13Y29) 35 (17Y48) 11.0 (7.9Y15.1) 19.8 (18.0Y22.3) 0.61 (0.61Y0.65)

Group averages reported as medians (Mdn) and interquartile ranges (25thY75th). Statistical comparisons based on Fisher’s and Mann-Whitney U tests. No differences found between the groups recruited in the two centers (P = 0.059Y1.000), except time since injury (P = 0.003). a Group available for analyses during T1, T2, and T3 (n = 16) compared with the group not available for these analyses (n = 9). b Paraplegia: lesion GTh1; motor complete/incomplete lesion. c Missing in n = 2 and n = 1, respectively. d P G 0.05. AIS, ASIA impairment scale; BMI, body mass index; MET, metabolic equivalent; NA, not applicable; PASIPD, Physical Activity Scale for Individuals with Physical Disabilities20; TSI, time since injury.

group (T1: 4.9% [3.7Y5.6]; T2: 5.3% [4.0Y6.5]; T3: 5.1% [3.2Y5.9]; P = 0.358), similar to block 2 (T1: 5.4% [3.5Y6.1]; T2: 5.4% [4.4Y7.5]; T3: 5.4% [3.4Y6.4]; P = 0.502). $T3 j T1 in block 1 was not significantly different between the groups (exercise vs. control: j0.2% [j0.7 to 0.4] vs. 0.8% [j0.4 to 0.8]; P = 0.115; ru = j0.39), similar to block 2 (0.1% [j0.4 to 0.4] vs. 0.2% [j0.7 to 0.8]; P = 0.717; ru = j0.13). Improvements in POpeak were also not found within the exercise group (T1: 52.6 W [47.5Y58.2]; T2: 51.9 W [39.5Y59.8]; T3: 49.9 W [45.4Y55.1]; P = 0.196), whereas $T3 j T1 was not significantly different between the groups (exercise vs. control: 0.1 W [j1.5 to 3.4] vs. 0.5 W [j3.7 to 6.9]; P = 0.620; ru = j0.19).

DISCUSSION Significant training effects on the propulsion technique were not found after 16 wks of lowintensity wheelchair exercise, twice a week, 30 mins per session in this group of inactive manual wheelchair users with long-term SCI. Changes during the

980

van der Scheer et al.

16-wk period seemed similar in the exercise and control groups, as indicated by high P values and small or negative effect sizes. An exception was push frequency: near-significant reductions were found in the exercise group. Variance was large in the exercise group’s changes over time. The relatively small sample size and missing data limited statistical power. Power was set at 0.80 to detect a meaningful difference. Post hoc power was 0.65 when calculating power based on the a priori equation.12 Therefore, a type 2 error cannot be excluded, but it seems unlikely, given the high P values and small or negative effect sizes. The authors speculate that substantial training effects would not appear in a larger sample. One explanation for the absence of substantial training effects, in contrast to the larger effects found in previous training studies,13,15,19 could be lack of physiologic adaptations. Adaptations were expected, such as reduced local muscular fatigue during the submaximal blocks due to improved peripheral vascular functioning, allowing the adoption

Am. J. Phys. Med. Rehabil. & Vol. 94, No. 11, November 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

