Energy cost of walking in children with cerebral palsy: relation to the Gross Motor Function Classification System

Energy cost of walking in children with cerebral palsy: relation to the Gross Motor Function Classification System Therese E Johnston* MSPT, Shriners Hospitals for Children, Philadelphia, PA; Stephanie E Moore DPT; Lance T Quinn DPT, Arcadia University, Glenside, PA; Brian T Smith MS, Shriners Hospitals for Children, Philadelphia, PA, USA. *Correspondence to first author at Shriners Hospitals for Children, 3551 North Broad Street, Philadelphia, PA 19140, USA. E-mail: tjohnston@shrinenet.org This study compared the energy cost of walking in children with cerebral palsy (CP) classified at different levels of the Gross Motor Function Classification System (GMFCS) with that in children with typical development. Sixteen female and 14 male children with CP (mean age 9 years 6 months, SD 2 years 4 months, range 6 years 4 months to 13 years 4 months) and 14 male and 13 female typically developing children (mean age 10 years, SD 1 year 6 months, range 7 years 1 month to 12 years 11 months) participated. Children with CP were classified at GMFCS level I, n=5; level II, n=10; level III, n=9; and level IV, n=6. Energy cost was assessed by the gas dilution method as each child walked around an oval track wearing a dilution mask. Significant differences were found across GMFCS levels (p<0.0001) and between adjacent levels (p 0.013). Children with CP displayed a higher energy cost of walking than the typically developing children (p<0.0001). A strong correlation (0.87) was found between the energy cost of walking and GMFCS level (p<0.01) when children with typical development were assigned a GMFCS level of zero to allow statistical analysis. This indicates increasing energy cost of walking with increasing severity of functional involvement. These differences in energy cost across GMFCS levels provide another distinguishing factor between GMFCS levels and further emphasize the importance of considering metabolic demand in determining treatment options. An important aspect of typical adult walking is conservation of energy. This is accomplished by minimizing the displacement of the center of gravity while progressing forward (Fisher and Gullickson 1978, Waters and Mulroy 1999) in a steady, minimal, wavelike pattern (Saunders et al. 1953). Other factors involved in an efficient walking pattern include stride length, walking speed, changes in vertical position, accelerations of body segments, forces across joints, muscle action, and the external environment (Williams 1985). Yet perhaps the most intriguing factor is the self-optimizing principle which states that a person will free-walk at a self-selected speed that minimizes energy use (Jeng et al. 1997), termed the economical walking speed. Children with cerebral palsy (CP) have been shown to expend greater energy during walking than their typically developing peers when assessed by measuring either oxygen uptake or heart rate (Campbell and Ball 1978; Rose et al. 1990, 1993). This increase has been shown when walking at a given speed (Rose et al. 1993) as well as at the child s self-selected economical walking speed (Rose et al. 1990). Economical walking speed has been shown to increase with age for children with typical development but not for children with CP, who may maintain or decrease this speed as they become older (Rose et al. 1990). Energy expenditure has also been shown to decrease with age for typically developing children (Campbell and Ball 1978, Waters et al. 1988) but to increase for children with CP (Campbell and Ball 1978). The increase for children with CP might be related to a greater reliance on wheelchairs to maintain functional mobility with aging and the inability to adapt to the changes in body weight and size that occur with growth (Campbell and Ball 1978). One area not yet investigated is the relationship between energy expenditure and function. Gross motor functional abilities of a child with CP can be classified according to the Gross Motor Function Classification System (GMFCS; Palisano et al. 1997). The GMFCS is a five-level functional grading system for children with CP aged between 12 months and 12 years and is based on typical self-initiated performance (Wood and Rosenbaum 2000). The GMFCS has been shown to have content validity (Palisano et al. 1997), good interrater reliability (0.75 to 0.79; Palisano et al. 1997, Wood and Rosenbaum 2000), and excellent intrarater reliability (0.93; Wood and Rosenbaum 2000). Palisano et al. (2000) and Rosenbaum et al. (2002) also found the GMFCS to be predictive of future gross motor function outcomes. If a relationship exists between energy expenditure and gross motor function, functional distinctions between adjacent GMFCS levels might be further supported. The purpose of this study was threefold: (1) to compare the energy cost of walking in children with CP of different GMFCS levels; (2) to compare the energy cost of walking in children with CP with that in children with typical development; and (3) to identify whether a relationship exists between energy cost and GMFCS levels, with typically developing children included. Method Thirty children with CP (16 females, 14 males; mean age 9 years 6 months, SD 2 years 4 months, range 6 years 4 months to 13 years 4 months) participated in the study (Table I). Each of these children was a participant in another prospective study that used energy cost of walking as an outcome measure, and 34 Developmental Medicine & Child Neurology 2004, 46: 34 38

