Spinal Cord (2009) 47, 62–66

& 2009 International Spinal Cord Society All rights reserved 1362-4393/09 $32.00
www.nature.com/sc

ORIGINAL ARTICLE

Effects of 6 months of regular passive movements on ankle joint

mobility in people with spinal cord injury: a randomized controlled
trial
LA Harvey1, RD Herbert2, J Glinsky1, AM Moseley2 and J Bowden1
1
Rehabilitation Studies Unit, Northern Clinical School, Faculty of Medicine, The University of Sydney, Sydney, New South Wales,
Australia and 2Musculoskeletal Division, The George Institute for International Health, Sydney, New South Wales, Australia

Study design: Assessor-blinded within-subject randomized controlled trial.

Objective: To determine the effects of 6 months of regular passive movements on ankle joint mobility
in people with spinal cord injury.
Setting: Community, Australia.
Methods: A total of 20 people with tetraplegia living in the community had one ankle randomized to
a control group and the other to an experimental group. Carers administered passive movements to
participants’ experimental ankles for 10 min, 10 times a week for 6 months. The control ankles were left
untreated. The primary outcome was passive ankle dorsiflexion range of motion.
Results: Adherence was high (mean adherence rate of 96%). Ankle dorsiflexion range of motion
decreased by a mean (s.d.) of 21 (4) in control ankles and increased by 21 (4) in experimental ankles.
The mean (95% confidence interval, CI) effect on ankle dorsiflexion range of motion was 41 (95% CI,
2–61).
Conclusion: Regular passive movements have small effects on ankle joint mobility. It is unclear if these
effects are clinically worthwhile.
Spinal Cord (2009) 47, 62–66; doi:10.1038/sc.2008.71; published online 24 June 2008

Keywords: ankle; stiffness; spinal cord injury; rehabilitation

Introduction
Contractures are a disfiguring and disabling complication of This places a significant burden on people with SCI as well as
spinal cord injury (SCI).1 They prevent the performance of their carers.
motor tasks and result in unsightly deformities. They also It is widely believed that passive movements are an
predispose individuals to pain and sleep disturbances.1 essential part of the ongoing physical care of people with
Contractures are caused by structural adaptation of muscles SCI. They are primarily justified by the deleterious effects of
and soft tissues spanning joints, and are frequently accom- paralysis on joint mobility and by anecdotal observations
panied by spasticity.2,3 regarding the benefits of regular passive movements. How-
A range of interventions are used to treat and prevent ever, the effectiveness of passive movements has never been
contractures.4 Perhaps the most widely used intervention is confirmed with a clinical trial and the results of the few high
the application of passive movements. Passive movements quality studies in non-disabled populations have not
involve manually and repeatedly moving joints through demonstrated improvements in passive range of motion
range. Typically a carer or physiotherapist applies a total of from regular passive movements.5–8 Therefore, the purpose
20–30 min of regular passive movements to all joints affected of this trial was to determine the effectiveness of passive
by paralysis. This often equates to a couple of minutes per movements administered over a 6-month period to the
joint for a person with tetraplegia. These interventions are ankles of people with SCI. An intensive treatment was
typically administered every day throughout a person’s life. selected (that is, 20 min of passive movements five times a
week) because we wanted to maximize the likelihood of
finding a treatment effect. Subsequent trials can address
Correspondence: Dr LA Harvey, Rehabilitation Studies Unit, Royal Rehabilita- questions about the relative effectiveness of different
tion Centre Sydney, PO Box 6, Ryde, New South Wales 1680, Australia.
dosages. This is a more efficient way of answering questions
E-mail: l.harvey@usyd.edu.au
Received 9 April 2008; revised 5 May 2008; accepted 11 May 2008; published about treatment effectiveness than starting with smaller
online 24 June 2008 dosages, running the risk of finding no treatment effect, and
 Passive movements for people with SCI
LA Harvey et al
63

therefore needing to repeat the same trial with larger

dosages.