www.ajpmr.com

Low-Intensity Wheelchair Training in Inactive SCI

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

981

T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4

57 49 50 52 66 57 63 64 61 71 70 66 0.34 0.38 0.38 0.31

46 35 30 29 48 41 62 54 49 59 62 63 0.32 0.35 0.34 0.30

25th 65 54 59 66 104 62 73 85 68 78 78 76 0.42 0.46 0.47 0.40

75th 8 8 8 5 8 8 8 5 8 8 8 5 8 8 8 5

n

0.197 0.484

0.072 0.263

0.034e 0.093

0.030e 0.069

Pb 52 56 54 50 56 57 58 59 62 67 65 65 0.40 0.43 0.40 0.42

Mdn 48 43 42 39 50 48 48 37 59 61 58 63 0.38 0.37 0.36 0.39

25th 64 61 59 60 63 71 75 77 64 73 68 67 0.47 0.49 0.46 0.43

75th

Control

8 8 8 6 8 8 8 6 8 8 8 6 8 8 8 6

n

0.687 0.327

0.417 0.263

1.000 0.779

0.687 0.208

Pb

1 2 j9

j27 j9 j6

j16 j16 j2

25th

0.04 0.01 0.06 0.01 j0.04 j0.06

11 15 j2

j9 0 3

j11 j10 0

Mdn

$a Exercise

0.06 0.13 0.00

12 20 3

j5 8 7

j4 j5 6

75th

j3 j4 j1

j3 j3 j14

j7 j10 j2

25th

0.03 j0.01 0.01 j0.02 0.00 j0.01

4 1 j 1

2 2 j13

j2 j3 1

Mdn

11 6 9

6 10 j2

3 2 2

75th

0.07 0.04 0.06

$a Control

Between-Group Analyses

0.798 0.065 0.161 0.792 0.505 0.050 0.721 0.429 0.798 0.505 0.130 0.329 0.234 0.959 0.234 0.177

Pc

0.03 0.38 j0.53

0.22 0.47 j0.40

0.59 0.13 j0.33

0.56 0.44 j0.13

rUd

Ex vs. Con

Group averages reported as medians (Mdn) and interquartile ranges (25thY75th). Definitions parameters (23Y25): Push frequency (pushIminj1) determined as the number of completed pushes per minute. Peak force (newtons) as three-dimensional forces within a push phase, averaged over all completed pushes. Contact angle (degrees) as the angle at the end of a push minus the angle at the start, averaged over all completed pushes. Push time (seconds) as the time of start to stop of positive torque over a push, averaged over all completed pushes. a $ is the difference scores within the groups ($T2 j T1, $T3 j T1, and $T4 j T3). b P at T3: Friedman analysis of variance over T1, T2, and T3; p at T4: Wilcoxon test over T3 and T4. c P at T1: Mann-Whitney U test to compare the groups at baseline; P at T2, T3, and T4: Mann-Whitney U test to compare the groups on difference scores. d Effect size for the comparison between the groups on difference scores (rU = 1 Y 2U / [nexercise  ncontrol]).26 e No significant differences in the post hoc Wilcoxon tests (P G 0.05/4). Push frequency showed T1 vs. T2: P = 0.025; T2 vs. T3: P = 0.889; T1 vs. T3: P = 0.050. Peak force showed T1 vs. T2: P = 0.069; T2 vs. T3: P = 0.025; T1 vs. T3: P = 1.000. Ex, exercise group; Con, control group.

Push time, secs

Contact angle, degrees

Peak force, N

Push frequency, pushIminj1

Mdn

Exercise

Within-Group Analyses

TABLE 2 Submaximal block 1: analyses of training effects on the propulsion technique in the participants without missing data at T1, T2, and T3 (n = 16)

982

van der Scheer et al.

Am. J. Phys. Med. Rehabil. & Vol. 94, No. 11, November 2015

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4

50 50 46 57 76 56 68 66 62 70 71 75 0.39 0.37 0.38 0.36

45 34 33 31 59 40 62 58 56 62 68 68 0.31 0.35 0.34 0.32

25th 72 55 58 58 101 65 75 85 75 76 78 76 0.44 0.47 0.49 0.41

75th 8 8 8 5 8 8 8 5 8 8 8 5 8 8 8 5

n

0.079 0.063

0.030g 0.094

G0.001f 0.094

0.120 0.094

P

b

55 56 56 48 60 59 50 64 66 68 68 67 0.43 0.43 0.41 0.43

Mdn 47 45 42 40 49 49 48 38 64 62 60 65 0.41 0.36 0.35 0.40

25th 59 61 59 57 73 73 82 75 70 72 73 73 0.50 0.54 0.50 0.45

75th

Control

8 8 8 6 8 8 8 6 8 8 8 6 8 8 8 6

n

0.967 0.844

0.967 0.563

0.355 0.563

0.794 0.313

P

b

0.03 0.03 0.00

4 5 0

j20 j7 j1

j11 j16 j1

Mdn

j0.01 0.00 j0.03

j1 3 j3

j22 j11 j9

j16 j20 j4

25th

$a Exercise

0.05 0.08 0.01

10 15 3

j15 1 8

j5 j4 6

75th

0.02 j0.03 0.01

3 j6 3

1 1 j10

j1 1 j2

Mdn

j0.04 j0.04 j0.05

j6 j7 j2

0 j3 j21

j3 j7 j5

25th

$a Control

0.06 0.05 0.04

10 8 7

7 10 19

4 4 1

75th

Between-Group Analyses

1.000 0.050 0.083 0.662 0.161 0.007e 0.195 0.662 0.442 0.721 0.234 0.527 0.234 1.000 0.234 0.537

