Engaging the Lower Extremity via Active Therapy Early (ELEVATE) Is Feasible and May Improve Gross Motor Function in Children with Spastic Bilateral Cerebral Palsy: A Case Series

Abstract

Purpose:

The feasibility of ELEVATE with respect to adherence and preliminary efficacy was determined for children with spastic bilateral cerebral palsy (CP) from encephalopathy of prematurity.

Methods:

A case series was used. Participants were randomized to receive ELEVATE immediately or delay the intervention by 3 months before receiving the intervention. The outcomes included feasibility measures of (1) number of children recruited, (2) percentage of sessions attended, (3) stride counts during the intervention, and preliminary efficacy measures of change over the intervention period in (4) Gross Motor Function Measure-66 (GMFM-66), and (5) kinematics and weight-bearing during treadmill walking.

Results:

Four boys under 3 years of age participated. All participants tolerated 60-minute intervention sessions four times/week for 12 weeks, and attended 75%–94% (min-max) of the targeted sessions. The median step count per session ranged from 833 to 2484 steps (min–max) during the final week of training. Participants showed an increase in GMFM-66 score of 2.4–7.5 points (min–max) over the 3-month intervention phase, as compared to a decrease of 1.7 for one participant and an increase of 1.3 for another over the delay period. Three participants demonstrated small improvements in their gait with the intervention.

Conclusions:

Engaging young children with bilateral CP in intensive rehabilitation targeting gross motor function was feasible and demonstrated preliminary efficacy. The results have guided the design of a larger clinical trial to assess efficacy of early, active interventions for children with spastic bilateral CP.

Cerebral palsy (CP) is the most common physical disability of childhood with an incidence of 1.5/1000 live births in high income countries.¹ The majority of children are classified as having spastic bilateral CP.^2,3 In some individuals with this subtype of CP, the lower extremities are much more affected than the upper extremities because of damage to white matter near the ventricles, a condition sometimes referred to as spastic diplegia. This presentation is characteristic of CP associated with preterm birth because this is a time of active proliferation of oligodendrocytes in the deep periventricular white mater, which is vulnerable to injury, especially as a result of changes in blood flow.^4,5 Periventricular white matter injury affects descending motor tracts close to the ventricles, especially the corticospinal tract (CST) to the lower extremities, and results in long-term impairments in gross motor function and walking.⁶

Early brain injury can lead to atypical brain development, especially within the first 2 years of life while the nervous system is still developing.^7,8 Animal models of early brain injury have demonstrated that early activity is essential for motor development, promoting neuroplasticity and improving function. For example, early, continuous environmental enrichment in a mouse model of early hypoxic brain injury promoted oligodendrocyte regeneration, myelination, and functional recovery.⁹ Early activity is also important for the development of the CST in other quadrupeds.¹⁰ Studies of kittens with early brain injury have shown that intensive motor rehabilitation while kittens are young improves the development of motor circuits and results in functional improvements, whereas training at an older age is less effective.¹¹

The efficacy of early, active interventions for children with CP is also becoming apparent. Interventions that incorporate motor learning principles and task specificity have demonstrated improvements in motor function.^12,13 Many of the studies in younger children have focused on the upper extremity,^14–16 but interventions are beginning to incorporate the lower extremity.^17–22 For example, incorporating lower extremity training into intensive therapy of the upper extremity for school-aged children with bilateral CP was shown to improve both fine and gross motor function and a systematic review of active exercises demonstrated improved gross motor function in ambulant school-aged children with CP.^23,24 Despite the increasing evidence for the efficacy of early, active interventions, there remains limited evidence for interventions focused on the lower extremity and gross motor function for young (i.e., preschool-age) children with CP.^25,26

We have demonstrated the efficacy of an early, intensive intervention targeting the lower extremity in very young children with unilateral CP as a result of perinatal stroke.²⁷ The current study was conducted to evaluate the feasibility of the ELEVATE intervention for preschool-aged children with bilateral CP, as a result of premature birth. We aimed to assess the number of eligible participants identified, the number of participants who enrolled in the study over 1 year and the adherence to the intervention. We also aimed to determine the tolerance of this population to the ELEVATE intervention, and to establish the preliminary effects of the intervention on gross motor function, activity, and gait.