Materials and methods

Participants
A total of 20 people with tetraplegia living in the community
were invited to participate in the trial. Participants were
included if they were wheelchair dependent, had mild to
moderate ankle stiffness (less than 1011 dorsiflexion with a
12 Nm torque applied to the ankle9 but an arc of at least 151
motion), had paralysis around both knees and ankles and
had carers able to provide the intervention.
A power analysis indicated that 20 participants (that is, 20
pairs of ankles) would provide a 95% probability of detecting a
treatment effect of 51, assuming an a of 0.05, a standard
deviation (s.d.) of the treatment effect of 51 and loss to follow-
up of 15%. It was decided prior to the conduct of the trial that
the smallest clinically worthwhile effect would be 51. This
figure was nominated after considering the time, cost and
effort associated with the intervention and the implications of
loss of ankle mobility on the lives of people with SCI.
All applicable institutional and governmental regulations
concerning the ethical use of human volunteers were
followed during the course of this research. The trial was
registered prior to commencement with the Australian New
Figure 1 The testing device. Image used with permission from
Zealand Clinical Trial Register (ACTRN12607000220460). www.physiotherapyexercises.com.

Intervention
Prior to commencement of the trial, a computer-generated The primary outcome was passive ankle dorsiflexion range
random number sequence was created by a person not of motion with the application of a 12 Nm torque. This
involved in recruitment to determine the allocation sche- torque was less than that specified in the original protocol
dule. Each participant’s allocation was placed in a sealed, (i.e., 17 Nm). It was changed soon after commencing the trial
opaque, sequentially numbered envelope to ensure conceal- because of concerns about the integrity of participants’
ment. The envelopes were not opened until after the plantarflexor muscles. Passive ankle dorsiflexion was also
participant had completed all pre-trial assessments. A measured with the application of six smaller torques (2, 3, 5,
participant was considered to have entered the trial once 7, 8 and 10 Nm). Measurements were made with a custom-
his/her envelope was opened. built device.10 The device consisted of a footplate attached to
The experimental ankles of participants were passively a wheel (Figure 1). The foot was firmly secured to the
moved by participants’ carers for 10 min in the morning and footplate. Both the foot and footplate rotated in response to
10 min in the evening, 5 days a week for 6 months. Carers a weight hung from a rope looped around the rim of the
were given written instructions and training on how to wheel. Dorsiflexion angle was derived from the difference in
administer the passive movements. Participants and carers inclination of the footplate and tibia measured with a digital
were also visited regularly to ensure the passive movements inclinometer.
were performed correctly. Participants or carers were required Testing always followed the same format. Participants lay
to record when and for how long the passive movements in a supine position. The right ankle was tested before the
were administered in a diary. In addition, participants were left ankle. Initially a 12 Nm torque was applied for 3 min to
contacted at least every second week and often every week. precondition the ankle. Dorsiflexion was then measured
At this time, recordings from the diaries were noted. The with the application of six progressively larger torques,
control ankles did not receive passive movements or starting with 2 Nm and finishing with 12 Nm torques. Each
stretches for the duration of the trial. torque was applied for approximately 20 s before measuring
ankle angle. The reliability of this procedure is high
Measurement (intraclass correlation coefficient of 0.95; 95% confidence
Participants were assessed at baseline and then 1 day after interval (CI), 0.92–0.98).10
the 6-month intervention period. No passive movements or The secondary outcomes were spasticity in ankle plantar-
stretches were administered in the 24 h prior to either flexor and knee hamstring muscles and a measure of
assessment. One assessor performed all but two assessments. participants’ perception of change. Spasticity in ankle
All assessors were blinded to group allocation. plantarflexor and knee hamstring muscles was measured

