The effect of Resistance training on Gait kinematics and Kinetics in

Children with Cerebral Palsy: A Systematic Review

By Daniel Yazbek

Abstract

Aim: This paper reports a systematic review of Progressive resistance training

(PRT) interventions for children with Cerebral Palsy. The sum of randomised

controlled trials (RCT’s) within this review, aims to quantify if regular strength training

increases gait velocity. Background: Secondary problems associated with Cerebral

Palsy affect normal gait mechanics compared to healthy people. Those with CP have

shown to be significantly weaker than their healthy counterparts. Muscle weakness

has been found to negatively affect walking speed and gait efficiency. Methods: A

comprehensive literature search identified all studies of those which contained the

key words Cerebral palsy (CP), Gait and Resistance training. It included 4 electronic

database journals and two internet search engines. Language was limited to English

and was dated from 1998 – 2012. Progressive resistance training studies and there

effects on gait parameters were selected for review. Results: An overall mean affect

0.06 (-0.33 – 0.46) showed that gait velocity favoured the intervention over the

control group. Conclusion: Muscle weakness may not be the only contributor to

poor gait performance. To achieve a greater overall mean increase in gait velocity,

resistive exercise design incorporating repetitions that involve simultaneous agonist

and antagonist muscle contraction during functional movement, should be combined

with other interventions such as gait training, balance and proprioception. If gait

velocity is to be maximised it is imperative to treat the cause of gait compensations

and to address impaired selective voluntary motor control, abnormal stretch reflexes

and to ensure sufficient heel strike at initial contact.

1

Introduction

Cerebral Palsy is a collective term to describe the different categories of neurological

deficits associated within the cerebral areas of the brain (Mockford & Caulton, 2010).

Many researchers believe that non-progressive disturbances such as damage to

blood vessels caused by disrupted oxygen to developing areas of the foetal of infant

brain to be the cause (O'Shea, 2008).

The functional consequences secondary to cerebral lesions include spasticity,

muscle weakness, impaired selective voluntary motor control, hyperreflexia, muscle

hypertonia and poor ability to ambulate (Diane L. Damiano, Laws, Carmines, & Abel,

2006). In those with Cerebral Palsy, it is not that muscle contraction is weak but

rather an inappropriate timing of muscle activation patterns in both agonist and

antagonist muscles, resulting in co-contraction of the joints affected (Prosser, Lee,

Barbe, VanSant, & Lauer, 2010; Prosser, Lee, VanSant, Barbe, & Lauer, 2010).

Gait pattern in young children and adults has been analysed and divided into four

groups (Dobson, Morris, Baker, & Graham, 2007; Winters, Gage, & Hicks, 1987). In

Spastic Hemiplegia, group one and two subjects both exhibit a plantar-flexion

contracture (equinus) on the contralateral side of the cerebral lesion. However, only

group two subjects exhibit foot equinus at all stages of gait including mid stance that

produces an external moment which forces the knee into exaggerated plantar

flexion-extension coupling resulting in knee hyperextension. Both groups

compensate by increasing hip and knee flexion on the affected side to provide an

increase in the swing phase of gait. Group three and four subjects exhibit the same

characteristics of pathological gait as group one and two, however hamstring and

quadriceps overactivity result in insufficient swing phase resulting in vaulting of the

unaffected foot to counter this effect. Hip flexion contracture in group four subject’s

prevent full extension after mid-stance resulting in increased lumbar lordosis to

preserve stride length.

Several researchers have demonstrated a relationship between generalized muscle

weakness and decreased performance in gait, but more specifically with increased

2

energy expenditure as a result of ambulation in those with CP (Bohannon, 1989;

Chen et al., 2012). Likewise, several studies have shown that those with cerebral

palsy exhibit highly inefficient gait patterns and increased energy expenditure, which

not all contributes to an increase in horizontal velocity of the body’s centre of mass

(Ballaz, Plamondon, & Lemay, 2010; Goldberg, Requejo, & Fowler, 2010; Russell,

Bennett, Kerrigan, & Abel, 2007). Logically, this implies that a suitable resistance

training program tailored to cerebral palsy patients may improve gait efficiency and

increase gait velocity.

In a review of the literature, correction of equinus deformity through surgical

interventions in an attempt to improve gait has shown not be successful (Shore,

White, & Kerr Graham, 2010). Unfortunately, greater incidence of equinus and

calcaneal deformity in children with hemiplegia were eminent following surgical

procedures. Furthermore, Stebbins et al (2010) analysed gait pelvic kinematics

before and after foot surgery on twelve subjects with spastic hemiplegia.

Interestingly, increased anterior pelvic tilt was evident in the CP group before and

remained uncorrected after surgery. This provides evidence that anterior pelvic tilt of

the pelvis may not always be a secondary compensation for foot equinus deformity

and may occur as a result of hip flexor tightness, weak abdominals or hip extensors.

This has implications as to whether resistance training may have a positive effect, as

failure to address the cause of compensatory mechanisms may lead to a resistance

training program that doesn’t approach the correct muscles around the joints of the

lower extremity.

In addition, children with Cerebral Palsy show to have seventy-percent reduction in

quadriceps’ rate of force development and knee extensor impulse during the loading

response of gait, compared to healthy controls (Moreau, Falvo, & Damiano, 2012). It

is of great interest whether a traditional resistance training program for the knee

extensors might counter this effect since the loading response of gait may require

the quadriceps to produce force at higher contraction velocity.