each child s baseline energy cost measure from the other study was used in this study. These children were recruited through an outpatient clinic located within Shriners Hospitals for Children, Philadelphia, PA, USA. The children were each assigned a GMFCS level by a physical therapist (Table II). Twenty-seven children with typical development also participated prospectively in this study (Table I). These children were recruited through colleagues of the authors or through families of patients within the hospital system. The parent(s) of each child signed an informed consent form approved by the governing Institutional Review Board, and each child provided verbal assent. Energy expenditure was measured by assessing the energy cost of walking by means of the gas dilution method (Harris and Wertsch 1994, Boyd et al. 1999) using a SensorMedics VMax29 metabolic cart (SensorMedics Corporation, Yorba Linda, CA, USA). This method was chosen because it has been shown to be more reliable (Bowen et al. 1998) and more responsive than other methods of assessing energy expenditure in children with CP (IJzerman and Nene 2002). Bowen et al. (1998) reported an average variability in scores of 13.2% across 5 days for children with CP using the gas dilution method, compared with 17.5 and 20.3% using other methods. Children were asked to refrain from eating for at least 2 hours before testing. Children with CP wore ankle foot orthoses during the test if an orthosis was typically used for walking. The volume of oxygen consumed per kilogram of body weight (ml/kg), also called energy consumption, was measured every 60 seconds by means of a dilution mask attached by a hose to the metabolic cart (Fig. 1). The volume of oxygen consumed was measured under four consecutive conditions: (1) sitting quietly for at least 5 minutes to establish 5 minutes of steady-state resting values; (2) walking at a selfselected walking speed for 2 minutes along the 24-meter oval track to allow the body to warm up; (3) walking at a self-selected walking speed for at least 5 minutes along a 24-meter oval track to establish 5 minutes of steady-state walking values (Boyd et al. 1999), and (4) sitting quietly to establish 3 minutes of steady-state recovery values (Fig. 2). Distance traveled per minute was recorded while walking along the track during the establishment of the five minutes of steady-state walking values. The energy cost of walking (ml/kg/m) was then calculated for each of the 5 minutes of walking by dividing the energy consumption per minute (ml/kg) by the distance traveled during that minute (m/min; Waters 1992, Harris and Wertsch 1994, Boyd et al. 1999, Waters and Mulroy 1999). Energy costs of walking values were analyzed with a oneway analysis of variance (ANOVA) to determine any significant differences between (1) children with CP in relation to GMFCS levels and (2) between children with CP of all levels and typically developing children. A p value of less than 0.05 was accepted for significance. A Bonferroni post-hoc analysis was applied to examine any differences in energy cost between adjacent GMFCS levels. For this analysis, a p value of less than 0.017 was accepted for significance because three comparisons were made (level I vs level II, level II vs level III, and level III vs level IV). A Spearman s correlation coefficient was used to examine the relationship between energy cost and GMFCS levels for all children. For this correlational analysis, the children with typical development were assigned an artificial GMFCS level of zero for statistical analysis, to allow their data to be included in the examination of this relationship. Results There was a main effect of energy cost of walking across GMFCS levels for the children with CP (p<0.0001): an increasing energy cost of walking was seen with increasing GMFCS level (Fig. 3). Bonferroni post-hoc analyses revealed significant differences between each adjacent GMFCS level (p 0.013). As a group (combined data), the children with CP displayed a higher energy cost of walking (mean 0.79, [SD 0.84] ml/kg/m) than the children with typical development (mean 0.23 [SD 0.03] ml/kg/m; p<0.0001). When the children with typical development were included with a GMFCS level of zero, there was a strong correlation (0.87) between energy cost of walking and GMFCS level (p<0.01). In a secondary analysis, children with CP classified as level I were compared with children with typical development by Table I: Characteristics of participating children Characteristics CP group (n=30) TD group (n=27) Age, y:m Mean (SD) 9:6 (2:4) 10:0 (1:6) Range 6:4 13:4 7:1 12:11 Sex, n Male 14 14 Female 16 13 Diagnosis, n Hemiplegia 2 Diplegia 23 Quadriplegia 5 GMFCS level, n I 5 II 10 III 9 IV 6 CP, cerebral palsy; TD, typically developing. Table II: Gross Motor Function Classification System (GMFCS) levels (Palisano et al. 1997) GMFCS level I II III IV V Description Walks without restrictions; limitations in more advanced gross motor skills Walks without assistive devices; limitations in walking outdoors and in community Walks with assistive devices; limitations in walking outdoors and in community Self-mobility with limitations; children are transported or use powered mobility outdoors and in community Self-mobility is severely limited, even with use of assistive technology Energy Cost of Walking Therese E Johnston et al. 35