Spinal Cord
 Passive movements for people with SCI
LA Harvey et al
64

with the six-point modified Ashworth scale.11,12 This test of 96%). The reasons given for nonadherence included
has limited reliability in people with SCI (mean k ¼ 0.37)12 hospital visits, pressure areas and lack of time. One
but is widely used as a clinical test of spasticity in the participant sustained an ankle fracture within a week of
community setting.13 At the completion of the trial, commencing the trial. This was unrelated to participation in
participants were asked to rate perceived Global Impression the trial but required cessation of passive movements for 14
of Change in both the treated and untreated ankle using a weeks.
15-point Likert scale where 7 indicates ‘very great deal Three participants spent the majority of the 6-month
worse’, 0 indicates ‘no change’ and þ 7 indicates ‘very great period in bed. The other participants mobilized in a wheel-
deal better’.14 In addition, participants were asked to rate the chair with their feet supported on footplates. Twelve
convenience or inconvenience of the intervention on a participants slept at night on their sides with their feet
10 cm visual analogue scale. At one end of the 10 cm scale unsupported and eight slept supine with both feet supported
were the words ‘very inconvenient’ and at the other end at 901. No participant used stationary bicycles or electrical
were the word ‘not at all inconvenient’. A higher score stimulation but one participant regularly stood with stand-
indicated less inconvenience. ing equipment. None of these co-interventions or any other
unidentified co-interventions were likely to affect one ankle
Data analysis more than the other.
The t-distribution was used to estimate 95% CI for between- Passive dorsiflexion of the experimental ankles increased
group (that is, between-leg) differences in change scores for slightly over the 6-month period from a mean (s.d.) of 88 (9)
ankle angle (posttest score minus pretest score). Paired t-tests to 911 (10) when measured with the application of 12 Nm
were used to test for significant differences. Probabilities of torque. The corresponding angle in control ankles decreased
less than 0.05 were considered significant. The ‘centile’ slightly over the same period from a mean (s.d.) of 89 (8) to
routine in Stata (v9.2; Statacorp, TX, USA) was used to derive 871 (9). Thus, the overall between-group mean difference was
the 95% CIs for median between-group differences for the 41 (95% CI, 2–6, P ¼ 0.002). Between-group differences for
modified Ashworth and Global Impression of Change data. passive dorsiflexion with the other five torques were similar
This method does not make assumptions about the distribu- (Table 1). Removal of the three least compliant participants’
tion of the data. All data were analyzed according to the data from the analyses had little effect on the results (mean
intention-to-treat principle. treatment effect with 12 Nm was 51, 95% CI, 3–7).
The overall between-group difference in median scores for
the modified Ashworth score of the hamstring and plantar-
Results flexor muscles were 0 (95% CI, 0–0) and 0 (95% CI, 0–1),
respectively (Table 2). These results were not statistically
Participants had American Spinal Injury Association (ASIA) significant.
neurological levels ranging from C2 to C7. Eleven partici- Participants reported a median (interquartile) Global
pants had ASIA A lesions and nine participants had ASIA B Impression of Change of three points (2–4) on the 15-point
lesions. The median (interquartile) age and time since injury scale in the treated ankle and 0 points (0–0) in the untreated
were 39 (34–44) and 8 years (4–14), respectively. Three ankle. The overall between-group difference in median
participants were women and seventeen were men. scores was two points (95% CI, 2–3). Some participants
The flow of participants through the trial is given in reported that the passive movements reduced ankle stiffness,
Figure 2. There was no loss to follow-up and complete data spasticity and edema.
sets were obtained from all participants. Participants did not find the passive movements incon-
The protocol dictated that each participant receive 260 10- venient. The median (interquartile) rating of inconvenience
min treatments over a 6-month period. In practice associated with the provision of passive movements was 8
participants received on average (s.d.) 250 treatments (70) points (6–8) where 10 is ‘not at all inconvenient’.
over a 6-month period (equivalent to an average adherence

20 participants (40 ankles)

Discussion
randomized
This is the first randomized controlled trial to examine the
effects of passive movements on joint mobility in people
with SCI. It is important because passive movements are
commonly administered to people with SCI on a regular
20 ankles in experimental 20 ankles in control
group group basis for the treatment and prevention of contractures.
There was a small but statistically significant effect of
passive movements on ankle mobility: on average the effect
of 260 10-min treatments of passive movements adminis-
20 ankles re-tested 20 ankles re-tested
tered over a 6-month period was 41 (95% CI, 2–6). At issue is
at six months at six months whether this effect is large enough to justify the widespread
and ongoing administration of passive movements for the
Figure 2 Flow of participants through the trial. treatment and prevention of contractures. Prior to the start

Spinal Cord
 Passive movements for people with SCI
LA Harvey et al
65

Table 1 Mean (s.d.) passive ankle dorsiflexion (degrees) of experimental and control legs at the beginning and end of the 6-month period

Control Experimental
Between-group difference
Stretch torque (Nm) Pre Post Pre Post

2 81 (7) 78 (9) 81 (9) 81 (10) 3 (95% CI, 1–6)

3 81 (7) 80 (9) 81 (9) 83 (9) 3 (95% CI, 1–5)
5 82 (10) 81 (9) 83 (9) 84 (9) 2 (95% CI, -1–4)
7 85 (7) 83 (9) 85 (9) 86 (10) 3 (95% CI, 1–5)
8 86 (7) 84 (9) 86 (9) 88 (10) 4 (95% CI, 1–6)
10 87 (7) 85 (9) 87 (9) 89 (10) 4 (95% CI, 2–6)
12 89 (8) 87 (9) 88 (9) 91 (10) 4 (95% CI, 2–6)

Passive ankle dorsiflexion was measured with the application of 2, 3, 5, 7, 8, 10 and 12 Nm. The overall mean (95% CI) between-group differences are also shown.
Data reflect both feet of all participants.