Overall, the main purpose of this review is to analyse whether resistance training

increases ambulatory function for those with spastic Cerebral Palsy. Furthermore, an

analysis of abnormal gait kinematics and kinetics during stages of gait in those with

Cerebral Palsy will potentially serve great value for Physical therapists and exercise

3

physiologists in designing appropriate resistance exercise programs to match the

demands that abnormal gait places on the neuromuscular system.

Methods

Identification and selection of literature

A comprehensive literature search was performed using electronic databases,

including Medline (1998 to Feb 2012); SPORTDiscus (1998 to Feb 2012); Web of

Science (1998 to Feb 2012); and Cinahl (1998 to Feb 2012). The following keywords

were used: ‘Cerebral Palsy’, ‘Gait’, ‘walking’ and ‘strength training’, ‘weight training’

or 'resistance training’. To optimize journal article selection, abstracts and full texts

were selected prior to searching. Any Title Abstract without full text identified in the

database, would be used in the two internet search engines, GOOGLE scholar and

Scirus to identify whether a full text of that title would appear.

A study was included if it met the following criteria:

Participants: Children with Cerebral Palsy (spastic hemiplegia & diplegia) <18

years

CP History: varying disability levels

Outcome measures: Gait velocity

Intervention: Strength or resistance training

Study Design: Single or double blinded Randomized controlled trials

For a study to be excluded, the intervention must be of any other, than resistance

exercise or PRT (e.g. cardiovascular exercise, Electro-stimulation, virtual reality,

etc.). Any study that had 5 or less subjects involved, were also excluded. Any paper

that failed to report outcomes at baseline testing prior to the intervention, were also

excluded.

Selection of data and analysis of quality

The quality of each study was determined by internal validity (intention to treat

analysis, blinding study designs, reporting of subject withdrawals within groups and

randomisation), external validity (inclusion/exclusion criteria) and power analysis

(sample size calculation). Assessment of quality was accomplished by the

attainment of a critical appraisal skills programme (CASP) (Guyatt et al., 1995). This

4

programme tool involves 10 screening questions that determine the quality of RCT’S.

Each question will be answered ‘yes’, ‘can’t tell’ and ‘no’ for each RCT, and will be

awarded1 point, zero points or a deduction point respectively, with a total score out

of 10.

Data analysis and synthesis

A mixed method approach was adopted for data synthesis: a narrative review of the

results supplemented by vote counting (Goodwin et al., 2008). Vote counting method

syntheses results by listing all outcomes of each study and identifying the direction of

effect for each outcome. The direction was rated positive if significant differences

were reported in favour of the intervention, negative if the difference supported the

control, equivocal if no significant difference was reported between groups.

Significance was set at p < 0.05.

Meta-analysis was undertaken using MetaEasy Excel add-in (StatAnalysis;

Kontopantelis & Reeves, 2009). A standardised effect size was calculated for each

study and expressed in standard deviation units. An overall effect was calculated

using the DerSimonial-Laird method (DerSimonial & Laird, 1986).

Results

The initial search identified 232 journal articles, and 16 studies remained after the

initial screening. Applying the exclusion/inclusion criteria resulted in only two

randomized controlled trial studies (Figure 1). Due to scarcity of RCT’s, four pre-test-

post-test experimental designs were incorporated so that we could avoid a

conservative approach in analysing the effects that resistance training has on gait in

CP patients. However, only three (two RCT’s & 1 experimental-control design) were

inputted into meta-analysis for synthesis of results. This experimental design was the

only study that incorporated a control group compared to the other non-RCT’s.

Table 1 summarises the quality of the 6 studies. Two studies were of moderate

quality (score ≥ 4 & ≤ 6) and four of low quality (score ≤ 3). All of the included studies

provided an extensive rationale for the use of resistance training as an intervention

for CP. Additionally, selection criteria was defined for all the studies. Only two

studies were RCT’s (Scholtes et al., 2012; Unger, Faure, & Frieg, 2006). Power and

5

sample size calculations were only reported in two studies (D. L. Damiano & Abel,

1998; Scholtes, et al., 2012).

Only Three studies reported assessor blinding of pre-testing results during post-

intervention testing (D. L. Damiano & Abel, 1998; Scholtes, et al., 2012; Unger, et al.,

2006). Four studies did not perform an intention to treat analysis (Eagleton, Iams,

McDowell, Morrison, & Evans, 2004; Eek, Tranberg, Zügner, Alkema, & Beckung,

2008; Scholtes, et al., 2012; Unger, et al., 2006). Instead, only those subjects upon

completion of the intervention were accounted, for determining main outcomes.

Three studies reported adequate concealment of randomisation (Lee, Sung, & Yoo,

2008; Scholtes, et al., 2012; Unger, et al., 2006).

6

‘

Figure 1. Progression of search for relevant studies

7

Searched databases (n = 4)

Search engines (n = 2)

Potentially relevant literature identified

(n = 232)

Using key words “Cerebral Palsy” “Gait”, “walking”, “strength training”, “resistance

training” or “exercise”

1998 – 2012; limited to English

Papers excluded

(n =216)

Not relevant to review

Papers excluded

(n = 10)

Not meeting inclusion criteria or within exclusion criteria

Papers reviewed

(n = 16)

Papers meeting inclusion criteria

(n = 6)

Included studies

(n = 6)

RCT’s = 2

Included participants

(n = 126)

R.T. = 83

Controls = 43

-Intervention involving adults (n = 2)

- Inappropriate intervention (n = 3)

-Other outcome measures (n =2)

-Inappropriate study design (n = 3)

Table 1 description quality of selected studies

Study Design Rationale Described?