Methods

Study design

A case series study was conducted at an academic center. Half of the participants were randomized to receive the intervention immediately and the other half underwent monthly assessments for 3 months before receiving the intervention, similar to a usual care, waitlist-control. All children were followed for monthly assessments for 3 months following the end of the intervention phase, and a final follow-up occurred within 3 months of the participants’ fourth birthday.

Participants

Inclusion criteria:

• 1.Age at entry to study between 8 months and 3 years old. The age range was chosen to include young children who could participate for at least the duration of the intervention, and to span and exceed the probable “critical period” for maturation of the CST, thought to occur before the age of 2 years.²⁸
• 2.Evidence of periventricular white matter injury on magnetic resonance imaging (MRI) or serial ultrasound (US) evidence of periventricular involvement.
• 3.Clinical evidence of bilateral spasticity affecting the lower extremities, as assessed using rapid passive movement of the ankle, knee, and hip joints by the screening physical therapist (PT).
• 4.Parent/guardian able to attend all assessment and intervention sessions.

Exclusion criteria:

• 1.Extensive brain injury beyond periventricular white matter injury.
• 2.Musculoskeletal, cognitive, or behavioral impairments that preclude participation in the protocol.
• 3.Epileptic seizures within the past 6 months that could interfere with the intervention. Only those with very frequent seizures were deemed too severe.
• 4.A diagnosis associated with neurological/developmental regression.
• 5.Botulinum toxin injection or surgery in the lower extremity in the past 6 months.

Clinical partners identified potentially suitable participants, who were then screened in person by a research PT. Evidence of periventricular white matter injury and absence of other extensive brain injury was confirmed by a pediatric neurologist. Final decision whether to include a participant was made by the study team. The PT who performed the in-person screening together with the principal investigator explained the study to the parent(s) and obtained consent prior to screening. Written informed consent was provided by parents of all participants.

Intervention

The ELEVATE intervention was delivered by a single research PT for 1 hour/day, 4 days/week for 12 weeks. The frequency and dosage of this intervention was modelled after the study we completed in children with perinatal stroke,^27,29 which resulted in a strong effect size. ELEVATE is intensive, child-initiated, play-based movement therapy of the lower extremities. The intervention occurred primarily over ground on various surfaces, including linoleum, carpet, mats, grass, and cement sidewalks. Children wore soft-soled shoes and no orthotics during the intervention to enhance the use of muscles in the feet and around the ankles. Activities included ascending and descending stairs and ramps, walking on stable and unstable surfaces, stepping over obstacles, balancing in standing, kicking, and squatting to pick up items. Manual support from the therapist was provided as needed to facilitate 60 minutes of child-initiated activity.

To increase exercise intensity and to augment movement error, weights were placed on the dorsum of the feet and the ankles of children with sufficient endurance to stay active for 50–60 minutes. Commercially available weights in increments of 110 g were used for the ankles, and one-quarter-inch chain links, ∼20 g each, were affixed to the dorsum of the feet with elasticized fabric. All participants had one lower extremity that was more affected than the other, so weights were applied to this side first. The weights initially increased gait asymmetry but the children walked more symmetrically as their strength and endurance increased. This is similar to enhancing asymmetry of walking using the split-belt treadmill in adult patients with stroke,³⁰ in that augmenting the asymmetry induces learning over time.³¹ Using weights to induce the asymmetry has the added benefit of strengthening the weak muscles. Children with unilateral CP are able to respond to asymmetric weighting of the limbs.³² To continue to challenge participants, foot and ankle weights were added to the less affected extremity and additional weights were added to the more affected extremity. Clinical decision-making was used to screen for potential overuse injuries and prevent injuries throughout the intervention.

There are many transitions related to walking, such as sit to stand, floor to stand, turning, stopping. While these tasks were not specifically targeted, they were incorporated throughout the intervention where possible. We feel that one of the most important elements in our intervention is motivation for moving, so the child's curiosity and enjoyment were emphasized. We allowed children to find transition that worked for them.

Outcomes

Recruitment

The number of children recruited in a 1 year period was compared with our experience with recruiting children with perinatal stroke. Recruitment rates similar to our previous experience was considered feasible.