Table 2 Median (interquartile) modified Ashworth scores for the With any intervention there is always the possibility that a
hamstring and plantarflexor muscles of experimental and control legs particular subgroup of patients respond better to the
at the beginning and end of the 6-month period
intervention than the majority. If these patients could be
Muscle Control Experimental Between-group identified, they could be specifically targeted for therapeutic
difference attention. For example, it could be that those with recent SCI
Pre Post Pre Post benefit more from regular passive movements than those
with long-standing SCI, or that those with early contracture
Hamstring 1 (0–2) 0 (0–1) 1 (0–2) 1 (0–1) 0 (0–0)
Plantarflexor 1 (0–2) 1 (0–2) 1 (0–2) 1 (0–1) 0 (0–1) benefit more than those with more established contracture.
Although these hypotheses sound plausible they are not
The overall between-group differences (95% CI) in median scores are also supported by the data. There was little person-to-person
shown. Data reflect both feet of all participants.
variation in the size of the treatment effect (the s.d. of the
treatment effect was just 51). That is, there was no evidence
that some individuals responded substantially more to
of the trial we articulated a belief that passive movements passive movements than others.
administered for 6 months needed to have an effect on ankle The results of this study are consistent with three
range of motion of at least 51 to be clinically worthwhile. randomized controlled trials in people with SCI and two in
This is conservative as others set the clinically worthwhile people with other neurological conditions indicating that
criterion at 101.15 Our point estimate of 41 therefore suggests sustained stretch, as typically applied in the clinical setting,
a treatment effect that is too small to be worthwhile. is not effective for the treatment and prevention of
Few would dispute the claim that 41 of ankle range of contractures.16–20 Although passive movements confer a
motion is of little functional importance on its own. slightly different mechanical stimulus to that provided
However, it is possible that the therapeutic effects of passive by sustained stretch, the two interventions are similar.
movements accumulate over time. That is, it is possible that The main difference is that passive movements involve
20 min of passive movements could prevent a loss of 41 every the administration of repetitive short-duration stretches.
6 months, equivalent to a loss of 801 every 10 years. An effect The results of this trial add to the mounting evidence
of this magnitude would justify the routine provision of indicating that stretch is not the potent stimulus it is
passive movements over an extended period of time. assumed to be, regardless of whether it is applied in a cyclic
However, it is not known whether the therapeutic effects repetitive manner or in a sustained way.
of passive movements are cumulative or not. Nor is it feasible Interestingly, although the effects of passive movements
to conduct a randomized trial over a 10-year period to clarify on ankle range of motion were small, nearly all participants
the issue. felt the passive movements were worthwhile. They com-
The passive movements administered in this trial were monly stated that the passive movements decreased spasti-
very intensive. In most contexts it would be difficult to apply city and ankle stiffness even though these claims were not
20 min of passive movements to any one joint on a long- substantiated in the data collected by blinded assessors.
term basis, and it would be nearly impossible to apply 20 min Participants did not find the intervention particularly
of passive movements to all the joints affected by tetraplegia. inconvenient.
Data on the relationship between dose of passive movements In conclusion, this randomized trial indicates that an
and effects are not available but it would seem reasonable to intensive 6-month program of passive movements increases
expect that the size of the treatment effect is positively range of motion by 41. This treatment effect is too small to be
related to dose. If this is the case, 2–3 min of passive intrinsically worthwhile. However, if the same effect can be
movements to a joint each day, as typically provided in the attained with 2–3 min of passive movements and if the
community setting, would be expected to have smaller effects accumulate with time, passive movements may be
treatment effects than demonstrated in this trial. It is not effective for the treatment and prevention of contractures
clear just how small the treatment effects would be. when provided on a routine and long-term basis.