Pc

0.00 0.38 j0.27

0.13 0.38 j0.27

0.78 0.41 j0.20

0.59 0.53 j0.20

rUd

Ex vs. Con

Group averages reported as medians (Mdn) and interquartile ranges (25thY75th). Definitions parameters (23Y25): Push frequency (pushIminj1) determined as the number of completed pushes per minute. Peak force (newtons) as three-dimensional forces within a push phase, averaged over all completed pushes. Contact angle (degrees) as the angle at the end of a push minus the angle at the start, averaged over all completed pushes. Push time (seconds) as the time of start to stop of positive torque over a push, averaged over all completed pushes. a $ is the difference scores within the groups ($T2 j T1, $T3 j T1, and $T4 j T3). b P at T3: Friedman analysis of variance over T1, T2, and T3; p at T4: Wilcoxon test over T3 and T4. c P at T1: Mann-Whitney U test to compare the groups at baseline; P at T2, T3, and T4: Mann-Whitney U test to compare the groups on difference scores. d Effect size for the comparison between the groups on difference scores (rU = 1 Y 2U / [nexercise  ncontrol]).26 e Significantly different at P G 0.05/3. f Post hoc Wilcoxon tests showed T1 vs. T2: P = 0.008; T2 vs. T3: P = 0.008; T1 vs. T3: P = 0.250. g No significant differences in the post hoc Wilcoxon tests (P G 0.05/4). T1 vs. T2: P = 0.191; T2 vs. T3: P = 0.074; T1 vs. T3: P = 0.039. Ex, exercise group; Con, control group.

Push time, secs

Contact angle, degrees

Peak force, N

Push frequency, pushIminj1

Mdn

Exercise

Within-Group Analyses

TABLE 3 Submaximal block 2: analyses of training effects on the propulsion technique in the participants without missing data at T1, T2, and T3 (n = 16)

FIGURE 2 Push frequency in the second block: T1 plotted against T2 and T3. A dot represents a participant’s push frequency. Diagonal line serves as a reference. For example, when a participant’s frequency reduced from T1 to T2, the data point will fall below the diagonal line.

of a more optimal propulsion technique.9,11Y13 Perhaps, such adaptations were lacking owing to the low intensity and twice-weekly frequency of the training. Exercise intensities and frequencies were higher in the previously studied, more effective training programs.13,15,19 However, these did not seem feasible for the inactive wheelchair users in this study.16,17 Alternatively, it could be that the steady-state submaximal exercise blocks were below a critical threshold for shifting to a more optimal propulsion technique.9 Perhaps, such a shift can only be observed in long-term wheelchair users performing fatiguing propulsion,9,11 in contrast to able-bodied novices.15 The absence of substantial training effects could also have been caused by a lack of motor adaptations. The authors speculated that such adaptations could result from the training’s long consecutive propulsion periods.12 These periods seem to be shorter than 3 mins in the daily life of most wheelchair users.18 Assuming that motor adaptations that benefit the propulsion technique are possible in long-term wheelchair users, they may require different or additional stimuli besides long propulsion periods. For example, biofeedback may need to be presented during treadmill propulsion.10 Alternatively, it could be that motor adaptations are limited in long-term wheelchair users. Perhaps, their propulsion technique is already optimal under their current capacity and wheelchair configuration. Although a significant training effect was found when comparing the exercise group with the control group in peak force at T2, such an effect did not appear at T3. The exercise group presumably finished familiarization to treadmill propulsion owing to the training at T2, in contrast to the control group. A difference in familiarization might have disappeared at T3. www.ajpmr.com

The exercise group’s variance in the propulsion technique over time suggests the existence of responders and nonresponders to the training. Potential responders and nonresponders could be expected to differ in characteristics such as lesion level and completeness, body weight, as well as initial propulsion technique. 27,28 The nearsignificant reductions in push frequency suggest that this parameter is relatively sensitive to distinguish responders and nonresponders. A 10% reduction in push frequency was used as a criterion to explore characteristics of potential responders and nonresponders to the training, but systematic differences cannot be derived from the small sample in the present study (Table 4).

