Clinical Rehabilitation

http://cre.sagepub.com/

Can a six-week exercise intervention improve gross motor function for non-ambulant
children with cerebral palsy? A pilot randomized controlled trial
Elizabeth Bryant, Terry Pountney, Heather Williams and Natalie Edelman
Clin Rehabil published online 30 July 2012
DOI: 10.1177/0269215512453061

The online version of this article can be found at:

http://cre.sagepub.com/content/early/2012/07/30/0269215512453061
A more recent version of this article was published on - Jan 21, 2013

Published by:

http://www.sagepublications.com

Additional services and information for Clinical Rehabilitation can be found at:

Email Alerts: http://cre.sagepub.com/cgi/alerts

Subscriptions: http://cre.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Version of Record - Jan 21, 2013

>> OnlineFirst Version of Record - Jul 30, 2012

What is This?

Downloaded from cre.sagepub.com at THAMMASAT UNIVERSITY on October 9, 2014

453061
2012
CRE0010.1177/0269215512453061Clinical RehabilitationBryant et al.

CLINICAL
REHABILITATION

Clinical Rehabilitation

Can a six-week exercise 0(0) 1­–10

© The Author(s) 2012
Reprints and permissions:
intervention improve gross sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0269215512453061

motor function for non- cre.sagepub.com

ambulant children with cerebral

palsy? A pilot randomized
controlled trial

Elizabeth Bryant1,2, Terry Pountney1,2,

Heather Williams1 and Natalie Edelman3

Abstract
Objective: To determine the effect of a six-week exercise intervention on gross
motor function for non-ambulant children with cerebral palsy.
Design: A parallel arm randomized controlled trial.
Setting: Four special schools.
Participants: Thirty-five children aged 8–17 with bilateral cerebral palsy; Gross Motor Function
Classification System levels IV–V.
Method: Participants were randomly allocated to a static bike group, a treadmill group or control group.
Participants in the bike and treadmill groups received exercise training sessions, three times weekly for
six weeks. The control group received their usual care. Blinded assessments were performed at baseline
and six weeks and followed up at 12 and 18 weeks.
Outcome measures: Gross Motor Function Measures GMFM-66, GMFM-88D and GMFM-88E.
Results: At six weeks significant differences were found in GMFM-88D scores between the bike group
and the control group, and the treadmill group and the control group (P < 0.05). The mean change
(SD) in GMFM-88D score was 5.9 (6.8) for the bike group; 3.7 (4.4) for the treadmill group and 0.5
(1.9) for the control group. No significant differences were found for GMFM-66 or GMFM-88E scores
between the bike group and control group, or the treadmill group and control group, although trends
of improvement were observed for both exercise groups. The improvements observed declined during
the follow-up period.

1Chailey Heritage Clinical Services, Sussex Community NHS Corresponding author:

Trust, East Sussex, UK Terry Pountney, Chailey Heritage Clinical Services, Beggars
2Clinical Research Centre for Health Professions, University of Wood Road, North Chailey, Nr Lewes, East Sussex BN8 4JN,
Brighton, East Sussex, UK UK
3Centre for Health Research, University of Brighton, Falmer, Email: [email protected]
Brighton, East Sussex, UK

Downloaded from cre.sagepub.com at THAMMASAT UNIVERSITY on October 9, 2014

2 Clinical Rehabilitation 0(0)

Conclusion: This study provides preliminary evidence that exercising on a bike or treadmill may provide
short-term improvements in gross motor function for non-ambulant children with cerebral palsy. This
needs to be tested in a large-scale randomized trial.

Keywords
Bicycle training, treadmill training, cerebral palsy, children, non-ambulant
Received: 4 January 2012; accepted: 5 June 2012