Spinal Cord
 Passive movements for people with SCI
LA Harvey et al
66

Acknowledgements 9 Moseley A, Crosbie J, Adams R. High- and low-ankle flexibility

and motor task performance. Gait Posture 2003; 18: 73–80.
This project was funded by The University of Sydney’s 10 Harvey L, Byak A, Ostrovskaya M, Glinsky J. Reliability of a device
Research and Development Grants Scheme. Rob Herbert is designed to measure ankle mobility. Spinal Cord 2003; 41:
supported by a fellowship from the Australian NHMRC. We 559–562.
11 Gregson JM, Leathely M, Moore AP, Sharma AK, Smith TL,
thank our 20 dedicated participants as well as their families Walkins CL. Reliability of the tone assessment scale
and carers. This study was sponsored by The University of and the modified Ashworth scale as clinical tools for
Sydney, the Royal Rehabilitation Centre Sydney and Southern assessing poststroke spasticity. Arch Phys Med Rehabil 1999; 80:
Cross Community Healthcare. 1013–1016.
12 Haas BM, Bergstrom E, Jamous A, Bennie A. The inter rater
reliability of the original and of the modified Ashworth scale for
the assessment of spasticity in patients with spinal cord injury.
References Spinal Cord 1996; 34: 560–564.
1 Grover J, Gellman H, Waters RL. The effect of a flexion 13 Cardenas DD, Ditunno J, Graziani V, Jackson AB, Lammertse D,
contracture of the elbow on the ability to transfer in patients Potter P et al. Phase 2 trial of sustained-release fampridine in
who have quadriplegia at the sixth cervical level. J Bone Joint Surg chronic spinal cord injury. Spinal Cord 2007; 45: 158–168.
Am 1996; 78: 1397–1400. 14 Barrett B, Brown D, Mundt M, Brown R. Sufficiently important
2 Dalyan M, Sherman A, Cardenas DD. Factors associated with difference: expanding the framework of clinical significance. Med
contractures in acute spinal cord injury. Spinal Cord 1998; 36: Decis Making 2005; 25: 250–261.
405–408. 15 Robinson W, Smith R, Aung O, Ada L. No difference between
3 McDonald MF, Garrison MK, Schmit BD. Length-tension proper- wearing a night splint and standing on a tilt table in preventing
ties of ankle muscles in chronic human spinal cord injury. ankle contracture early after stroke: a randomised trial. Aust
J Biomech 2005; 8: 2344–2353. J Physiother 2008; 54: 33–38.
4 Herbert RD. The prevention and treatment of stiff joints. In: 16 Ben M, Harvey L, Denis S, Glinsky J, Goehl G, Chee S et al. Does
Crosbie J, McConnell J (eds). Key Issues in Musculoskeletal 12 weeks of regular standing prevent loss of ankle mobility and
Physiotherapy. Butterworth Heinemann: Oxford, 1993, bone mineral density in people with recent spinal cord injuries?
pp 114–141. Aust J Physiother 2005; 51: 251–256.
5 Milne S, Brosseau L, Robinson V, Noel M, Davis J, Drouin H et al. 17 Lannin NA, Horsley SA, Herbert R, McCluskey A, Cusick A.
Continuous passive motion following total knee arthroplasty. Splinting the hand in the functional position after brain
Cochrane Database Syst Rev 2003 Iissue 2. Art. No.: CD004260. impairment: a randomized, controlled trial. Arch Phys Med Rehabil
doi:10.1002/14651858.CD004260 (see www.cochrane.org/reviews/ 2003; 84: 297–302.
en/ab004260.html). 18 Turton AJ, Britton E. A pilot randomized controlled trial of a daily
6 Kay S, Haensel N, Stiller K. The effect of passive mobilisation muscle stretch regime to prevent contractures in the arm after
following fractures involving the distal radius: a randomised stroke. Clin Rehabil 2005; 19: 600–612.
study. Aust J Physiother 2000; 46: 93–101. 19 Harvey LA, Batty J, Crosbie J, Poulter S, Herbert RD.
7 Harvey LA, Herbert RD. Muscle stretching for treatment and A randomized trial assessing the effects of 4 weeks of daily
prevention of contracture in people with spinal cord injury. stretching on ankle mobility in patients with spinal cord injuries.
Spinal Cord 2002; 40: 1–9. Arch Phys Med Rehabil 2000; 81: 1340–1347.
8 ter Woerds W, de Groot PC, van Kuppevelt DH, Hopman MT. 20 Harvey LA, Byak AJ, Ostrovskaya M, Glinsky J, Katte L, Herbert RD.
Passive leg movements and passive cycling do not alter arterial leg Randomised trial of the effects of four weeks of daily stretch on
blood flow in subjects with spinal cord injury. Phys Ther 2006; 86: extensibility of hamstring muscles in people with spinal cord
636–645. injuries. Aust J Physiother 2003; 49: 176–181.

Spinal Cord