Limitations The relatively small sample size and missing data limited statistical power. Although the existence of a type 2 error cannot be excluded, the generally high P values and small effect sizes suggest that substantial training effects would not appear in a larger sample. Possible bias might have occurred in the analyses owing to missing data and the three participants who did not complete the training. Systematic bias due to missing data is not suggested by the similarity between the participants included in and excluded from the analyses (Table 1). Bias due to the participants not completing the training also seems unlikely: results in auxiliary analyses excluding these participants were similar to those shown in Tables 2 and 3. Extra weight of the instrumented and dummy wheels (11.4 kg) may have influenced the propulsion technique.29 An influence on changes over Low-Intensity Wheelchair Training in Inactive SCI

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

983

TABLE 4 Potential responders and nonresponders based on the criterion of a 10% reduction in reduced push frequency after 16 wks of training

Group size Men/women Paraplegia/tetraplegia Complete/incomplete AIS A/B/C/D C4Y6/C7Y8/Th1Y9/Th10YL5

Age, yrs Height, m Body mass, kg BMI, kg/m2 TSI, yrs Age at onset SCI, yrs PASIPD (MET hr I wkj1) POpeak, W Push frequency at T1 (push per minute)

Responders

Nonresponders

N

n

5 4/1 3/2 3/2 3/0/2/0 2/0/2/1

2 2/0 2/0 2/0 2/0/0/0 0/0/1/1

Median (Range)

Median (Range)

47 (39Y65) 180 (165Y187) 86 (66Y134) 28 (23Y38) 17 (14Y34) 25 (16Y50) 8.6 (5.2Y12.0) 53.6 (10.3Y72.9) 46 (44Y79)

62 (57Y66) 188 (183Y193) 102 (83Y121) 29 (22Y36) 21 (14Y28) 40 (29Y51) 7.5 (7.5-NA) 51.3 (43.6Y59.0) 48 (43Y53)

The sample (n = 7) consists of the examined exercise group (Tables 1Y3), excluding a participant who did not complete the training. AIS, ASIA impairment scale; BMI, body mass index; MET, metabolic equivalent; NA, not applicable; PASIPD, Physical Activity Scale for Individuals with Physical Disabilities20; TSI, time since injury.

time is not expected because the extra weight did not differ over time. Similarly, changes over time are not expected to differ owing to the differences in wheelchair weight between the exercise and control groups.

Clinical Implications and Future Research Higher exercise intensities or frequencies may be required for physiologic adaptations that can allow a more optimal propulsion technique. However, higher exercise intensities are not considered safe for inactive or deconditioned populations,14 whereas more than twice a week center-based exercise might not be feasible for inactive, community-dwelling wheelchair users with SCI.16,17 Perhaps, more frequent exercise is feasible in supervised exercise at home or at a local fitness center.30 Alternatively, it could be that low-intensity wheelchair training without additional interventions is an insufficient stimulus for long-term wheelchair users to adopt a more optimal propulsion technique. Perhaps, some participants responded to the training, but the sample was too small to derive conclusions about potential responders and nonresponders. This might become clearer when using single-case designs and multiple measurements occasions before, during, and after an intervention. Analyses on interindividual differences are recommended when a sufficiently sized sample can be recruited.28

984

van der Scheer et al.

Even small improvements in the propulsion technique are assumed to be clinically relevant for reducing the risk for upper-joint damage during daily wheelchair propulsion.1,27 Any intervention effective for this purpose should therefore be explored. Perhaps, this requires an individually tailored approach using a combination of training, biofeedback, and optimizing wheelchair configuration.1,10,13,27,28

CONCLUSIONS It seems that substantial training effects on the propulsion technique cannot be expected in inactive people with long-term SCI performing 16 wks of low-intensity wheelchair exercise, twice a week, 30 mins per session. Perhaps, inducing a more optimal propulsion technique in long-term wheelchair users requires higher exercise intensities or frequencies, but these are not considered safe or feasible for inactive or deconditioned populations.1,16,17 A limitation in these conclusions is that it cannot be excluded that the nonsignificant findings were caused by a lack of statistical power. The possibility of responders and nonresponders to the training is suggested by the large variance in changes in the propulsion technique over time. It remains to be investigated what type of interventions could help different people with SCI to adopt a propulsion technique that can prevent or reduce the risk for upper-body joint damage, pain, and pathology.