using a one-tailed t-test, assuming unequal variances. There was a significant difference between these two groups (p<0.0001). Discussion A difference was found between the energy cost of walking at different GMFCS levels, providing another possible distinguishing factor between these levels in addition to the functional components described within the GMFCS itself. This difference further supports the specific GMFCS levels that were developed to capture clinically meaningful distinctions between children with CP in terms of their motor function. In addition, children classified as GMFCS level I displayed a greater energy cost than children with typical development, suggesting that their impairments were significant enough to affect energy expenditure during walking. Reasons for the potential differences across levels might include the increasing severity of impairments such as strength, co-contraction, or spasticity and inefficient energy transfers between body segments, all of which have been shown to be related to energy expenditure during walking for children with CP (Mann 1983, Olney et al. 1987, Gage 1993, Kramer and MacPhail 1994, Unnithan et al. 1996). Kramer and MacPhail (1994) found that knee extensor strength was related to energy expenditure, with the relationship becoming stronger at faster walking speeds. Lower extremity muscle co-contraction or spasticity while walking might increase energy expenditure by causing the displacement of the center of gravity to vary, leading to a less efficient walking pattern (Mann 1983, Gage 1993). Finally, alterations in segmental energy transfer might contribute to increased energy expenditure because children with CP use atypical patterns of movement, which might interfere with the natural exchange or transfers of potential and kinetic energy between body segments (Olney et al. 1987). It has been shown that these deficient energy transfers account for as much as 87% of the variability in the energy cost of walking for children with CP (Unnithan et al. 1999). Some of these same variables have also been related to function in children with CP. Kramer and MacPhail (1994) found a VO 2 /kg (ml/kg/min) 30 25 20 15 10 Baseline Warm-up Exercise Recovery 5 0 0 5 10 15 20 25 Minutes Figure 2: Sample of data collected using metabolic cart. ml/kg/m 3.5 3 2.5 2 1.5 1 0.5 0 0.23 0.28 0.44 TD I II III IV (n=27) (n=5) (n=10) (n=9) (n=6) GMFCS level 0.63 2.17 Figure 1: A child with cerebral palsy participates in data collection. Hose connects mask to computer, enabling measurement of volume of oxygen consumed while walking. Figure 3: Means and standard deviations of energy cost of walking values per GMFCS level and for children with typical development (TD). By post-hoc testing, differences were significant between each adjacent GMFCS level: level I versus level II (p=0.0105), level II versus level III (p=0.0130), and level III versus level IV (p=0.0003). 36 Developmental Medicine & Child Neurology 2004, 46: 34 38

modest relationship (0.57 to 0.69) between knee extensor strength and Gross Motor Function Measure (GMFM) scores in adolescents with mild CP. Engsberg and Ross (2002) found a significant relationship between strength and function, which accounted for up to 68% of the variance seen between spasticity, strength, gait, and GMFM in 95 children with spastic diplegia. They also identified a significant relationship between spasticity and function but reported that it contributed only up to 17% of the variance. Damiano et al. (2001) reported a relationship between spasticity as measured by passive resistance and GMFM scores in children with spastic CP of various GMFCS levels. The GMFCS has been shown to be a good predictive tool for future functional abilities as measured by the GMFM (Palisano et al. 2000, Rosenbaum et al. 2002). Wood and Rosenbaum (2000) have suggested that the GMFCS might be a useful tool in evaluating the impact of intervention on children with CP by determining whether treatment provides the child with the tools to advance beyond what would be expected from the GMFCS level. Rosenbaum et al. (2002) suggest that developing children with CP might improve their gross motor performance by improving in areas such as endurance and energy efficiency. This challenges clinicians to examine and consider interventions that might lead to a decrease in energy expenditure. Interventions could involve the use of orthoses, assistive devices, muscle strengthening, or other interventions designed to enhance motor function. In our study, children classified as GMFCS level IV used walkers, whereas those classified as level III used a cane, crutches, or a walker. Studies of energy expenditure in children with CP have been limited to comparing anterior and posterior walkers. Park et al. (2001) reported that energy expenditure was less with a posterior walker, whereas Mattsson and Andersson (1997) reported no difference. Other research has examined this topic in different populations. Melis et al. (1999) reported that adults with incomplete spinal cord injury who used a walker walked more slowly than those using forearm crutches or a cane. Given that the denominator for