Treadmill Training in Pediatrics
A Literature Review
Audrey Kaplonski, PT
University of Medicine and Dentistry of New Jersey
School of Health Related Professions
In partial fulfillment of the requirements for IDST 6400
Craig Scanlan, EdD, RRT
December 9, 2002
Physical therapists who work in pediatrics know that every parent’s ultimate goal for their child is ambulation. Even in those children with little active movement, parents usually hold out hope that their child will walk. Over the years pediatric physical therapists have used different techniques to help the children in their care reach as much functional independence as possible. The use, however of treadmill training with body weight support is a new area in which pediatric physical therapists (PT) are beginning to venture to look for even more answers.
There have been numerous studies on adults with hemiplegia and spinal cord injuries on the effectiveness of body weight supported treadmill training, (BWSTT). Adult studies have hypothesized that partial BWSTT works in the adult population due to the presence of a central pattern generator. A central pattern generator, in response to a stimulus, causes neuronal linkages in the spinal cord to generate the rhythmical motor activity necessary for gait (Mudge & Rochester, 2001). However, in the pediatric population, ambulation is a not a skill to be relearned, but rather a skill to be learned as new. P. Zelazo, N. Zelazo, and Kolb (1972) showed that infants who were stimulated to use their walking and placing reflexes during the first eight weeks of life, walked sooner than infants who did not receive this stimulation. Therefore, one can hypothesize that in order to achieve some level of ambulation, children need to practice walking. This is a departure from some of the more traditional methods of treating children, where the importance of developing control of movements free from reflexive patterns is emphasized, rather than a task specific approach.
The purpose of this review is to examine the literature relating to the use of treadmill in training in pediatrics and to determine what affect it has on the development of ambulation in the pediatric population.
An OVID search was done using both the CINAHL and MEDLINE databases. CINAHL was searched from 1982 until October of 2002 and MEDLINE was searched from 1966 to October 2002. These dates encompass all the available citations in the respective databases. The terms that were mapped and combined included: body weight support, treadmill training, gait analysis, ambulation, locomotion, brain damage, cerebral palsy, and Down syndrome (See Appendix). Although this review emphasized the pediatric population, the words children and pediatrics were not mapped since citations were limited enough without using these terms. Studies were included in this review if they met the criteria of studying the pediatric population (operationally defined as under the age of 21) and included treadmill training with some type of support.
Reference lists of chosen studies were also examined. Articles that appeared to meet the criteria for inclusion were obtained and reviewed.
The Internet was searched for pertinent studies, using the search engines, Google, Alta Vista, Northern Light and Infoseek. This proved to be a poor source of studies, with most sites only offering general information or duplication of studies obtained through Ovid searches. No studies were added to the review from this source.
Seven articles were chosen to be included in this review. Greater weight was given to randomized clinical trials, followed by cohort studies and case control studies. Due to the lack of information on the use of BWSTT in the pediatric CP population, one case report was also included to provide additional information on this topic.
Studies of BWSTT in the adult population have centered on the use of this technique on adults with hemiplegia and spinal cord injuries. However, the transference of the technique of BWSTT to the pediatric population requires an understanding of infant development and of the differences in CNS injuries in the pediatric and adult population. It has been shown (Brouwer, Ashby, 1991; Gottlieb, Myklebust, Penn, & Agarwal, 1982; Mykelbust, Gottlieb, Penn, & Agarwal, 1982) that patients with cerebral palsy tend to demonstrate reciprocal excitation rather that reciprocal inhibition during the performance of rapid ankle dorsiflexion. The latency period in which this excitation occurs suggests that in this population, damage to the brain creates a secondary, developmental disorder in the spinal cord. However, the locomotor activities of newborn children have been well documented, and these rhythmical activities occur even in anencephalic infants (Mudge & Rochester, 2001). Therefore, the idea of activation of an existing spinal or supraspinal generator is feasible when working with this population
Unlike in adults, the pediatric population is working with an immature nervous system. Development of motor skills requires the production of neuronal projections. There is indication, given the stepping patterns of anencephalic infants, that the neuronal mechanisms which produce primitive locomotion are laid down during development and that these mechanisms are located at or below the brainstem (Forssberg, 1985). If this is indeed the case, then the practice of ambulation may help develop those neuronal connections in cases where there is damage in the brain prior to or shortly after birth, rather than as in adults, where the existing pathways are disrupted.