Power/sample size calculations

presented?

Selection criteria

described?

Assessor/participants blinded?

Adequate concealment of randomisation?

Intention to treat analysis

performed?

*CASP Quality score

Scholtes et al. (2012) Matched RCT Parallel Yes Yes Yes Yes Yes no (per protocol

analysis) 6 (mod)

Unger et al. (2006) Matched RCT Parallel Yes No Yes Assessor blinded to pre-test results. Yes Not reported 4 (mod)

Damiano et al. (1998)

Pre-test-post-test experimental design Yes Yes Yes Assessor blinded to

pre-test results Not reported Not required 3 (Low)

Eagleton et al. (2004)

Pre-test-post-test experimental design Yes No Yes Not reported Not reported Not reported 1 (Low)

Eak et al. (2008) Pre-test-post-test experimental design Yes No Yes Not reported No Not reported 2 (Low)

Lee et al. (2008) Pre-test-post-test experimental design Yes No Yes Not reported Yes Not required 2 (Low)

* CASP score out of maximum of 10

8

Table 2 Variations and outcomes of selected studiesStudy Subjects Resistance Training Duration Outcome measures Major findings

Scholtes et al. (2012)

1) 24 PRT ambulant spastic CP 10.4 ± 1.1 yrs. 2) 25 CON CP

10.3 ± 2.3yrs

PRT - Leg Press, Sit to stand, half knee raise,

lateral step up, 3 sets 8 reps, load ↑ 5% once reached 8RM CON -

Physio care

36 sessions (3x/wk., 12

weeks)

POM - timed 10m walk test (cadence, velocity & stride length) & timed stair

test. SOM - Spasticity, ROM, AMP (Wingate) & isometric strength

Sig ↑in comfortable & fast walking speed, ↔ cadence and ↑ stride length. Sig ↑isometric muscle strength 8% & leg

press 14%, ↔anaerobic power & spasticity, sig ↓KF ROM in PRT > CON

Unger et al. (2006)

1) 21 PRT CP patients 15.86 yrs., 1 orthotic, 2 assistive

devices (crutch & wheelchair) 2) 10 CON CP patients

16.38yrs

Resistance exercises not reported PRT - 8-12

exercises of 28 station circuit - upper and

lower extremities CON - no explanation

1-3 x per week/8 weeks

Stride length, cadence & velocity, knee angle at heel strike, crouch gait at mid-stance & self-perception Q-

airre

Sum of joint (H,K,A)⁰ at midstance ↓ sig (p < 0.05) compared to CONS, ↔ stride length, velocity & cadence compared to

CON & Knee⁰ at heel-strike sig ↓

Damiano et al. (1998)

1) 6 Diplegics 8.3yrs, 5 Hemiplegics 9.2yrs 2) no Con

Resistance exercises not reported, 4 sets 5 reps

65% 1RM Isometric strength. Hemiplegics unilateral & diplegics

bilateral training

18 sessions (3x/wk., 6

weeks)

Gait velocity, stride length, cadence, % stance & % double support gait,

strength, GMFM & EEI

Hemiplegics & Diplegics sig↑ stride velocity & cadence, no sig ↑EEI & Stride

length, trend ↑% mid stance phase, trend ↓% double support

Eagleton et al. (2004)

1) 7 PRT CP patients (12-20 yrs.) 2) no CON

Free weights & machines, 80% 1RM 10 reps, trunk, hip, knee &

18 sessions (3x/wk., 6

weeks)

muscle strength, Stride length, cadence, velocity & EEI Sig ↑ in all outcome measures (p < 0.05)

9

ankle flex & ext.

Eak et al. (2008)

16 subjects with Spastic Diplegia (2 females, 14 males) GMFCS level 1, n = 10 (12.2 ± 1.8yrs) & level 2, n = 6 (13 ± 2

yrs.)

Resistance exercises not reported. 3 sets 10 reps progressive increase in load from 1st - 3rd set

24 sessions (3x/week, 8

weeks)

3-D Gait analysis, Muscle strength, GMFM assessment, joint ROM and

spasticity

No sig ↑ gait velocity, sig ↓ cadence, ↑trend stride length, Spasticity ↔, sig ↑ in all hip muscle groups and knee flexors, sig ↑ hip ext. moment & plantar flexing

generating power at push-off

Lee et al. (2008)

1) 9 PRT (4 Diplegic & 5 hemiplegic) 6.3 ± 2.1 yrs. 2) 8

CON (5 Diplegic & 3 Hemiplegic) 6.3 ± 2.9 yrs.

PRT - Squat to stand, lateral step up, and stair

walk up and down. 2 sets 10 reps load 0.25,

0.45 or 0.9kg adjustable weight cuffs as

progression CONS - Physio, ROM exercise &

Gait training

15 session (3x/week, 5

weeks)

Gait (velocity, stride length, cadence, % single & double limp support)

Muscle strength, spasticity & GMFM

Sig ↑ max hip extensor strength, Sig ↑ gait velocity & stride length, ↑ trend for

% single leg support & ↓ trend for % double limb support, Sig ↑ lateral step

ups & squat to stand, Sig ↑ GMFM score D & E compared to CONS, ↔ spasticity

 

10

Participants

Three studies lacked control groups (D. L. Damiano & Abel, 1998; Eagleton, et al.,

2004; Eek, et al., 2008). Exclusion and inclusion criteria were similar in four studies

(Eagleton, et al., 2004; Eek, et al., 2008; Scholtes, et al., 2012; Unger, et al., 2006).