Introduction
Children with cerebral palsy who are unable to walk walk on a treadmill with a support harness.15
independently are particularly prone to muscle Therefore, the rationale for this study was to iden-
weakness,1 which contributes to pain, deformity tify effective interventions, where none currently
and functional loss.2 exist, with a view to offering a choice of activity for
A number of previous studies have demonstrated this population.
that strength and exercise interventions for children The aim of this study was to determine the effect
with mild to moderate cerebral palsy can result in an of a six-week exercise intervention, using a static
increase in muscle volume,3 muscle strength4–6 or bicycle or treadmill, on gross motor function ability
can improve gross motor function.6–10 However, the in non-ambulant children with cerebral palsy.
effectiveness of such interventions remains unclear
as, while meta-analyses of randomized controlled
trials have found increases in strength compared to Method
control groups, the differences were not statistically
significant.11,12 The authors of both reviews sug- Participants were recruited from four special
gested the need for interventions of longer duration schools in the south of the UK. Participants aged
and of sufficient intensity to result in significant 8–17 years, with cerebral palsy at GMFCS17 levels
strength gains. IV and V, able to pedal on an adapted static bicycle
Some studies have examined the effect of tread- and walk with partial body weight support on a
mill training, using partial body weight support, for treadmill were invited to participate in the study.
children with mild to severe cerebral palsy,10,13–16 They were excluded if they had undergone ortho-
reporting improvements in gross motor function10,14,15 paedic surgery to the spine or lower limbs within
and increases in walking speeds.13–16 the last year, or if they had cognitive or behavioural
There are very few studies that have investigated impairment preventing understanding or compli-
the effect of exercise interventions in non-ambulant ance with instructions.
children with cerebral palsy as routine strength and Participants were randomly allocated to the bike,
exercise training protocols (such as weight training treadmill or control group at each site. The study
or circuit training) are not feasible for this group. administrator provided nine sealed envelopes (three
Treadmill training and static bikes are not routinely for each group) to each site, which were randomly
used in rehabilitation for children who have limited placed by a third party in the participant files and
motor ability, at Gross Motor Function Classification were opened after the baseline assessment.
System (GMFCS)17 levels IV and V. However, Ethical approval was granted by Brighton West
some children with severe motor functional disabil- Research Ethics Committee. An overview of the
ity may be able to pedal independently when given randomized controlled trial methodology is shown
additional support on a static bike9 or may be able to in Figure 1.

Downloaded from cre.sagepub.com at THAMMASAT UNIVERSITY on October 9, 2014

Bryant et al. 3

Referred
(n=37)

Excluded (n=2):
1 = Inclusion criteria not met
1 = Did not wish to participate

Block randomization at four sites

(n=35)

Static bike group Treadmill group Control group

(n=11) (n=12) (n=12)

Week 0 Week 0 Week 0

Baseline assessment Baseline assessment Baseline assessment
(n=11) (n=12) (n=12)

Exercise intervention Exercise intervention

Withdrawn (n=2)
1= Hospitalization Losses (n=1) Losses (n=1)
1= Hip pain 1= Sick 1= Sick

Week 6 Week 6 Week 6

Post intervention assessment Post intervention assessment Post intervention assessment
(n=11) (n=9) (n=11)

Losses (n=1) Losses (n=2)

Withdrawn (n=1) 1 = School 2 = School
1 = Botox injections absence absence

Week 12 Week 12 Week 12

Follow-up assessment Follow-up assessment Follow-up assessment
(n=10) (n=9) (n=10)

Losses (n=1)
Refused final
assessment

Week 18 Week 18 Week 18

Follow-up assessment Follow-up assessment Follow-up assessment
(n=9) (n=10) (n=12)

Figure 1. Schematic diagram of the intervention and assessment process.