Finally, the study of BWSTT needs to address the issues of improved strength and endurance, variables that may also improve skill development in the pediatric population.
This study had difficulty, however, with internal validity, as there was no accounting for the natural maturation that could occur in subjects this young. This maturation could have had a positive influence on the GMFM scores.
As this was a feasibility study, statistics were descriptive in nature, and no causality can be drawn. However, this study demonstrates the need for sensitive outcome measures, provides guidance in planning for future efficacy studies, and shows a positive impact from the use of BWSTT.
Six nonambulatory children and four children with CP who had some degree of ambulation received BWSTT as well as traditional PT for three months in a study performed by Schindl, Forstner, Kern, and Hesse (2000). This study, with some modifications, built upon the Richards (1997) feasibility study. The children ranged in age from six to 18 years. Once again the degree of support provided by the harness was based on the child’s needs. Outcome measurements included the Functional Ambulation Categories (FAC) and the standing and walking components of the GMFM. The patients were asked for their subjective input regarding the use of BWSTT and the majority found the training motivating and joyful. Two subjects in the nonambulatory category found the training exhausting. Eight of the subjects found improvement in their motor abilities post training, such as transfers and sit to stand maneuvers. One subject, who was in the nonambulatory group prior to the BWSTT, was able to give up his wheelchair usage at home after the three months of training. One subject in the ambulatory group was able to climb a flight of stairs, a previously unachieved goal.
Baseline scores on the FAC and GMFM were static for six weeks prior to initiation of the BWSTT. Mean scores on the FAC increased from 1.1 to 1.9 with a p=. 02. On the standing portion of the GMFM the mean increased from 10.9 to 15.9 with a p=. 018, and the walking score improved from 9.8 to 14.1 with a p= .007. The improvement in the standing score demonstrates a 47% change, while that of the walking score shows a 50% change.
This study looks at older children with CP as the children ranged in age from six to eighteen years. At this age, children who have not achieved independent ambulation are frequently thought not to have the potential to be ambulators. Yet, in this small, baseline study, the majority of the children have not only shown some improvement in their ambulation skills, but also in other functions that would improve their ability to care for themselves.
In a case report by McNevin, Coraci, and Schafer (2000) different outcome variables were used to assess the efficacy of partial unweighting in changing the gait of a seventeen-year-old adolescent with CP. The subject was a community ambulator with forearm crutches. The authors looked at the effects of partial weight support versus full weight on the dependent variables of heart rate, blood pressure, and perceived exertion. They were able to show that partial weight support allowed the subject to have lower HR and BP readings that were consistent with her perceived exertion rating. They also felt that since the patient was able to walk faster in a partially unweighted condition, the amount of force required to propel the patient forward was contributing to her gait inefficiency, rather than the timing of force production.
Although this case report adds little to our knowledge of how BWSTT influences the development of ambulation, it is useful in that it shows that the use of BWSTT can help in improving gait efficiency, as demonstrated by its influence on heart rate and blood pressure.
Thelen and Ulrich (1991) studied the use of treadmill training in the normal developing infant, to determine when alternating stepping movements which are similar to upright bipedal locomotion occur. They initially identified thirteen infants, seven girls and six boys, although four did not complete the study. Starting at one month of age, the infants were tested twice monthly on a treadmill with two individual parallel belts. All the infants who remained in the study completed it at least through month seven. An adult, allowing the infant to take as much weight on their feet as they could tolerate, supported the infants on the treadmill. The infants all had 11 trials on the treadmill in the following order: Trials 1 and 9 were baseline trials with the belts remaining stationary. Trials 2-8 had the belts slowly moving in increasing speed. Trials 10 and 11 were mixed speed trials with the belts moving at different speeds. A motion analysis system was used to monitor the foot movements.