However, one study reported that hemiplegic children had to demonstrate at least

20% asymmetry in strength values in a minimum of two of eight muscle groups

tested on their more involved extremity in comparison with contralateral extremity

and Diplegics to have at least 50% weakness from healthy normal (D. L. Damiano &

Abel, 1998).

Only one study reported to exclude those patients with fixed contracture at the hip

and knee joints (Lee, et al., 2008). Three studies failed to report categorisation group

of CP subjects (Eagleton, et al., 2004; Scholtes, et al., 2012; Unger, et al., 2006).

Only one of the three experimental-control groups reported statistical p-values

illustrating no baseline significance between groups (Scholtes, et al., 2012). One

study contained an inappropriate difference in sample size between the intervention

and control group (Unger, et al., 2006). Number of subjects within the intervention

group almost doubled that of the control group.

Two studies failed to recruit an appropriate sample size and resorted to using non-

parametric statistics for data analysis (Eagleton, et al., 2004; Eek, et al., 2008).

Interventions

Two studies included an individual analysis of each child’s muscle strength and gait

pattern abnormalities to program appropriate muscle strengthening exercises (Eek,

et al., 2008; Unger, et al., 2006). However, these studies did not report actual

strengthening exercises to be carried out. Only two out of the six studies explained

the strengthening exercises that were to be carried out through the intervention (Lee,

et al., 2008; Scholtes et al., 2008). These two studies were almost homogenous in

regard to resistive exercise selection but varied greatly in the duration of the entire

intervention with five and thirty-six week sessions carried out respectively. Only one

study reported individualized strengthening exercises that were to be carried out

depending on CP categorisation, being Diplegic or hemiplegic (D. L. Damiano &

11

Abel, 1998). Hemiplegics strengthened the unaffected unilateral lower extremity and

diplegics strengthened both extremities.

In regards to the three studies involving controls, only one study failed to report what

type of intervention they received (Unger, et al., 2006). Interestingly, Lee et al (2008)

reported in detail the intervention the controls carried out, however this intervention

was comprised of ROM, gait training and exercise, which was not matched by the

experimental group. In regard to exercise selection, two studies included resistive

exercises for the trunk (Eagleton, et al., 2004; Unger, et al., 2006). In addition, these

were the only studies that reported concurrent lower and upper body strengthening.

Outcomes

The outcomes for all the studies have been summarised (table 2) and presented in

narrative form. Most studies differed in their intervention, duration and outcome

measures. The most common outcome measures between the six studies were Gait

velocity.

12

Gait velocity

As can be seen in figure 2, three of the six highly scored studies produced an overall

effect to suggest that progressive resistance exercise most likely results in a small

but significant overall mean increase in gait velocity (standardised mean difference:

0.0625; 95% CI: -0.3323 - 0.4573). Testing for heterogeneity was not significantly

different (X² = 1.26, df = 2, p = 0.53)

Two studies reported no change in gait velocity (Eek, et al., 2008; Scholtes, et al.,

2012; Unger, et al., 2006).

13

Overall 95% CI

Study

Figure 2. Meta-analysis for Gait velocity

Favours InterventionFavours control

Standardised Mean Difference (Lower, Upper 95% CI)

0.16 (-0.8 – 1.1)

-0.26 (-0.94 – 0.43)

0.24 (-0.32 – 0.8)

0.06 (-0.33 – 0.46)

Stride length

Two studies did not improve stride length as an outcome (D. L. Damiano & Abel,

1998; Scholtes, et al., 2012; Unger, et al., 2006), whereas one study reported a trend

increase (Eek, et al., 2008). However, three studies found a significant increase in

stride length (Eagleton, et al., 2004; Lee, et al., 2008; Scholtes, et al., 2012).

Cadence

Only one study reported a reduction in cadence (Eek, et al., 2008), whereas two

studies showed no difference after intervention (Scholtes, et al., 2012; Unger, et al.,

2006)

Other gait parameters

Only Two studies reported Energy expenditure Index (EEI). One study found no

significant differences in energy expenditure after the intervention (D. L. Damiano &

Abel, 1998). According to Eagleton et al (2004), there was a decrease in EEI in four

of the seven subjects despite increased gait velocity after the intervention. However,

three of the seven subjects increased there EEI due to increases in gait velocity,

suggesting no change in energy expenditure.

Two studies reported both a trend decrease in % of double support phase and a

trend for the increase in % mid-stance support phase of the gait cycle (D. L.

Damiano & Abel, 1998; Lee, et al., 2008).

Strength

Two studies reported an increase in strength via manual muscle testing (Eagleton, et

al., 2004; Lee, et al., 2008). Unger et al (2006) failed to report strength

measurements at post testing. Eak et al (2008) found an increase in all hip muscle

groups and knee flexors via myometer testing. Only one study reported an increase

in strength by 14% using an exercise based machine (leg press) that was used

during the intervention (Scholtes, et al., 2012). Damiano et al (1998) found increase

in strength on the affected side in hemiplegics with no change on the unaffected

side, whereas diplegics increased their strength in targeted muscles with a trend

decrease in antagonist muscle strength.