Attendance

The total number of intervention sessions attended was documented, and was compared with our previous trial in children with perinatal stroke.

Active time during intervention

Active time was documented by a real-time observer for each intervention session. Bouts (measured in seconds) of walking, running, standing, and ascending and descending stairs were recorded as active time and any time spent sitting or being carried was recorded as rest time.

Stride counts during intervention

Stride counts were recorded using the StepWatch activity monitor (Orthocare Innovations, Seattle, WA) to quantify the amount of activity during intervention sessions. A StepWatch was applied to the ankle of the more affected lower extremity at the beginning of each session and was removed immediately following the hour of intervention. All participants had one lower extremity that was deemed to be more affected, typically via parental report and confirmed by the PT administering the ELEVATE intervention. No target level was set for these children, but our expectation was that by the final month of training, they would be able to take at least 1000 steps during the hour of training, similar to children with unilateral CP.²⁷

Safety

Safety of the ELEVATE intervention was assessed by documenting any harms or adverse events throughout the intervention period. This could have included, but was not limited to, injurious falls or overuse injuries in the participants.

GMFM

The Gross Motor Function Measure-66 (GMFM-66) was used to assess change in gross motor function. Sixty-six tasks spanning five dimensions were scored, including lying and rolling, sitting, crawling and kneeling, standing, and walking/running/jumping. Reliability, validity, and responsiveness to change have been established for children ≥6 months of age.³³ GMFM-66 assessments were performed by pediatric PTs who were blinded to the child's group assignment. A GMFM-66 assessor was assigned for each child and assessments occurred twice at baseline, separated by a week, and monthly thereafter (Figure 1A). The two baseline assessments were averaged for each participant, to provide a better representation of the child's motor capability.

Figure 1.

Participant flow chart.

The flow chart shows the number of participants screened, enrolled, allocated, and followed.

ELEVATE = Engaging the Lower Extremity Via Active Therapy Early.

Weight-bearing and kinematics of walking

Walking assessments were performed using a custom-built split-belt treadmill with a force plate under each belt. Children walked with soft-soled shoes, played with toys on a table in front of them and received manual support on their lateral thorax, as needed. Gait assessments were performed twice at baseline and monthly thereafter throughout the study.

Walking kinematics were captured with 3-D Investigator (NDI, Waterloo, Ontario, Canada). Markers were placed at the top of the iliac crest, greater trochanter, knee joint line, lateral malleolus, and head of the 5^th metatarsal, bilaterally. Customized MatLab script calculated percent weight-bearing (mean vertical force during stance phase) and knee and ankle joint angles throughout the gait cycle.

In each assessment, children walked for two trials of approximately 1 minute each at three different speeds, for a total of six trials per assessment. Speeds of 0.2, 0.4, and 0.6 m/s were used for children who were not walking independently and 0.4, 0.6, and 0.8 m/s were used for children who were. We focused on the speed that included at least 10 steps at all assessments throughout the study, typically the median speed of .4 m/s or .6 m/s.

The percent of a child's weight borne by the lower extremities in walking was calculated using the following formula:

$$\%W=\{[(V{F}_r+V{F}_l)/2]\times 100\}/BW$$ (1)

where %W is the percent of body weight the child supports on their feet, VF_r and VF_l are the average vertical forces during the stance phase from the right and left force plates, respectively, and BW is total body weight as measured on a scale. Independent walkers will approach or exceed 100% weight bearing.

Knee and ankle joint angles were calculated for the right and left side for each child at every time point. The average joint angle and standard deviation throughout the step cycle was calculated from bouts of at least 10 steps, with the stride cycle normalized in time from foot-contact to foot-contact represented by 100 data points. Full knee extension was characterized as 0° and positive degrees indicate knee flexion. A neutral ankle angle was represented as 0°, so measurements greater than 0° indicate ankle dorsiflexion and less than 0° indicates ankle plantar flexion (e.g., −5° indicates 5° of plantar flexion from a neutral ankle position).

Sample size

No sample size was established a priori since the objective was not to establish the efficacy of the intervention. Instead, the primary aim was to address the feasibility of ELEVATE for this population of children.

Randomization

Randomization was conducted via sealed opaque envelopes to determine if participants would receive the ELEVATE intervention immediately or after a 3-month delay. Each participant completed two baseline assessments 1 week apart before their parent chose an envelope.