The authors identified alternating steps as those steps that were initiated within 20-80 % of the step cycle of the opposite leg. Generally, the infants displayed a trend in developing the alternating steps: initially there were only a few, followed by a short decline in the number of steps, followed by a steady increase lasting several months. These changes were not only noted in the trials with the belts moving at the same speed, but also in those trials where the belts moved at different speeds.
The idea that infants can walk on a treadmill with support, with a stepping behavior never seen without the treadmill, lends credibility to the idea that early ambulation is constrained by strength and balance factors. If the infants did not have the support or the use of the treadmill, they would have to maintain their weight fully on one leg and maintain their balance while swinging the other leg forward. This activity requires considerable extensor strength and skill, which is not seen in the infant’s early development. Other factors, such as motivation, voluntary control of the legs, and the use of visual information to guide exploration appear to develop earlier, and allow the infant the potential to use a treadmill with support to walk.
Building upon the study of Thelen and Ulrich (1991), B. Ulrich, D. Ulrich, and Collier (1992) looked at the development of stepping on a treadmill with seven 11-month old children with Down syndrome. Following Thelen and Ulrich’s (1991) protocol, the infants were supported by an experimenter and were given eight trials on the treadmill. Trials one and eight were baseline, with no movement of the treadmill. Trials two through seven consisted of two trials at three different belt speeds. The order of the trials was randomized.
One of the infants demonstrated no steps during the test, and the data from this infant was subsequently excluded from the study. The remaining six infants all demonstrated alternating steps when supported on a moving treadmill. A MANOVA was performed to determine whether the number of alternating steps taken varied with belt speed. There was significant difference in the number of alternating steps taken in the stationary belt trials and the moving belt trials. However, there was no difference in the number of alternating steps taken at the three moving speeds. Therefore, the infants only required the stimulus of the moving treadmill, with no one speed being better than the others.
The results of this study show that the use of BWSTT in infants with Down syndrome allow the infants to take alternating steps long before they can walk independently. This demonstrates that at least one major component of ambulation is in place and available for use earlier in life.
B. Ulrich, D.Ulrich, Collier, and Cole (1995) continued the work on BWSTT in infants with Down syndrome. Nine infants with Down syndrome (DS) were enrolled in the study (two subsequently dropped out), and followed longitudinally for a period varying from four to 21 months. Age at the beginning of the study ranged from 8 to 11 months. Following the same protocol as B. Ulrich, D. Ulrich, and Collier (1992) the infants were followed monthly while engaging in treadmill training. In addition to the analysis of stepping, anthropometric measurements were also taken. The results demonstrated that infants with DS were able to perform alternating stepping patterns when supported on a motorized treadmill long before they could walk independently. It was found that at the time when alternating stepping became the infant’s dominant step type, they also had a reduction in thigh and calf skin folds with a rise in body weight and had developed the ability to make forward progress in prone and to pull to stand independently. The authors hypothesize that this set of variables reflects increasing lower extremity and pelvic strength and control. This may help suggest when BWSTT may be utilized in the developmental sequence.
In one final study, the effect of BWSTT on the development of walking in children with DS was looked at in a randomized clinical trial (Ulrich, D., Ulrich, B., Angulo-Kinzler, Yun, 2001). Thirty infants with DS were randomized to either receive treadmill training and conventional PT or to receive conventional PT only. The parents were instructed in how to perform the treadmill training in their home. The effectiveness of the intervention was measured by the length of time the child needed between learning to sit independently (criteria for entry into the trial) and transitioning to stand, walking with assistance and walking independently. The study showed that the group receiving treadmill training learned to walk with help (p= .03 with an effect size of .80) and independently significantly faster (p =. 02 with an effect size of .83) than the control group. The intervention group demonstrated independent walking 101 days sooner than the control group. By using power analysis prior to the initiation of the study, the authors were able to include enough subjects to demonstrate a group difference.