14

Spasticity

There was no increase in spasticity reported in three studies (Eek, et al., 2008; Lee,

et al., 2008; Scholtes, et al., 2012)

Range of motion

Three studies reported changes in ROM after the intervention. A reduction in knee

flexion angle at heel strike was reported in two studies (Scholtes, et al., 2012; Unger,

et al., 2006). Eak et al (2008) found a significant increase in hamstring length but

only through goniometric measurement. Two Studies reported no change in

antagonistic muscle length after strengthening the agonist (D. L. Damiano & Abel,

1998; Lee, et al., 2008).

Discussion

Although a small but significant overall mean increase in gait velocity, it is clear that

there are factors other than muscle weakness which might contribute to an

insufficient walking speed.

Unger et al (2006) reported increased knee angle at heel strike which may have

been a consequence of increased knee extension moment during swing phase as a

result from increased inhibition of hamstrings. Resistance training programs for

agonist muscles have shown to reduce antagonist activation to allow greater net

force output (Carolan & Cafarelli, 1992; Tillin, Pain, & Folland, 2011). However,

Damiano, Martellota, Sullivan, Granata & Abel (2000) did not show correlation

between co-contraction and weakness which seems counter-intuitive because

antagonist activity could diminish the net force contribution of the agonist, unless the

agonist simultaneously increased its activity. Even though potential negative effects

of excessive co-contraction include greater total muscle activation during net force

production and altered movement quality and quantity, co-contraction may still be a

useful compensatory strategy in CP to increase joint stability, limit degrees of

freedom, or to allow the motor system to respond to perturbations. In Unger et al

(2006), resistance training may have increased agonist muscle recruitment with

corresponding relaxation of antagonists. However this may have reduced joint

15

stability in CP and compromised force regulation. Therefore, we need to understand

the purpose of co-contraction before setting appropriate resistance training programs

for those with CP, as increased joint stiffness may either reduce net force output or

increase joint stability. This mechanism may only be true if increased knee angle at

heel strike was a consequence of increased relaxation of antagonists.

Goldberg, Requejo & Fowler (2010) demonstrated that Children with good selective

voluntary motor control (SVMC) are more capable of moving out of synergy during

the swing phase of gait (hip flexion with knee extension), while children with poor

SVMC are constrained to move in synergy (simultaneous hip and knee flexion).

Swing phase inappropriate SVMC may be the reason Unger et al (2006) showed no

increase in stride length or gait velocity despite reduced crouch at mid stance. This

suggests that resistance training may have improved mid-stance centre of mass, but

SVMC might become a much higher contributor to the swing phase.

Although the effects of resistance training in Unger et al (2006) might have improved

crouch at mid stance, there are other factors within pathological gait that needs to be

addressed. Russell, Bennett, Kerrigan & Abel (2007) showed that CP patients have

increased knee flexion at double support phase and increased plantar flexion in

single limb stance contributing to wasting of energy and increased centre of mass

(COM) oscillation. Only by addressing equinus foot, there may not need increased

contralateral knee flexion during double support to minimize joint compression stress

upon initial contact, thereby minimizing oscillation and improving gait velocity. Unger

et al (2006) did not include specific exercises for musculature around the ankle and

foot and it has been shown that greater plantar flexion-dorsiflexion range of motion

contributes to increased gait velocity and increased ankle range of motion (Ballaz, et

al., 2010). Additionally, Van der Krogt, Doorenbosch, Becher & Harlaar (2009) found

that with increasing walking speed also increased equinus at foot contact and that

the potential increase in walking speed might have been cancelled out due to

amplified stretch reflexes contributing to increase upward rather than horizontal

momentum therefore creating no difference in gait velocity. Intervention strategies

aimed to ensure heel initial contact may reduce gravitational potential energy

(reduction of the triceps surae stretch reflex) and kinetic energy at impact of

contralateral leg, therefore minimizing vertical oscillation and knee flexion during

double leg stance phase.

16

Furthermore, there was a weak correlation between gait joint kinetics and isometric

strength using a dynamometer (Dallmeijer, Baker, Dodd, & Taylor, 2011). Twenty-

five subjects with bilateral Spastic cerebral Palsy with significant reduction in

isometric strength still exhibited joint moment curves similar to normal values during

gait. Ankle joint moments during gait in CP exceed greatly than that during isometric

strength testing. This indicates that isometric strength doesn’t contain information

regarding dynamic muscle function and that muscle spasticity may in some way

activate a velocity dependant stretch reflex during gait which is not possible during

static conditions. This suggests that abnormal neural control mechanisms that

amplify joint moments may retard horizontal momentum, such that seen in bounce

gait. Therefore, focussed efforts should be in understanding how those with CP can

utilize the velocity dependant stretch reflex to maximize horizontal momentum.

Interestingly, the muscles that may have been targeted in Unger et al (2006), which

contributed to reduced crouch gait may have not positively contributed to terminal

stance or pre-swing. It was shown by Goldberg, Ounpuu & Delp (2003) that knee

extension moments during swing phase were the same between controls and those

with Spastic diplegia. This meant that stiff knee gait may not always be a function of

rectus femoris spasticity but rather inadequate hip flexor moments during pre-swing

which was shown to be significantly weaker in those with CP. Physical therapists

must be mindful that programming for resistance training should take into account all

phases of gait and how they interact with each other to produce normal function.

Exercises that increase hip flexor power at pre swing may have resulted in an

increase flexed knee during swing phase, reduced segment inertia, hip hiking and

consequently increased efficiency and velocity.