Downloaded from cre.sagepub.com at THAMMASAT UNIVERSITY on October 9, 2014

4 Clinical Rehabilitation 0(0)

Static exercise bikes (Vision Fitness E3200 •• Static bike protocol (total time <30 minutes):
Upright Cycle), adapted to provide extra postural participants were asked to pedal for as long as
support, were used for the bike group. Trunk, feet they could with a load of 75% of the maximum
and wrist supports were secured when required. achieved in the modified graded exercise test,
Standard treadmills (Tunturi T90) with minimal and the time recorded. After a rest of a few min-
operating speed of 0.5 km/h were used for the tread- utes they were asked if they felt they had recov-
mill group. ered and were ready to continue. They were
Participants were hoisted on and off the static then asked to pedal as fast as they could in short
exercise bike using their designated sling, or onto bursts against the maximum resistance deter-
the treadmill using a ‘standing sling’ (Liko mined in the initial graded exercise test. The
Mastervest) that remained in place for the session number and highest speed of the short bursts
providing partial body weight support using an were recorded. A warm-up and cool-down
overhead tracking hoist. opportunity was provided by the assistant rotat-
Participants in the bike and treadmill groups ing the pedals for participants.
received exercise training three times a week for •• Treadmill protocol (total time <30 minutes): the
six weeks. This duration was chosen as the chil- treadmill speed was increased by 0.1 km/h every
dren have a natural break in their schooling every 10 seconds, and the participant was asked to
six weeks. Sessions lasted approximately 30 min- continue walking until he or she was unable to
utes, including transfers between wheelchair and go faster and began to stumble. The ‘fastest
bike or treadmill. Control group participants walking speed’ was recorded as fastest 10-sec-
received their usual physiotherapy activities dur- ond walk before stumbling. After a rest of a few
ing this period, such as stretching and exercises on minutes they were asked if they felt they had
a mat, standing with a standing frame or recovered and were ready to continue. They
swimming. were then asked to walk for as long as possible
To evaluate the starting level of exercise neces- at 75% of the last ‘fastest walking speed’. A
sary for each child, and to monitor the progress in warm-up and cool-down was provided by walk-
their ability from baseline to six weeks, partici- ing for 30 seconds at the slowest speed.
pants in the bike group performed a modified
graded exercise test based on the McMaster All- The length of time on the bike or treadmill varied
Out Progressive Continuous Cycling Test.18 For the between children according to ability but aimed to
treadmill group assessments of the ‘fastest walking be increased at each session. Verbal encouragement
speed’ achieved on the treadmill were recorded was given to participants to perform their best.
for each participant. These assessments have been To evaluate motor function the Gross Motor
used previously for children with cerebral Function Measures22 GMFM-66 and dimensions D
palsy.9,19,20 (Standing) and E (Walking, Running & Jumping) of
The training protocols used were adapted from the GMFM-88 were measured. One researcher
the American College of Sports Medicine guide- (HW) acted as assessor for all but three of the study
lines for healthy adults,21 which were modified to assessments and was blinded as to which arm of the
bring them within the ability of the participants. study each participant was allocated.
Subsquently, in order to progressively increase the The outcome measures were assessed on four
difficulty of walking on the treadmill, the speed of occasions for each participant, pre intervention
the treadmill was increased, as an alternative to the (week 0), immediately post intervention (week 6)
usual method of increasing the inclination, because and two follow-up assessments (weeks 12 and 18).
fixed ankle foot orthoses were worn by some Secondary outcome measures (speed and duration
participants (which make walking on a slope of exercise) were evaluated for those in the bike and
difficult). treadmill groups at week 0 and week 6.

Downloaded from cre.sagepub.com at THAMMASAT UNIVERSITY on October 9, 2014

Bryant et al. 5

A sample size calculation was performed in order were analysed using mixed design analysis of vari-
to determine the sample size. A sample size of n = ance (ANOVA). An overall group-by-time interac-
30 (10 cases in each arm) was calculated to detect a tion tested whether the average time course differed
difference in GMFM-88E score of 3 between con- between groups. Where data were not normally dis-
trol and experimental arms with power set at 90% tributed the square root of the value was computed
and a significance level of 0.025. This expected dif- for the ANOVA test. For ease of interpretation, raw
ference between arms was based on previous data are used in all tables.
research.9 The significance level was selected to
reduce the chance of type 1 error, as the planned
analysis comprises two independent t-tests, each
Results
comparing the difference in GMFM-E score
between the control arm and experimental arm. Thirty-five participants with bilateral cerebral
Thirty-five participants were recruited to allow for palsy (21 female, 14 male) aged between 8 and 17
attrition. years (mean age 13 years 9 months, SD 2 years 3
Statistical analysis was conducted using SPSS, months) participated in the study. Table 1 displays
version 16 (SPSS Inc., Chicago, IL, USA). the characteristics of participants for each group.
Exploratory analyses assessed data normality and Figure 1 shows a flow diagram of the intervention
the effectiveness of randomization. Significance and assessment process. Three participants did not
level was set at P < 0.05 for all analyses, the risk of complete the study: one participant from the bike
type I error having been corrected for already by group was withdrawn as he was given botulinum
setting the P-value to 0.025 within the sample size toxin injections into his quadriceps; two partici-
calculation. pants from the treadmill group were withdrawn,
Changes in outcome measurements immediately one owing to hospitalization for gastric problems
after the intervention (between baseline at week 0 and one who chose to withdraw due to a reoccur-
and week 6) were calculated, and the data analysed rence of long-standing hip pain.
using unpaired Student’s t-test or Mann–Whitney U During the six-week intervention period the
as appropriate. All analyses compared bike group maximum number of training sessions each partici-
and control group, and treadmill group with control pant could attend was 18. Table 1 shows the number
group. The study was not powered to analyse differ- of training sessions attended by participants. No
ences between the bike and treadmill groups as that adverse incidents were recorded.
was not the aim of the study. Table 2 shows the baseline data, and the change
To investigate whether changes observed in out- in outcome measures post intervention. At baseline
come measurements immediately post intervention no significant differences were found in motor func-
(week 6) were retained at weeks 12 and 18, data tion (GMFM scores) between the bike group and