Analytical methods

Descriptive statistics (median, minimum, and maximum) for primary and secondary outcome measures were performed using Microsoft Excel and SPSS 20.0 software. Results are summarized in box-plots.

Results

Recruitment and participant flow

Recruitment began in January 2017 and ended in December 2017. The final 3 months follow-up assessments occurred in January 2019. The final 4-year-old follow-up assessment was completed in January 2020.

Figure 1 shows the flow of participants throughout the study. Within the designated recruitment year, five potential participants were referred to the study by clinical partners. One family did not reply to contact from the research team and the remaining four children enrolled in the study. Based on the randomization, two children began the intervention immediately and two delayed the onset of the intervention by 3 months. All participants received the intended intervention. Two participants (P3 and P4) missed one assessment during the post-intervention 3-month follow-up period due to illness and scheduling difficulties. Three participants completed the 4-year-old follow-up. One participant (P3) did not complete the 4-year-old follow-up due to scheduling difficulties.

Participant characteristics

Table 1 shows the baseline demographic and clinical characteristics for all participants. All participants were male and all were born prematurely (gestational age at birth ranging from 25 to 30 weeks). Their age at entry to the study was corrected for gestational age if they were under 2 years old.

Table 1.

Participant Characteristics

Code	GA at birth (wks)	Age at BL (mo)	Sex	GMFCS level	Group	GMFM-66 begin Rx (percentile)	GMFM-66 end Rx
P1	29	25	Male	II	Immed.	49.9 (50)	52.3
P2	30	35	Male	II	Delay	50.2 (50)	55.3
P3	25	15*	Male	I	Delay	58.8 (NA)	66.3
P4	27	32	Male	II	Immed.	51.6 (50-75)	58.8

Asterisk (*) indicates adjusted age.

BL = baseline; Code = participant code; GA = gestational age; GMFCS = Gross Motor Function Classification scale; Group = participant randomized to Immed. (Immediate training) or Delay (Delayed training); GMFM-66 begin Rx (percentile) = GMFM-66 score at the beginning of training (percentile score for corresponding GMFCS level and age); GMFM-66 end Rx = GMFM-66 score at the end of the 12 week intervention; NA = not available because percentiles scores unknown under the age of 2 years old.

Participant 1 (P1)

This participant began the ELEVATE intervention immediately following baseline assessments and was not an independent walker. He tolerated a small weight on the dorsum of each foot beginning in the second week of training and weight at the ankles bilaterally beginning in the seventh week of training. P1 showed the smallest change in their GMFM-66 score over the intervention period, which may be related to inconsistent engagement in the intervention. This participant became frustrated with the intervention and was often unwilling to walk with the training PT for prolonged periods and long distances, also reflected in the smallest number of strides per session among the participants (see below).

Participant 2 (P2)

During the intervention, we began using weights on the dorsum of the feet in the second week of training and consistently used ankle weights beginning in the ninth week of training. P2 began the intervention phase with impaired trunk control, difficulty clearing the toes while walking and very limited endurance. With the intervention, he was able to support more of his weight throughout the 60-minute sessions and began to stand independently for over 30 seconds, walk with one hand support for balance, and take short bouts of steps independently.

Participant 3 (P3)

P3 showed limited improvement on the GMFM-66 over the delay period, but substantial improvement over the intervention phase. P3 was an independent walker, tolerated weights on the dorsum of the feet beginning in the first week of training and bilateral ankle weights beginning in the second week of training. This participant had difficulty walking on unstable surfaces and over obstacles at the beginning of the intervention phase but was able to make quick changes in direction, step over obstacles and up shallow stairs without support by the end of the intervention. This participant was keen to participate in the intervention and responded well to progressive challenges.

Participant 4 (P4)

P4 was not walking independently, but consistently keen to participate and was able to keep active for a prolonged time during the intervention sessions, even with weights on the dorsum of the feet. He showed substantial improvement in his GMFM-66 score over the intervention period.