Am. J. Phys. Med. Rehabil. & Vol. 94, No. 11, November 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

ACKNOWLEDGMENTS

For their assistance in measurements, the authors thank Anita Fijen and Joke Sprik of rehabilitation centers Heliomare and University Medical Center Groningen Beatrixoord as well as a large group of (under)graduate students including Thijs Hoogantink, Karin Lubberding, Jacqueline Pot, and Marco Soesman. The authors also thank the technical personnel of the Center of Human Movement Sciences (University of Groningen, University Medical Center Groningen) for their technical assistance. List of the names of ALLRISC group members: Hans Bussmann, PhD, Department of Rehabilitation Medicine and Physical Therapy, Erasmus MC University Medical Centre, Rotterdam, the Netherlands; Willemijn Faber, MD, Heliomare, Rehabilitation Center, Wijk aan Zee, the Netherlands; David Gobets, MD, Heliomare, Rehabilitation Center, Wijk aan Zee, the Netherlands; Thomas Janssen, PhD, MOVE Research Institute Amsterdam, Faculty of Human Movement Sciences, VU University Amsterdam, the Netherlands and Amsterdam Rehabilitation Research Centre|Reade, Amsterdam, the Netherlands; Marcel Post, PhD, Rehabilitation Center De Hoogstraat, Utrecht, the Netherlands; Linda Valent, PhD, PT, Heliomare, Rehabilitation Center, Wijk aan Zee, the Netherlands; and Ferry Woldring, PT, University Medical Center Groningen, Center for Rehabilitation, Department of Rehabilitation Medicine, the Netherlands.

REFERENCES 1. Boninger ML, Koontz AM, Sisto SA, et al: Pushrim biomechanics and injury prevention in spinal cord injury: recommendations based on CULP-SCI investigations. J Rehabil Res Dev 2005;42(suppl):9Y19 2. Jensen M, Hoffman A, Cardenas D: Chronic pain in individuals with spinal cord injury: a survey and longitudinal study. Spinal Cord 2005;43:704Y12 3. Eriks-Hoogland I: Shoulder Impairment in Persons with a Spinal Cord Injury (thesis): appendices 1 and 2. Available at: http://www.rug.nl/staff/l.h.v.van.der.woude/ thesisinge.pdf 2014. Accessed December 28, 2014 4. Pentland WE, Twomey LT: Upper limb function in persons with long term paraplegia and implications for independence: Part II. Paraplegia 1994;32:219Y24 5. Requejo PS, Mulroy SJ, Haubert LL, et al: Evidencebased strategies to preserve shoulder function in manual wheelchair users with spinal cord injury. Top Spinal Cord Inj Rehabil 2008;13:86Y119 6. Post MW, van Asbeck FW, van Dijk AJ, et al: Services for spinal cord injured: availability and satisfaction. Spinal Cord 1997;35:109Y15 7. van der Woude LH, de Groot S, Postema K, et al: Active LifestyLe Rehabilitation interventions in aging spinal cord injury (ALLRISC): a multicentre research program. Disabil Rehabil 2013;35:1097Y103 www.ajpmr.com