Table 1. Group characteristics

Group Participants Gender Age (years) GMFCS level Type of cerebral Training sessions
palsy attended

(n) Mean and SD Dyskinetic Spastic Mean and SD

Bike 11 M = 6, F = 5 14.3 (1.9) IV = 8; V = 3 4 7 14.6 (3.1)
Treadmill 12 M = 3, F = 9 13.5 (2.6) IV = 8; V = 4 3 9 13.8 (4.2)
Control 12 M = 5, F = 7 13.8 (2.3) IV = 7; V = 5 7 5 n/a
Total 35 M = 14, F = 21 IV = 23; V = 12 14 21
GMFCS, Gross Motor Function Classification System.

Downloaded from cre.sagepub.com at THAMMASAT UNIVERSITY on October 9, 2014

6 Clinical Rehabilitation 0(0)

–0.2 (2.9) –1.6 (5.6) –1.3 (4.1)

–2.3 (5.3)b –3.9 (4.4)b 0.2 (0.8)
–1.5 (2.1) –1.7 (1.8)b –0.7 (2.5)
the control group (P > 0.05), or between the tread-

(n = 12)
Control
mill group and the control group (P > 0.05). One

Mean change from week 6 to week Mean change from week 6 to

third of all participants failed to score on the
GMFM-88D measure, and half of all participants

Treadmill
(n = 10)
failed to score on the GMFM-88E measure.
Immediately post intervention, at six weeks, a
week 18 (SD)

significant improvement was found in the GMFM-

88D (standing) scores for participants in the static
(n = 9)

bike group compared to the control group (u =

Bike

24.5, P = 0.02), and for the treadmill group com-

pared to the control group (u = 29.5, P = 0.04).
–0.4 (1.6) –1.0 (4.7) –0.7 (3.1)
–3.5 (6.1)b –2.9 (4.8)b –0.0 (1.2)
–1.5 (3.4) –1.5 (1.8)b 0.6 (2.2)
The mean change (SD) in GMFM-88D score
(n = 10)
Control

between baseline and six weeks for the bike group

was +5.9 (6.8); for the treadmill group was +3.7
(4.4) and for the control group was +0.5 (1.9). The
Treadmill

GMFM-88E (walking, running, jumping) scores

(n = 9)

were higher for both treatment groups compared to

the control group, however the difference was not
statistically significant (P > 0.05). The mean
(n = 10)
12 (SD)

change (SD) in GMFM-88E score between base-

Bike

line and six weeks for the bike group was +2.5
Table 2. The mean (SD) baseline data and change in outcome measures for all groups

(4.1); for the treadmill group was +1.1 (2.0) and

39.4 (6.3) 34.5 (12.3) 34.1 (10.2) 1.0 (2.6) 1.5 (5.3) 0.3 (2.5)
1.7 (2.5) 5.9 (6.8)a 3.7 (4.4)a 0.5 (1.9)
2.4 (4.4) 2.5 (4.1) 1.1 (2.0) 0.2 (1.8)

for the control group was +0.2 (1.8). Small

(n = 11)
Treadmill Control
Mean change from baseline to

increases in the GMFM-66 score were observed

group-by-time interaction for data between weeks 6, 12 and 18 (P < 0.05).

for participants between baseline and six weeks in

both the treadmill and bike groups, compared to
(n = 9)

the control group, although the differences were

not statistically significant (P > 0.05).
week 6 (SD)