Attendance and active time during intervention

The number of ELEVATE intervention sessions completed by each participant was 42 (P1), 45 (P2), 36 (P3), and 39 (P4). All participants were able to tolerate the intervention activity, albeit inconsistently, with a median ranging from 50 to 56 minutes per session in the final month of the intervention. The average number of active minutes during each month of the intervention for each participant is shown in Figure 2A. The average percent of training time spent in over ground walking, treadmill walking, standing, and other activities (e.g., jumping) for each participant is shown in Figure 2B.

Figure 2.

Intervention details.

(A) Active time (minutes/session) for each participant for each month of the intervention. Asterisk (*) indicates outlier. (B) Average percent of intervention time (over the 12 week intervention) spent in over ground walking, treadmill walking, standing, and other activities for each participant.

Strides during intervention

Stride counts were used as a measure of the degree of mobility during intervention sessions. Strides in a single session ranged from 552 to 2,974 among all participants. The average number of strides during the first, second, and third month of the intervention for each participant is shown in Figure 3. The median strides per session in the final week of the intervention ranged from 833 to 2,484.

Figure 3.

Strides during intervention sessions. Strides during intervention sessions for each participant in each month of the intervention.

Asterisk (*) indicates outlier.

Gross motor function

The monthly GMFM-66 scores for each participant are shown in Figure 4. Change scores were calculated for each child over the ELEVATE intervention and ranged from an increase of 2.5 to 7.5 over the 3-month intervention. The change in GMFM-66 score over the 3-month delay period was a decrease of 1.7 for P2 and an increase of 1.3 for P3. The GMFM-66 scores for the two participants classified as GMFCS level II (P2 and P4) who attended the 4-year-old follow-up were 58.9 and 64.6, which corresponds to the 75^th and between the 90^th and 95^th percentile for 4-year-old children, respectively, as reported in reference curves.³⁴ The GMFM-66 score at the 4-year-old follow-up for the participant classified as GMFCS level I was 76.0. This corresponds to the 75^th percentile for 4-year-old children classified as GMFCS level I.³⁴

Figure 4.

GMFM-66 results.

The total GMFM-66 score are indicated for each participant for each assessment. Vertical dashed lines segment the 3-month time periods in the study. P1 and P4 received the intervention immediately on entry to the study, and P2 and P3 received the intervention after a three-month delay period.

GMFM-66 = Gross Motor Function Measure-66.

Weight-bearing and kinematics of walking

Weight-bearing during walking was expressed as a percentage of the child's body weight, where increases in weight-bearing indicate more proficient walkers. Throughout the intervention period, weight-bearing (expressed as % body weight) increased by 3.8% for P1, 11.0% for P3 and 16.1% for P4 and decreased by 0.2% for P2 (Figure 5). The increase over the 3-month delay period was 13.1% for P2 and 7.2% for P3.

Figure 5.

Weight-bearing while walking.

The percent of their body weight that each child could independently support prior to and following the intervention, as measured during treadmill walking.

The right and left knee and ankle joint angles in the sagittal plane throughout the gait cycle were used to assess changes in the walking pattern and range of motion with training (Figure S1). There were very few changes with training, especially for P4. P1 demonstrated less variability in the range of motion at the ankle between steps and slightly greater range of motion in both dorsiflexion and plantarflexion. P3 adopted a slightly longer proportion of time in the stance phase on the right following the intervention, which improved the knee and ankle joint angle symmetry between sides.

Adverse events

No harms or adverse events occurred during the ELEVATE intervention.

Discussion

This case series demonstrates the feasibility to recruit and engage young children with spastic bilateral CP in intensive rehabilitation targeting the lower extremities. All participants were able to participate in 1-hour intervention sessions four times per week and tolerated challenges including adding weights to the ankles and feet, reducing the external support required for walking (e.g., from two hands held to one or slight support at the shoulder or pelvis) or varying the environment (e.g., introducing stairs, unstable surfaces, etc.), as applicable. All participants showed increases in GMFM-66 score over the intervention phase, ranging from 2.4 to 7.5. The minimum clinically important difference (MCID) for the GMFM-66 is only known for children older than 4 years old. Medium and large change is estimated to be 1.7 and 2.7 for GMFCS I, and 1.0 and 1.5 for GMFCS II, respectively.³⁵ Three participants also demonstrated increased weight-bearing in the treadmill walking assessments.