This development of early ambulation in the DS in a randomized clinical trial shows that the use of BWSTT is effective. As it has been shown that ambulation and exploration of the environment has a positive impact on cognitive and social development, the implications of this treatment on children with DS is important. Additionally, further investigation needs to be done on the impact of this early ambulation on the quality of gait as well as cardiovascular conditioning.
Results of the studies in this review have shown some positive effects from the use of BWSTT in the pediatric population. The limited subjects in the studies on children with CP make it difficult, however to draw any type of definitive conclusions. Although the studies showed statistically significant findings, particularly in the standing and walking sections of the GMFM (Richards et al. 1997, Schindl et al. 2000), there was, however, no attempt made to randomize the treatment to determine if the positive effects were from the use of the BWSTT, natural maturation, or the continued effect of conventional physical therapy. The argument can be made, however, particularly in the Schindl study that looked at an older population, that the lack of progress prior to the initiation of the study provided a stable baseline from which to measure gains.
Dietz and Berger (1983) have shown that the coactivation pattern seen in children with CP is very similar to the leg muscle activity described in stepping newborns. Therefore, it can be assumed that in these children the locomotor pattern cannot mature because of impaired supraspinal influences. However, if the impact of BWSTT is shown to have a greater affect on neuronal connections in the spinal cord, it can then be said that gait improvements may be effective using this technique.
The studies on children with Down syndrome and the one study on normal infants provide an interesting counterpoint to the CP studies. Even though these studies worked with small subject groups, they were able to build one study upon another and show that the use of BWSTT in infants not only demonstrated increased alternating stepping, but also in the final RCT (Ulrich, D. et al. 2001) showed a significant increase in the age when independent walking was achieved.
The question still remaining to be answered however is what is to be achieved through the use of BWSTT. The use of BWSTT permits the early initiation of partially supported stepping without requiring good postural stability and full weight bearing (Dobkin, 1999). The use of treadmill training therefore, is a viable tool in the PT’s arsenal, for those children who are unable to achieve ambulation any other way. The studies presented in this review, however leave several questions unanswered. The long-term impact of treadmill training has not been researched, and there have been no studies that have attempted to see if children were able to maintain the gains that they made. Additionally, there is no information in the pediatric literature that shows whether the use of BWSTT translates into overground walking. Van Ingen Schenau (1980) showed, through the use of Galilean invariance, that except for air resistance, there should be no mechanical difference between the two concepts. As the studies have shown, a small percentage of the children with CP actually became more proficient ambulators, although gains have also been shown in other areas.
This review of BWSTT in pediatrics looked at one randomized clinical trial, two quasi-experimental studies, three non-experimental studies and one case report, all prospective in nature. These studies encompassed fifteen children with CP, forty-six children with DS and thirteen children without a disability. Although various outcome measurements were used, the primary focus of the literature review was on the improvement of gait or stepping patterns. The quasi-experimental studies of children with CP placed their emphasis on improvement in the GMFM, while the non-experimental studies on children with DS and children without a disability looked more closely at the improvement of stepping patterns. The randomized clinical trial on children with DS used the Bayley Scale of Infant Development as an outcome measure, while the case report used perceived exertion as the outcome measure.
No matter what outcome measure was used, there was improvement shown with the use of BWSTT. Statistically significant changes were seen in the all of the experimental studies. Although no causal relationship can be assumed, in the non-experimental studies there was a positive change in the amount of alternating steps taken with the use of BWSTT. Additionally, in the case study, the subject demonstrated a decrease in heart rate and blood pressure with the use of BWSTT. Therefore, although most of the samples studied were small, the likelihood of a positive outcome appears strong and a recommendation can be made for use of this treatment in the pediatric population.