Both treadmill and bike groups demonstrated

(n = 11)

significant improvements in their exercise capabil-

difference between baseline and week 6 (P < 0.05).
Bike

ity as shown in Table 3. The bike group showed a

significant increase in their maximal speed (P =
0.016), and minutes exercised (P = 0.004). Similar
(n = 12)
Control

improvements were seen in the treadmill group for

their maximal speed of walking (P = 0.011) and
minutes walked (P = 0.004).
1.5 (1.3) 2.5 (2.8)
1.7 (2.9) 2.6 (4.7)

Follow-up assessments revealed a decline in

Treadmill
(n = 12)

the improvements observed immediately post

intervention in gross motor function scores for
Baseline data

both treatment groups. Participants in the bike

(n = 11)

group demonstrated a significant reduction in

Bike

GMFM-88D (standing) scores compared to the

control group (group-by-time interaction F[2,36]
= 3.8, P = 0.030). A similar pattern was seen with
GMFM-88D
GMFM-88E
GMFM-66

bSignificant
aSignificant

participants in the treadmill group compared to

Group

the control group (group-by-time interaction

Downloaded from cre.sagepub.com at THAMMASAT UNIVERSITY on October 9, 2014

Bryant et al. 7

Table 3. The mean (SD) speed and duration of exercise pre and post intervention for both treatment groups

Group Speed (kph) Duration (minutes)

Pre (week 0) Post (week 6) Pre (week 0) Post (week 6)

Bike 6.1 (4.0) (n = 11) 8.9 (4.1) (n = 11) 2.2 (1.8) (n = 11) 5.8 (2.8) (n = 11)
Treadmill 1.35 (0.4) (n = 12) 1.86 (0.6) (n = 9) 2.2 (1.6) (n = 12) 7.4 (4.1) (n = 9)

Table 4. Change in outcome measures for participants from both exercise groups by their level of motor disability
(Gross Motor Function Classification System levels)

Change in scores between baseline and GMFCS level IV (n = 15) GMFCS level V (n = 4)
week 6 for both exercise groups
Mean (95% CI) Mean (95% CI)
GMFM-66 0.2 (–1.3 to 1.7) 5.2 (–3.9 to 14.4)
GMFM-88D 4.4 (1.6 to 7.3) 6.4 (–7.1 to 19.9)
GMFM-88E 1.4 (–0.3 to 3.2) 3.1 (–3.2 to 9.4)
CI, confidence interval.

F[2,35] = 3.5, P = 0.041). With regards to the Discussion

GMFM-88E (walking, running, jumping) scores
participants in the treadmill group showed a sig- Research on exercise interventions for non-ambu-
nificant reduction compared to the control group lant children with cerebral palsy is very limited.9,10
(group-by-time interaction F[2,34] = 4.4, P = The primary aim of this study was to identify exer-
0.020). Participants in the bike group demon- cise interventions effective in improving gross
strated a reduction in GMFM-88E score at follow- motor function ability in non-ambulant children
up, however the difference was not statistically with cerebral palsy. This study found that a six-
significant compared to the control group. week exercise programme either on a static bike or
Table 4 shows a breakdown of the change in treadmill can lead to significant short-term improve-
scores for participants in the exercise groups by ments in motor function (GMFM-88D) compared to
their level of motor disability. While it would appear standard physiotherapy care. Significant improve-
that those more severely disabled showed greater ments were also observed in exercise capability;
improvements, no statistical analysis was performed being able to cycle or walk at a faster speed and for
on the data due to the uneven sample sizes. longer duration. The improvements observed in
After the intervention the participants and phys- motor function declined during the 12-week follow-
iotherapists were asked of their views and experi- up period.
ences of the trial. The children enjoyed the Participants in both the bike and treadmill groups
opportunity to exercise on the static bike or the demonstrated significant improvements in mobility
treadmill. Exercising at this intensity was a new with regards to their ‘standing ability’ (GMFM-
experience for the majority of them, who had not 88D) immediately post intervention, compared to
been ‘out of breath’ before and had not felt tired the control group. Some children were able to stand
from exercise previously. Some of the young people holding on to a support for longer periods. These
and physiotherapists reported that participating in small changes may be of value to the children and
the exercise programme had helped them with their carers in the realms of independence and personal
standing and walking skills. care. Similar improvements in GMFM-88D scores