Feasibility considerations

The referral rate was comparable to our prior experience recruiting children with perinatal stroke. Here, five individuals were referred over the 1-year period, which is slightly lower than the recruitment rate seen during our previous trial for children with perinatal stroke (mean of seven children per city per year²⁷). Our recruitment is dependent on practicing clinicians referring, and special interest groups distributing study information. Given that this is a new population for our team, we feel that achieving approximately six to seven children per year should be possible in the future. We are actively engaging new partners for recruitment in our new and on-going RCT for these children. The age at referral was generally later than the children with perinatal stroke, and this could be a missed opportunity (see Section: Early Intervention below).

The children attended 36–45 of the 48 intended sessions, which is within 1 SD of the number of sessions attended by the children with unilateral CP in our previous trial of the same length (Mean ± SD = 41 ± 6 sessions). This dosage of intervention produced strong results in the previous trial,²⁷ and was considered sufficient to determine the efficacy of the intervention for these children.

Intervention considerations

The ELEVATE intervention is designed to progressively challenge each participant, which can include increasing stride counts, challenging standing and walking balance, introducing foot and ankle weights, and incorporating skills such as stairs, rapid direction changes, and avoiding obstacles. We did not have an a priori expectation for increases in stride counts/session, but our prior experience in children with unilateral CP suggested average increases over 3 months to be ∼600 strides/session for independent walkers, and ∼1000 strides/session for non-independent walkers.²⁷ The type of challenge in the training depended on the child's gross motor function at the time. For example, P1 and P2 increased their active time per session in the second month of the intervention, but the average number of stride counts remained the same or dropped slightly during this period. In this case, the participants were not as mobile during their second month of training, but the intervention was more challenging than during the first month of the intervention (e.g., walking along extremely unstable surfaces or ascending and descending stairs as compared to walking down a hallway). Importantly, the use of stride counts does not capture other forms of training intensity such as challenges to balance, which may be especially important in children not yet very mobile. These children were slightly more limited in function compared to our previous trial. Better ways to quantify challenge during training are needed, since our method to time categories of activity is very labour intensive, and likely unfeasible in the future.

Despite the difficulties in quantifying training intensity, the GMFM-66 scores increased with training reflecting an improvement in gross motor function. Notably, P3 and P4 increased the average number of strides/session throughout the intervention phase and saw large increases in GMFM-66 scores and percent body weight supported, potentially suggesting the importance of quantity of movement.

The personality of the child also influences how much challenge they are willing to embrace during the intervention. For example, P1 was considerably more difficult to engage compared to the other three participants (see Results). We have had similar experiences with children with unilateral CP, in which some children were less willing to try new and challenging tasks. The creativity of the therapist and their relationship with the child is especially critical in those situations. Thus, the effectiveness of ELEVATE will depend on these and other child-specific and therapist-specific factors.

The gross motor function of the children in this study was mid- to high-functioning in their respective GMFCS levels. P1, P2, and P4 fell between the 50th and 75th percentile for children with GMFCS Level II, while P3 could not be classified because of his younger age at the start of training (i.e., there are no distributions for children <2 years old). Nevertheless, even for a 2-year old, P3's baseline GMFM-66 score would be close to the 50th percentile for GMFCS Level I. In general, children with encephalopathy of prematurity and evidence of periventricular white matter injury can vary greatly in their motor function, ranging between GMFCS Levels I to IV.³⁶ Thus, our small sample consisted of children with higher motor function in this population. It will be important in the future to determine the type of child most suited for this form of training within this diverse population.³⁷ It is possible that children with very low or very high levels of function will benefit less. In future trials, it would be valuable to include stratification of children according to baseline function, to ensure equal distribution of abilities between groups.

Aside from P3, who was classified as GMFCS level 1, the children in this study demonstrated lower endurance during the hour-long intervention sessions, as compared to children with unilateral CP. This is consistent with a recent systematic review that found that increases in energy expenditure and oxygen cost are correlated with GMFCS level.³⁸ The greater fatigue in these children may be related to limited experiences with standing and walking with support as well as the compensatory movements employed to maintain balance and mobility.