Neither study within this review applied rate of force development protocols.

However, several studies have analysed voluntary joint moments in those children

with CP. Compared to controls, they demonstrated differences in moment generation

profiles, including decreased maximum voluntary isometric contractions, decreased

maximum rates of moment development, relaxation and increased time needed to

generate and reduce moments (Downing, Ganley, Fay, & Abbas, 2009; Tammik,

Matlep, Ereline, Gapeyeva, & Pääsuke, 2008). Furthermore, Moreau & Damiano

(2012) noted that rate of force development and impulse were more important than

maximal force development in the loading response, as this phase requires less than

17

two-hundred milliseconds to produce optimal force. Although it would be tempting to

improve rate of force development in CP patients, there may be many increased

risks associated with it such as bone fractures and muscle-tendon strains. It may be

even more dangerous to incorporate this type of training method if a therapist

assumes weakness and no other parameter to be a contributing cause for

dysfunctional gait.

Conversely, Lee et al (2012) showed that increased muscle strength related to an

improvement in gait velocity. Several studies have shown that Muscle weakness

assessed from isometric testing correlated to low gross motor function classification

scores, which in turn correlated with reduced gait efficiency and velocity (D. L.

Damiano, Kelly, & Vaughn, 1995; Eek & Beckung, 2008; Rose & McGill, 2005;

Thompson, Stebbins, Seniorou, & Newham, 2011). However, Lee et al (2012)

showed that manual muscle testing of the hip extensors were the only muscle group

to show a correlation to increased gait velocity and stride length. This may be due to

the fact that manual muscle testing only evaluates isolated muscle groups at

localized joints. Additionally, pathological gait is defined by interactions of multiple

limitations, co-contractions and muscle synergies and that manual muscle testing

focuses on primary and secondary problems, while pathological gait is characterized

by compensation mechanisms (tertiary problems) to overcome the primary and

secondary problems (Rose & McGill, 2005). Therefore, it may not be that muscle

strength associated with manual testing caused an increase in stride length and

velocity but rather the improvement in balance and proprioception associated with

the functional sit to stand and lateral step ups.

Although Unger et al (2006) disregarded weakness to be a rate limiter in improving

gait, perhaps eight weeks of training was not sufficient to realize strength gains. In

addition, Unger et al (2006) failed to report post strength measures and had there

been an increase in strength, then only can we say that weakness was a rate limiter

in improving gait. It is possible that Unger et al (2012) programmed resistance

station circuits that may have been comprised of machine based rather than

functional movements of which reported in Lee et al (2012) and Scholtes et al

(2012). Moreover, it may be that increased gait velocity found in Lee et al (2012) was

related due to the fact that those with fixed contracture at the hip and knee joints

were excluded from the subject criteria. It is interesting to note that the control group

18

within Lee et al (2012) received physical therapy, range of motion exercises and gait

training that were unmatched by the intervention group, despite increased gait

velocity favouring the latter. This might be due to the powerful effect that functional

resistive exercise has on gait rather than the unlikely harmful exercise effects carried

out by the control group. In fact it has been shown that gait training results in

increased walking speed over short distances, improved gross motor skills such as

static and dynamic balance and general gait parameters (Mutlu, Krosschell, & Spira,

2009; Willoughby, Dodd, & Shields, 2009). Therefore, it is illogical to assume that the

intervention carried out by the control group had negative effects on gait which

favoured the intervention group.

Damiano et al (1998) suggested that biomechanical limitation of active and passive

motion caused by spasticity, inadequate muscle length, or abnormal dynamic activity

limited an increase in stride length for those with CP. Additionally, stride length was

not related to strength in lower limbs. Conversely, Lee et al (2012) found an increase

in stride length which rejected the findings of Damiano et al (1998). Failure to

compare and contrast resistance exercise intervention between the two studies is

partly due to Damiano et al (1998) not reporting the resistance exercises to be

carried out. Therefore we assume that the subjects within Lee et al (2012) improved

stride length partly due to increased hip extensor strength which contributed to an

increase hip flexor movement at terminal stance, which resulted in increased angle

of hip flexion at swing phase and therefore stride length. Secondly, increased

ipsilateral strength in single limb support from hip girdle strengthening might have

contributed to larger contralateral limb progression. Subjects within Damiano et al

(1998), increased there velocity by an increase in cadence. This increase in cadence

is probably why energy expenditure index did not differ before and after the

intervention. Children who increased there velocity after training became less

efficient before strengthening, whereas children who had minimal increase in velocity

increased efficiency. Therefore, efficiency is not always related to gait velocity and

programming for resistance training should determine whether the outcome would be

increased cadence or stride length as each parameter may affect efficiency

independently. An increase in velocity through increase in stride length produces

greater energy efficiency during gait. In addition, We would have not seen an

19

increase in gait velocity in Damiano & Abel (1998) if muscle target criteria were such

that only weakest muscles were strengthened.