Downloaded from cre.sagepub.com at THAMMASAT UNIVERSITY on October 9, 2014

8 Clinical Rehabilitation 0(0)

were reported following a six-week intervention As would be expected from an exercise interven-
using a static bike9 and a three-month treadmill tion, participants in both treatment groups signifi-
intervention.10 cantly improved in their exercise capability; being
With regards to changes in walking ability there able to cycle or walk at a faster speed and for longer
was a trend towards initial improvements in the duration at the end of the six-week intervention.
GMFM-88E score for both treatment groups imme- Participants in the treadmill group increased the
diately post intervention, although the differences mean time spent walking by 5 minutes. Similar
were not significantly different compared to the increases in walking time were also reported by
control group. Similar findings have been reported Willoughby et al.16 with a 7-minute increase follow-
previously following strength6 or exercise training ing a nine-week intervention; and by Schindl et al.10
interventions.8 In contrast, Schindl et al.10 found with a 6-minute increase following a 12-week inter-
significant improvements in GMFM-88E scores vention. The improvement observed in mean walk-
(mean increase of 4.3) after 36 treadmill training ing speed of 0.52 kph in the current study was
sessions over 12 weeks. This would suggest a lon- slightly lower than the improvement reported by
ger duration of treadmill training is required to elicit Willoughby et al.16 of 0.75 kph, which may be due to
significant improvements in walking ability. the difference in the length of the interventions. For
The pattern of small but statistically insignificant the bike group, the mean increase in cycling time
improvements in the GMFM-66 scores observed in was 3.6 minutes and cycling speed was 2.5 kph.
the current study between the bike and control The initial improvements in motor function imme-
group, and the treadmill and control group, are simi- diately post intervention were not retained after ces-
lar to other studies with ambulant children.4,7 The sation of the intervention at the six-week follow-up
lack of effectiveness may also reflect a difference assessment, which is consistent with other stud-
between the outcome measures as the GMFM-88 ies.4,9,24 This detraining effect has also been observed
gives credit for each new ‘movement’, while the in healthy children.25 This would suggest further
GMFM-66 attempts instead to credit each new research is needed to explore the sustainability of
‘skill’.22 Furthermore the GMFM-66 has been effects through either ongoing ‘maintenance’ exer-
shown to be less sensitive to changes in motor func- cise programmes, or exercise programmes of longer
tion for children aged 5 and over.23 duration. Recent recommendations for resistance
The improvements observed in gross motor training protocols suggest at least 12 weeks duration
function scores immediately post intervention for to maximize the likelihood of a training effect in chil-
participants in both treatment groups compared to dren and adolescents with cerebral palsy.12
the control group suggest that both modes of exer- At baseline a number of children were unable to
cise have the potential to lead to improved mobility. initiate any of the tasks in the GMFM measures for
While participants in the bike group appeared to standing (GMFM-88D) and walking, running,
show greater improvements than participants in the jumping (GMFM-88E) and subsequently had a
treadmill group for their standing (GMFM-88D) score of zero. Both of these measures have limita-
and walking ability (GMFM-88E), no statistical tions in detecting functional change in individuals
analysis was performed as the study was not pow- with severe motor disabilities due to the potential
ered to analyse differences between these groups. ‘floor effect’ at the lower limits of the measure.
The trends for improvements in motor function Unfortunately, we are not aware of any alternative
suggest there may have been increases in lower measure suitable for children with severe motor dif-
limb muscle strength which would help the children ficulties, which is sensitive enough to detect changes
performing the functional activities such as stand- in gross motor function over time. Future research
ing and walking. Future studies would warrant to develop a more appropriate outcome measure for
investigating changes in muscle properties and this population would be welcomed.
muscle activity in order to identify the possible At the start of the intervention some children in
mechanisms for these improvements. the bike group were unable to pedal continuously