The purpose of this literature review was to discuss the implications of treadmill training in the pediatric population. Despite the lack of articles pertaining to the use of BWSTT in both children with CP and DS, the evidence that is there points to positive effects from the use of this technique. Schindl et al (2000) and Richards et al (1997) both demonstrated improvements in the GMFM in children with CP, a tool frequently used by pediatric PT’s to judge a child’s progress. The possibility of achieving improved ambulation or improved functional skills in a population where even small gains can improve the quality of life leads one to conclude that the use of BWSTT can be an important tool in a therapist’s repertoire. The lack of an RCT in this sample, however, makes one hesitant to use this as the only technique to improve function. Until further studies are completed, therapists should continue to utilize those techniques that have worked in the past as well as add the use of BWSTT.
The question remains however, on how to best implement BWSTT in the day to day school or clinic setting. The equipment is large and expensive and there is frequently not enough time to provide a child with both conventional and BWSTT. The choice is difficult to make, when no study has directly compared the outcomes of BWSTT to the outcomes from conventional therapy alone. In the studies presented above, most included conventional therapy along with the BWSTT, making it difficult to determine exactly what caused the improvement in the outcome measures.
As therapists, however, we are able to use our clinical judgment to determine when a child may have the potential to ambulate or to become more functional in their gross motor tasks. The need for a therapist's input, both prior to initiation of the BWSTT and during the training is essential. In both Schindl et al (2000) and Richards et al (2001) therapists provided the BWSTT for the children, stressing proper alignment and muscle activation with every step that was taken. In the studies of children with DS, parents or researchers frequently provided the BWSTT. With these children, however, the issues of tonal changes and spasticity are not so prevalent.
The choice between BWSTT and conventional therapy will frequently need to be made in the day to day practice of pediatric physical therapy. Those children who are older and are no longer showing substantial progress with conventional therapy should be using BWSTT in an attempt to progress their motor skills. In the younger, pre-ambulatory population, BWSTT will make an excellent adjunct to conventional therapy. Until it can be shown that BWSTT translates to overground walking, or until larger RCT’s are done that show improved outcome measures with BWSTT without conventional therapy, both techniques need to integrated to provide children with the opportunity to reach their full potential.
This review has demonstrated the importance of treadmill training in the development of gait in the pediatric population. However, it also demonstrated the lack of research in this area. Despite the thoroughness of the literature search, only three articles pertaining to its use in the CP population and three in the DS population were found. As these small, for the most part non-randomized, studies have demonstrated, however, there is the potential for this to become an important treatment technique. The necessity of increased research in this area cannot be stressed enough. As clinicians, we should all be striving to achieve the best outcome for our patients. As physical therapists, we need to perform the research to help us reach these outcomes and provide evidence-based practice.
Brouwer, B., Ashby, P. (1991, June). Altered corticospinal projections to lower limb motoneurons in subjects with cerebral palsy. Brain, 114, 1395-407.
Dietz, V., Berger, W. (March 1983). Normal and impaired regulation of muscle stiffness in gait: A new hypothesis about muscle hypertonia. Experimental Neurology, 79, 680-7.
Dobkin, B.H. (1999). An overview of treadmill locomotor training with partial body weight support: A neurophysiologically sound approach whose time has come for randomized clinical trials. Neurorehabilitation and Neural Repair, 13, 157-165.
Forssberg, H. (1985). Ontogeny of human locomotor control: 1. Infant stepping, supported locomotion, and transition to independent locomotion. Experimental Brain Research, 57, 480-493.
Gottlieb, G.L., Myklebust, B.M., Penn, R.D., Agarwal, G.C. (1982). Reciprocal excitation of muscle antagonists by the primary afferent pathway. Experimental Brain Research, 46, 454-6.
McNevin, N.H., Coraci, L., Schafer, J. (2000 April). Gait in the adolescent cerebral palsy: The effect of partial unweighting. Archives of Physical Medicine and Rehabilitation, 81, 525-528.