8. van der Woude LH, de Groot S, van Drongelen S: Evaluation of manual Wheelchair performance in everyday life. Top Spinal Cord Inj Rehabil 2009;15:1Y15 9. Rice I, Impink B, Niyonkuru C, et al: Manual wheelchair stroke characteristics during an extended period of propulsion. Spinal Cord 2009;47:413Y7 10. Rice IM, Pohlig RT, Gallagher JD, et al: Handrim wheelchair propulsion training effect on overground propulsion using biomechanical real-time visual feedback. Arch Phys Med Rehabil 2013;94:256Y63 11. Rodgers MM, McQuade KJ, Rasch EK, et al: Upperlimb fatigue-related joint power shifts in experienced wheelchair users and nonwheelchair users. J Rehabil Res Dev 2003;40:27Y37 12. van der Scheer JW, de Groot S, Postema K, et al: Design of a randomized-controlled trial on low-intensity aerobic wheelchair exercise for inactive persons with chronic spinal cord injury. Disabil Rehabil 2013;35:1119Y26 13. Rodgers MM, Keyser RE, Rasch EK, et al: Influence of training on biomechanics of wheelchair propulsion. J Rehabil Res Dev 2001;38:505Y11 14. Garber CE, Blissmer B, Deschenes MR, et al: American College of Sports Medicine position stand. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults: guidance for prescribing exercise. Med Sci Sports Exerc 2011;43: 1334Y59 15. de Groot S, de Bruin M, Noomen SP, et al: Mechanical efficiency and propulsion technique after 7 weeks of low-intensity wheelchair training. Clin Biomech (Bristol, Avon) 2008;23:434Y41 16. Hicks AL, Martin KS, Ditor DS, et al: Long-term exercise training in persons with spinal cord injury: effects on strength, arm ergometry performance and psychological well-being. Spinal Cord 2003; 41:34Y43 17. Valent LJ, Dallmeijer AJ, Houdijk H, et al: Effects of hand cycle training on physical capacity in individuals with tetraplegia: a clinical trial. Phys Ther 2009;89:1051Y60 18. Tolerico ML, Ding D, Cooper RA, et al: Assessing mobility characteristics and activity levels of manual wheelchair users. J Rehabil Res Dev 2007;44:561Y71 19. Dallmeijer AJ, van der Woude LHV, Pathuis CS: Adaptations in wheelchair propulsion technique after training in able-bodied subjects. In: van der Woude LHV HM, van Kemenade CH, (eds): Biomedical Aspects of Manual Wheelchair Propulsion. Amsterdam, the Netherlands, IOS Press, 1999, pp. 224Y6 20. de Groot S, van der Woude LH, Niezen A, et al: Evaluation of the physical activity scale for individuals with physical disabilities in people with spinal cord injury. Spinal Cord 2010;48:542Y7 21. Borg GA: Psychophysical bases of perceived exertion. Med Sci Sports Exerc 1982;14:377Y81 22. van der Grinten MP, Smitt P: Development of a practical method for measuring body part discomfort. In:

Low-Intensity Wheelchair Training in Inactive SCI Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

985

Kumar A (ed): Advances in Industrial Ergonomics and Safety. London, United Kingdom, Taylor & Francis, 1992, pp 101Y14 23. de Groot S, Vegter R, Vuijk C, et al: WHEEL-I: development of a wheelchair propulsion laboratory for rehabilitation. J Rehabil Med 2014;46:493Y503 24. de Groot S, Dallmeijer AJ, van Asbeck FW, et al: Mechanical efficiency and wheelchair performance during and after spinal cord injury rehabilitation. Int J Sports Med 2007;28:880Y6 25. Vegter R, de Groot S, Lamoth C, et al: Initial skill acquisition of handrim wheelchair propulsion: A new perspective. IEEE Trans Neural Syst Rehabil Eng 2014; 22:104Y13 26. Kerby DS: The simple difference formula: an approach to teaching nonparametric correlation. Innovative Teaching 2014;3:1Y9

986

van der Scheer et al.

27. Cowan RE, Boninger ML, Sawatzky BJ, et al: Preliminary outcomes of the SmartWheel Users’ Group database: a proposed framework for clinicians to objectively evaluate manual wheelchair propulsion. Arch Phys Med Rehabil 2008;89:260Y8 28. Vegter RJ, Lamoth CJ, de Groot S, et al: Inter-individual differences in the initial 80 minutes of motor learning of handrim wheelchair propulsion. PLoS One 2014;9:e89729 29. de Groot S, Vegter RJ, van der Woude LH: Effect of wheelchair mass, tire type and tire pressure on physical strain and wheelchair propulsion technique. Med Eng Phys 2013;35:1476Y82 30. Rimmer JH, Henley KY: Building the crossroad between inpatient/outpatient rehabilitation and lifelong community-based fitness for people with neurologic disability. J Neurol Phys Ther 2013;37:72Y7

Am. J. Phys. Med. Rehabil. & Vol. 94, No. 11, November 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.