Downloaded from cre.sagepub.com at THAMMASAT UNIVERSITY on October 9, 2014

Bryant et al. 9

unsupervised so were assisted during the exercise shows the potential of two modes of exercise for
sessions until they were able to pedal independently improving standing ability (with regard to standing
(which took several weeks for some participants). transfers) and exercise capability for this popula-
This would suggest the physiotherapists may have tion. Both modes of exercise were acceptable and
overestimated the ability of some children at the enjoyable for the children for whom there are few
time of recruitment. In retrospect, lengthening the other exercise opportunities.
duration of the intervention to include a familiariza-
tion/training period would have been of benefit to
these participants in order to optimize the potential Clinical messages
benefits of the programme. Indeed Verschuren et
al.12 recently suggested incorporating a training •• Exercise interventions are feasible for
period in order to practise the task (i.e. developing non-ambulant children with cerebral
the force to initiate the pedal turning) followed by palsy.
the introduction of more complex tasks for the inter- •• Such interventions can lead to short-term
vention (i.e. pedalling as fast as possible). improvements in gross motor function
On completion of the study the children reported such as with standing abilities.
enjoying the opportunity to exercise on the bike or •• A detraining effect is likely to be seen
treadmill, which was a new experience for the within six weeks of cessation of an exer-
majority of them. cise intervention.
There were several potential limitations to this
study. The major limitation being the small sample
Acknowledgements
size as we were expecting a big effect size.9 Some of
the non-significant findings may reflect a type 2 The authors would like to thank the participants and phys-
error in that the study may have been insufficiently iotherapy staff at the specialist schools: Chailey Heritage
powered to identify a smaller (but still clinically Clinical Services (Lewes); Valence (Westerham); Treloars
significant) effect size. This is indicated by the non- (Alton) and Riverside (Bromley) for their assistance with
significant trends for improvement in the GMFM- the exercise sessions and with data collection. The NIHR
88E data immediately post intervention. Research Design Service-South East provided assistance
The study was not powered to analyse whether with the study design.
there was greater benefit with one mode of exercise This article presents independent research commis-
over the other. In retrospect perhaps this should sioned by the National Institute for Health Research
have been addressed, however the aim of the study (NIHR) under the Research for Patient Benefit funding
was to identify effective interventions where none stream. The views expressed in this publication are those
currently exist, and to offer a choice. Treadmills and of the authors and not necessarily those of the NHS, the
static bikes are not routinely used for children with NIHR or the Department of Health.
severe motor functional disabilities. Therefore in
order to compare the effectiveness of the two modes
Funding
of exercise a larger scale parallel arm randomized
controlled trial is required. This work was supported by the National Institute for
We did not measure the participants’ weight and Health Research (grant number PB-PG-0807-14074).
height so therefore are unable to generalize the findings
from this study. In addition, heart rate recordings were
References
not undertaken while exercising so we do not know the
1. Eek MN and Beckung E. Walking ability is related to mus-
workload intensity, although the children were actively cle strength in children with cerebral palsy. Gait Posture
encouraged to work harder at each session. 2008; 28: 366–371.
In conclusion, while no significant improve- 2. Lin JP. The assessment and management of hypertonus
ments were found in walking ability, this study in cerebral palsy: a physiological atlas (‘road map’). In:

Downloaded from cre.sagepub.com at THAMMASAT UNIVERSITY on October 9, 2014

10 Clinical Rehabilitation 0(0)

Scrutton D, Damiano D and Mayson M (eds) Management 14. Mayston M, Beattie K and Gelkop N. The effect of a tread-
of the motor disorders of children with cerebral palsy, sec- mill training on fitness and function in children with spastic
ond edition. Clinics in Developmental Medicine No. 161. cerebral palsy: preliminary study. Dev Med Child Neurol
London: MacKeith Press, 2004, pp. 86–89. 2006; 48(suppl 105): 1B.1.
3. McNee A, Gough M, Morrissey MC and Shortland AP. 15. Mattern-Baxter K, Bellamy S and Mansoor JK. Effects of
Increases in muscle volume after plantarflexor strength intensive locomotor treadmill training on young children
training in children with spastic cerebral palsy. Dev Med with cerebral palsy. Pediatr Phys Ther 2009; 21: 308–319.
Child Neurol 2009; 51: 429–435. 16. Willoughby KL, Dodd KJ, Shields N and Foley S. Efficacy
4. Scholtes VA, Becher JG, Comuth A, Dekkers H, Van of partial body weight-supported treadmill training com-
Dijk L and Dallmeijer AJ. Effectiveness of functional pro- pared with overground walking practice for children with
gressive resistance exercise strength training on muscle cerebral palsy: a randomised controlled trial. Arch Phys
strength and mobility in children with cerebral palsy: a ran- Med Rehabil 2010; 91: 333–339.
domized controlled trial. Dev Med Child Neurol 2010; 52: 17. Palisano R, Rosenbaum P, Walter S, Russell D, Wood E and
107–113. Galuppi B. Development and reliability of a system to clas-
5. Morton JF, Brownlee M and McFadyen AK. The effects sify gross motor function in children with cerebral palsy.
of progressive resistance training for children with cerebral Dev Med Child Neurol 1997; 39: 214–223.
palsy. Clin Rehabil 2005; 19: 283–289. 18. Bar-Or O and Rowland TW. Pediatric exercise medicine:
6. Dodd JK, Taylor NF and Graham HK. A randomised clini- from physiologic principles to health care application.
cal trial of strength training in young people with cerebral Champaign, IL: Human Kinetics, 2004, p. 347.
palsy. Dev Med Child Neurol 2003; 45: 652–657. 19. Van den Berg-Emons R, van Baak M, de Barbanson D,
7. Fowler EG, Knutson LM, DeMuth SK, et al. Pediatric Speth L and Saris W. Reliability of tests to determine peak
endurance and limb strengthening (PEDALS) for children aerobic power, anaerobic power and isokinetic muscle
with cerebral palsy using stationary cycling: a randomised strength in children with spastic cerebral palsy. Dev Med
controlled trial. Phys Ther 2010; 90: 1–15. Child Neurol 1996; 38: 1117–1125.
8. Verschuren O, Ketellar M, Gorter JW, Helders P, Uiterwaal 20. Maltais D, Oded-Bar O, Pierrynowski M and Galea V.
C and Takken T. Exercise training program in children and Repeated treadmill walks affect physiologic responses in
adolescents with cerebral palsy. Arch Pediatr Adolesc Med children with cerebral palsy. Med Sci Sports Exerc 2003;
2007; 161: 1075–1081. 35: 1653–1661.
9. Williams H and Pountney T. Effects of a static bicycling 21. American College of Sports Medicine. Position stand. The
programme on the functional ability of young people with recommended quantity and quality of exercise for develop-
cerebral palsy who are non-ambulant. Dev Med Child Neu- ing and maintaining cardiorespiratory and muscular fitness,
rol 2007; 49: 522–527. and flexibility in healthy adults. Med Sci Sports Exerc 1998;
10. Schindl MR, Forstner C, Kern H and Hesse S. Treadmill 30: 975–991.
training with partial body weight support in nonambulatory 22. Russell DJ, Rosenbaum PL, Avery LM and Lane M. Gross
patients with cerebral palsy. Arch Phys Med Rehabil 2000; Motor Function Measure (GMFM-66 & GMFM-88) User’s
81: 301–306. Manual. Clinics in Developmental Medicine No.159. Lon-
11. Scianni A, Butler JM, Ada L and Teixeira-Salmela L. Mus- don: MacKeith Press, 2002.
cle strengthening is not effective in children and adolescents 23. Russell DJ, Avery LM, Rosenbaum PL, et al. Improved
with cerebral palsy: a systematic review. Aust J Physiother scaling of the Gross Motor Function Measure for children
2009; 55: 81–87. with cerebral palsy: evidence of reliability and validity.
12. Verschuren O, Ada L, Maltais D, Gorter JW, Scianni A Phys Ther 2000; 80: 873–885.
and Ketelaar M. Muscle strengthening in children and 24. Bania T, Dodd KJ and Taylor N. Habitual physical activity
adolescents with spastic cerebral palsy: considerations for can be increased in people with cerebral palsy: a systematic
future resistance training protocols. Phys Ther 2011; 91: review. Clin Rehabil 2011; 25: 303–315.
1130–1139. 25. Ingle L, Sleap M and Tolfrey K. The effect of a complex
13. Foley S and Dodd K. The effect of a treadmill training pro- training and detraining programme on selected strength and
gramme on walking in school-aged children with cerebral power variables in early pubertal boys. J Sports Sci 2006;
palsy. Dev Med Child Neurol 2006; 48(suppl 105): 1B.2. 24: 987–997.

Downloaded from cre.sagepub.com at THAMMASAT UNIVERSITY on October 9, 2014