The gross motor function of children with unilateral CP who received the ELEVATE intervention continued to improve for the 3-month follow-up after the intervention.²⁷ In contrast, the children with bilateral CP in this study appear to maintain but not improve their GMFM-66 scores throughout the three-month follow-up (Figure 4). In contrast to children with unilateral CP, classified as GMFCS I, most of the participants in this study were not independently ambulating at the end of the training period, and were likely not able to make significant progress in their gross motor development through self-training at home upon completion of the intervention. It will be important to evaluate the impact of multiple bouts of intensive rehabilitation for children with bilateral CP in future studies, which may provide more opportunity for improvement.

Early intervention

Three out of four participants in this study were over the age of 2, which is consistent with referral practices in Canada for these children with bilateral CP, in which the average age of diagnosis was 24.1 months.³⁹ Despite the evidence for early intervention to promote motor function,^13,40 delayed diagnosis and referral highlights the difficulty of initiating intensive rehabilitation when these children are young and presents a challenge for studying the effect of early administration of ELEVATE.⁴⁰ An accurate diagnosis can be made within the first year of life using magnetic resonance imagining, and the Hammersmith, General Movement Assessment or the Developmental Assessment of Young Children, which will create more opportunities to implement and evaluate early interventions.^41,42 This is promising, since early interventions that include task-specific motor training such as GAME¹⁸ and Small Step⁴³ have demonstrated improvements in motor function in children with bilateral CP. A recent clinical practice guideline also strongly recommends early intervention and task-specific motor training that incorporates challenging motor tasks with repeated practice.¹³

Our study has several limitations. Spastic bilateral CP represents a diverse group of people and our small sample did not reflect this heterogeneity. Although we would not expect a sex or gender difference in the efficacy of the ELEVATE intervention at this age, all the children in this pilot study were male, by chance. Also, three of the four participants were all close to 3 years old and classified as GMFCS level II. The ability to adapt to varied abilities is a strength of the ELEVATE intervention because it is child-centered and designed to progressively challenge each child at their functional level. It will be important to include younger children, boys and girls, and children with varied functional presentations in future studies. For example, children with foot or ankle deformities that may require bracing, and how ELEVATE would accommodate those children.

Another potential limitation of this small pilot is that only one research PT at one academic center provided the intervention. To increase generalizability, future efficacy studies should incorporate multiple therapists over multiple training sites.

The time commitment of four times per week for 12 weeks is substantial, and likely not feasible for many families, especially since health insurance plans typically do not cover this frequency and duration. Thus, in children with unilateral CP, we are testing a slightly modified delivery model of ELEVATE, which uses a parent-therapist partnership to jointly deliver the intervention (i.e., 2 days each).⁴⁴ In a qualitative study in parallel with this on-going trial, we are determining the parents’ perspective on this partnership model.⁴⁵

Conclusion

The current study demonstrates that it is feasible to recruit and carry out ELEVATE in a small number of children with bilateral CP with relatively high function (GMFCS Level I or II). We have further demonstrated preliminary efficacy in these four children showing improved gross motor function over the 12 weeks intervention. This information has contributed to the larger RCT, which is currently underway (clinicaltrial.gov, NCT03672877).

Key Messages

What is already known on this topic

• •Early, active interventions have been shown to improve motor function of the upper extremity for children with cerebral palsy (CP), but there remains limited evidence for interventions focused on the lower extremity and gross motor function for young (i.e., preschool-age) children.
• •The ELEVATE intervention has shown preliminary efficacy for children with unilateral CP but feasibility and preliminary efficacy has not been evaluated in children with spastic bilateral CP.

What this study adds

• •We have demonstrated that it is feasible to engage young children with spastic bilateral CP in intensive rehabilitation targeting the lower extremities and gross motor function.
• •The ELEVATE intervention shows preliminary efficacy in terms of improved gross motor function and gait for these children.

Jaynie Yang, is a professor in the Department of Physical Therapy, University of Alberta. She is a physical therapist with doctoral and post-doctoral training in the biomechanics and neuroscience of human walking. In the last 30+ years of work at the University of Alberta, her research has focused on how the nervous system controls walking in people, and ways to retrain walking in individuals with neurological insults. She is best known for her work in: a) using the stepping and crawling behaviour of infants to understand the neural control of human walking, and b) applying knowledge from preclinical studies to improve the rehabilitation of walking in clinical populations, including adults with spinal cord injury and children with perinatal brain injuries.