Although not reported, Unger et al (2006) and Eak et al (2008) implemented ‘specific

resistance exercises’ based on gait pattern abnormalities which did not correlate to a

significant increase in walking speed. This explains that the clinical evaluation of gait

in Cerebral Palsy is far too complex to ensure appropriate resistance exercise

design. Both these studies involved patients who exhibited crouch gait. Physical

therapists may assume that knee extensor strengthening be imperative in reducing

knee flexion moment, improving crouch gait and patella tendon strain. However, van

der Krogt et al (2010) demonstrated that by increasing the degree of crouch gait in

stance position contributed to an increase in stiff-knee gait. This was explained by

insufficient hip extension which caused increased gravitational moments about the

shank acting to extend the knee as it lay more in a horizontal position relative to the

floor. Conversely, with a reduction in crouch gait, lead to increased hip extension

which allowed increased gravitational flexion moments about the thigh segment

which in turn passively increased knee flexion. Therefore, stiff knee gait during swing

can occur purely as the dynamical result of crouch, rather than from altered muscle

function, pathoneurological control or muscle weakness. Consequently, It should be

realized the clinical gait assessment in designing resistance exercise has its

limitations.

As mentioned, Eak et al (2008) ignored the relevance of muscle weakness to

impaired gait. Although there was significant increase in plantar flexion generating

power at push off and increase hip extension moment, there was no change in gait

velocity and there was a reduction in cadence. Perhaps the reduction in cadence

was reflected for the trend increase in stride length which may have been a function

of increased hip and knee joint stability from the resistance exercises. The increase

in joint kinetics was reflected possibly from increased hip and knee stability during

stance so that it enables sufficient power to be produced by the gastrocnemius.

However if cadence were to be maintained, there may have been increased gait

velocity. This implies that exercise therapists should aim to maintain both gait

characteristics as resistance exercise may increase stride length at the expense of

cadence (or vice versa) thereby contributing to little or no increase in gait velocity.

20

This paper presents many limitations based on the poor quality of studies available

to select for review. Although Meta-analysis revealed no significant difference

between studies in detecting heterogeneity, it may be unreliable due to the small

amount of studies inputted. Therefore, It is still difficult to analyse the greatest effect

at which the independent variables influenced gait velocity.

Whether duration, intervention protocol, CP category or inclusion criteria influences

gait velocity is still a question that needs to be addressed. Furthermore, we cannot

fully interpret and analyse the difference in strength interventions as more than half

of the selected studies failed to report actual exercises to be carried out. The scarcity

of randomized controlled trials within this field of study may reflect the apprehensive

attempt to include resistive exercise in those with CP due to the fear of exercise

related spasticity or contracture. However, this review showed that exercise did not

contribute to negative effects such as muscle spasticity. Therefore future research

should be aimed at developing higher quality studies (RCT’S) that keep intervention

duration, CP category and inclusion criteria constant so that we can more confidently

suggest that resistance training does have an impact on gait velocity.

Conclusion

As a result, resistance training for those with CP should be part of an exercise

program. However, this paper clearly presents that muscular strength is only one

minor aspect to improving gait velocity. If gait velocity is to be maximised, physical

therapists must determine which aspects relating to gait need to be addressed and

subsequently improved upon. In addition, the phases of gait are interdependent on

each other for optimal walking efficiency and therefore a dysfunctional phase of gait

might be the effects caused by the phase that preceded it.

21

22

References

Ballaz, L., Plamondon, S., & Lemay, M. (2010). Ankle range of motion is key to gait efficiency in adolescents with cerebral palsy. Clinical Biomechanics (Bristol, Avon), 25(9), 944-948.

Bohannon, R. W. (1989). Is the measurement of muscle strength appropriate in patients with brain lesions? A special communication. Physical Therapy, 69(3), 225-236.

Carolan, B., & Cafarelli, E. (1992). Adaptations in coactivation after isometric resistance training. Journal Of Applied Physiology (Bethesda, Md.: 1985), 73(3), 911-917.

Chen, C. L., Lin, K. C., Wu, C. Y., Ke, J. Y., Wang, C. J., & Chen, C. Y. (2012). Relationships of muscle strength and bone mineral density in ambulatory children with cerebral palsy. Osteoporosis International: A Journal Established As Result Of Cooperation Between The European Foundation For Osteoporosis And The National Osteoporosis Foundation Of The USA, 23(2), 715-721.

Dallmeijer, A. J., Baker, R., Dodd, K. J., & Taylor, N. F. (2011). Association between isometric muscle strength and gait joint kinetics in adolescents and young adults with cerebral palsy. Gait & Posture, 33(3), 326-332.

Damiano, D. L., & Abel, M. F. (1998). Functional outcomes of strength training in spastic cerebral palsy. Archives Of Physical Medicine And Rehabilitation, 79(2), 119-125.

Damiano, D. L., Kelly, L. E., & Vaughn, C. L. (1995). Effects of quadriceps femoris muscle strengthening on crouch gait in children with spastic diplegia. Physical Therapy, 75(8), 658-667.

Damiano, D. L., Laws, E., Carmines, D. V., & Abel, M. F. (2006). Relationship of spasticity to knee angular velocity and motion during gait in cerebral palsy. Gait & Posture, 23(1), 1-8.

Dobson, F., Morris, M. E., Baker, R., & Graham, H. K. (2007). Gait classification in children with cerebral palsy: a systematic review. Gait & Posture, 25(1), 140-152.

23

Downing, A. L., Ganley, K. J., Fay, D. R., & Abbas, J. J. (2009). Temporal characteristics of lower extremity moment generation in children with cerebral palsy. Muscle & Nerve, 39(6), 800-809.

Eagleton, M., Iams, A., McDowell, J., Morrison, R., & Evans, C. L. (2004). The effects of strength training on gait in adolescents with cerebral palsy. Pediatric Physical Therapy: The Official Publication Of The Section On Pediatrics Of The American Physical Therapy Association, 16(1), 22-30.