Mudge, S., Rochester L. (2001, July). Neurophysiological rationale of treadmill training: Evaluating evidence for practice. New Zealand Journal of Physiotherapy, 29, 6-15.
Myklebust, B.M., Gottlieb, G.L., Penn, R.D., Agarwal, G.C. (1982 October). Reciprocal excitation of antagonistic muscles as a differentiating feature in spasticity. Annals of Neurology, 12, 367-74.
Richards, C.L., Malouin, F., Dumas, F., Marcoux, S., Lepage, C., Menier, C. (1997). Early and intensive treadmill locomotor training for young children with cerebral palsy: A feasibility study. Pediatric Physical Therapy, 9, 158-165.
Schindl, M.R., Forstner, C., Kern, H., Hesse, S. (2000 March). Treadmill training with partial body weight support in nonambulatory patients with cerebral palsy. Archives of Physical Medicine and Rehabilitation, 81, 301-306.
Thelen, E., Ulrich, B.D. (1991). Hidden skills: A dynamic systems analysis of treadmill stepping during the first year. Monographs of the Society for Research in Child Development, 56(1, Serial No.223).
Ulrich, D.A., Ulrich B.D., Angulo-Kinzler, R.M., Yun, J. (2001, November). Treadmill training of infants with Down syndrome: Evidence –based developmental outcomes. Pediatrics, 108,e84. Retrieved October 7, 2002 from UMDNJ database.
Ulrich, B.D., Ulrich D.A., Collier D.H. (1992 March). Alternating stepping patterns: Hidden abilities of 11-month-old children with Down syndrome. Developmental Medicine and Child Neurology, 34, 233-9.
Ulrich, B.D., Ulrich D. A., Collier D.H., Cole E.L. (1995 January). Developmental shifts in the ability of infants with Down syndrome to produce treadmill steps. Physical Therapy, 75, 14-23.
van Ingen Schenau, G.J. (1980). Some fundamental aspects of the biomechanics of overground versus treadmill locomotion. Medicine and Science in Sports and Exercise, 12, 257-61.
Zelazo, P.R., Zelazo, N.A., Kolb, S. (1972, April 21). “Walking” in the newborn. Science, 176, 314-15.
Breniere, Y., Bril, B. (1998, August). Development of postural control of gravity forces in children during the first five years of walking. Experimental Brain Research, 121, 255-62.
Filion, G., Richards, C., Malouin, F. (1993, November). Changes induced by weight support on gait in normal children during treadmill walking [Abstract]. Archives of Physical Medicine and Rehabilitation, 74, 1265-66.
Finch, L., Barbeau, H., Arsenault, B. (1991, November). Influence of body weight support on normal human gait: Development of a gait retraining strategy. Physical Therapy, 71, 842-56.
Myklebust, B.M., Gottlieb, G.L., Agarwal, G.C. (1986, August). Stretch reflexes in the normal infant. Developmental Medicine and Child Neurology, 28, 440-9.
Pillar, T., Dickstein, R., Smolinski, Z. (1991, Fall). Walking reeducation with partial relief of body weight in rehabilitation of patients with locomotor disabilities. Journal of Rehabilitation Research, 28, 47-52.
Example of an Ovid Search of CINAHL
Database: CINAHL <1982 to October Week 4 2002>
Search Strategy:
--------------------------------------------------------------------------------
1 body weight support.mp. or Weight-Bearing/ (309)
2 treadmill.mp. or Treadmills/ (1200)
3 ambulation.mp. or Walking/ (2193)
4 gait.mp. or GAIT ANALYSIS/ or GAIT/ (1836)
5 locomotion.mp. or LOCOMOTION/ (221)
6 1 or 2 (1462)
7 3 or 4 or 5 (3655)
8 brain injuries.mp. or Brain Injuries/ (3090)
9 cerebral palsy.mp. or Cerebral Palsy/ (1278)
10 Down syndrome.mp. or Down Syndrome/ (663)
11 8 or 9 or 10 (4983)
12 6 and 7 (344)
13 11 and 12 (11)