Eek, M. N., & Beckung, E. (2008). Walking ability is related to muscle strength in children with cerebral palsy. Gait & Posture, 28(3), 366-371.

Eek, M. N., Tranberg, R., Zügner, R., Alkema, K., & Beckung, E. (2008). Muscle strength training to improve gait function in children with cerebral palsy. Developmental Medicine And Child Neurology, 50(10), 759-764.

Goldberg, E. J., Requejo, P. S., & Fowler, E. G. (2010). Joint moment contributions to swing knee extension acceleration during gait in children with spastic hemiplegic cerebral palsy. Journal Of Biomechanics, 43(5), 893-899.

Guyatt, G. H., Sackett, D. L., Sinclair, J. C., Hayward, R., Cook, D. J., & Cook, R. J. (1995). Users' guides to the medical literature. IX. A method for grading health care recommendations. Evidence-Based Medicine Working Group. JAMA: The Journal Of The American Medical Association, 274(22), 1800-1804.

Lee, J. H., Sung, I. Y., & Yoo, J. Y. (2008). Therapeutic effects of strengthening exercise on gait function of cerebral palsy. Disability And Rehabilitation, 30(19), 1439-1444.

Mockford, M., & Caulton, J. M. (2010). The pathophysiological basis of weakness in children with cerebral palsy. Pediatric Physical Therapy: The Official Publication Of The Section On Pediatrics Of The American Physical Therapy Association, 22(2), 222-233.

Moreau, N. G., Falvo, M. J., & Damiano, D. L. (2012). Rapid force generation is impaired in cerebral palsy and is related to decreased muscle size and functional mobility. Gait & Posture, 35(1), 154-158.

24

Mutlu, A., Krosschell, K., & Spira, D. G. (2009). Treadmill training with partial body-weight support in children with cerebral palsy: a systematic review. Developmental Medicine And Child Neurology, 51(4), 268-275.

O'Shea, T. M. (2008). Diagnosis, treatment, and prevention of cerebral palsy. Clinical Obstetrics And Gynecology, 51(4), 816-828.

Prosser, L. A., Lee, S. C. K., Barbe, M. F., VanSant, A. F., & Lauer, R. T. (2010). Trunk and hip muscle activity in early walkers with and without cerebral palsy--a frequency analysis. Journal Of Electromyography And Kinesiology: Official Journal Of The International Society Of Electrophysiological Kinesiology, 20(5), 851-859.

Prosser, L. A., Lee, S. C. K., VanSant, A. F., Barbe, M. F., & Lauer, R. T. (2010). Trunk and hip muscle activation patterns are different during walking in young children with and without cerebral palsy. Physical Therapy, 90(7), 986-997.

Rose, J., & McGill, K. C. (2005). Neuromuscular activation and motor-unit firing characteristics in cerebral palsy. Developmental Medicine And Child Neurology, 47(5), 329-336.

Russell, S. D., Bennett, B. C., Kerrigan, D. C., & Abel, M. F. (2007). Determinants of gait as applied to children with cerebral palsy. Gait & Posture, 26(2), 295-300.

Scholtes, V. A., Becher, J. G., Janssen-Potten, Y. J., Dekkers, H., Smallenbroek, L., & Dallmeijer, A. J. (2012). Effectiveness of functional progressive resistance exercise training on walking ability in children with cerebral palsy: a randomized controlled trial. Research In Developmental Disabilities, 33(1), 181-188.

Scholtes, V. A., Dallmeijer, A. J., Rameckers, E. A., Verschuren, O., Tempelaars, E., Hensen, M., & Becher, J. G. (2008). Lower limb strength training in children with cerebral palsy--a randomized controlled trial protocol for functional strength training based on progressive resistance exercise principles. BMC Pediatrics, 8, 41-41.

25

Shore, B. J., White, N., & Kerr Graham, H. (2010). Surgical correction of equinus deformity in children with cerebral palsy: a systematic review. Journal Of Children's Orthopaedics, 4(4), 277-290.

Tammik, K., Matlep, M., Ereline, J., Gapeyeva, H., & Pääsuke, M. (2008). Quadriceps femoris muscle voluntary force and relaxation capacity in children with spastic diplegic cerebral palsy. Pediatric Exercise Science, 20(1), 18-28.

Thompson, N., Stebbins, J., Seniorou, M., & Newham, D. (2011). Muscle strength and walking ability in diplegic cerebral palsy: implications for assessment and management. Gait & Posture, 33(3), 321-325.

Tillin, N. A., Pain, M. T. G., & Folland, J. P. (2011). Short-term unilateral resistance training affects the agonist-antagonist but not the force-agonist activation relationship. Muscle & Nerve, 43(3), 375-384. doi: 10.1002/mus.21885

Unger, M., Faure, M., & Frieg, A. (2006). Strength training in adolescent learners with cerebral palsy: a randomized controlled trial. Clinical Rehabilitation, 20(6), 469-477.

Willoughby, K. L., Dodd, K. J., & Shields, N. (2009). A systematic review of the effectiveness of treadmill training for children with cerebral palsy. Disability And Rehabilitation, 31(24), 1971-1979.

Winters, T. F., Jr., Gage, J. R., & Hicks, R. (1987). Gait patterns in spastic hemiplegia in children and young adults. The Journal Of Bone And Joint Surgery. American Volume, 69(3), 437-441.

26