the energy cost formula is walking speed, this identified difference in speed might affect energy cost. In adults with incomplete spinal cord injury or arthritis, the following order has been identified for energy cost from highest to lowest: walker, crutches, cane or crutch, no device (Waters 1992). The exact causal relationship cannot be surmised from these studies because those using a walker might have more impairments. More research is needed in this area for children with CP. Flexed knee gait has been shown to affect energy cost; rising cost is seen with increasing knee flexion angles in adults without disability (Winter 1983, Waters 1992). Waters (1992) reported an increase in energy cost due to a rise in energy consumption (ml/kg) as well as a decrease in walking speed. Duffy et al. (1997) only reported a change due to decreasing walking speed in children with typical development and suggested that other factors might contribute to increased energy cost in children with CP. More importantly for children with CP, weakened quadriceps muscles have been shown to be related to a crouched gait position (Damiano et al. 1995a). Strengthening of these muscles has been shown to lead to a decrease in the knee flexion angle (Damiano et al. 1995b), suggesting the potential for energy cost to also be affected. More research is needed to examine the relationships between energy cost and the above-mentioned variables, to determine whether decreasing the energy expended during walking, through various interventions, can assist a child in progressing beyond what would be expected from their current GMFCS level. An important group to study further is the more limited ambulators (GMFCS levels III and IV) who currently expend large amounts of energy to walk. Interventions that can allow them to be more energy efficient during walking might provide them with a desire to be ambulatory for a greater part of the day. Conclusion A difference in energy cost of walking across the GMFCS levels was found, providing another distinguishing factor between GMFCS levels. This difference is important because it reflects a difference in metabolic demand during walking: the more impaired children with higher GMFCS levels place a greater metabolic demand on their systems while walking. These findings have implications for therapeutic intervention and challenge clinicians to investigate treatment options that might reduce the energy demand during walking. DOI:10.1017/S0012162204000064 Accepted for publication 11th September 2003. Acknowledgements We thank John Gaughan PhD for his assistance with the statistical analysis, Margo Orlin PhD PT PCS for reviewing the manuscript, and Samuel CK Lee PhD PT, Samuel R Pierce PT MS NCS, Richard L Finson MS PT, Tracy Ronan DPT, Shannon Stuckey MPT, and Katherine W Parker SPT for their assistance with data collection. This study was funded by Shriners Hospitals for Children, grant no. 8530. References Bowen TR, Lennon N, Castagno P, Miller F, Richards J. (1998) Variability of energy-consumption measures in children with cerebral palsy. J Pediatr Orthop 18: 738 742. Boyd R, Fatone S, Rodda J, Olesch C, Starr R, Cullis E, Gallagher D, Carlin JB, Nattrass GR, Graham K. (1999) High- or- low technology measurements of energy expenditure in clinical gait analysis? Dev Med Child Neurol 41: 676 682. Campbell J, Ball J. (1978) Energetics of walking in cerebral palsy. Orthop Clin N Am 9: 374 377. Damiano DL, Kelly LE, Vaughan CL. (1995a) Effects of quadriceps femoris muscle strengthening on crouch gait in children with spastic diplegia. Phys Ther 75: 658 671. Damiano DL, Quinlivan J, Owen BF, Shaffrey M, Abel MF. (2001) Spasticity versus strength in cerebral palsy: relationships among involuntary resistance, voluntary torque, and motor function. Eur J Neurol 8: 40 49. Damiano DL, Vaughan CL, Abel MF. (1995b) Muscle response to heavy resistance exercise in children with spastic cerebral palsy. Dev Med Child Neurol 37: 731 739. Duffy CM, Hill AE, Graham HK. (1997) The influence of flexed-knee gait on the energy cost of walking in children. Dev Med Child Neurol 39: 234 238. Engsberg JR, Ross SA. (2002) Relationships between spasticity, strength, gait, and GMFM in spastic diplegia cerebral palsy. Dev Med Child Neurol 44 (Suppl. 91): 37. (Abstract). Fisher SV, Gullickson G. (1978) Energy cost of ambulation in health and disability: a literature review. Arch Phys Med Rehabil 59: 124 133. Gage JR. (1993) Gait analysis; an essential tool in the treatment of cerebral palsy. Clin Orthop Rel Res 288: 126 134. Harris G, Wertsch JJ. (1994) Procedures for gait analysis. Arch Phys Med Rehabil 75: 216 225. IJzerman MJ, Nene AV. (2002) Feasibility of the physiological cost index as an outcome measure for the assessment of energy expenditure during walking. Arch Phys Med Rehabil 83: 1777 1782. Energy Cost of Walking Therese E Johnston et al. 37

Jeng SF, Liao HF, Lai JS, Hou, JW. (1997) Optimization of walking in children. Med Sci Sports Exer 29: 370 376. Kramer JF, MacPhail HEA. (1994) Relationships among measures of walking efficiency, gross motor ability, and isokinetic strength in adolescents with cerebral palsy. Pediatr Phys Ther 6: 3 8. Mann R. (1983) Biomechanics in cerebral palsy. Foot Ankle 4: 114 119. Mattsson E, Andersson C. (1997) Oxygen cost, walking speed, and perceived exertion in children with cerebral palsy when walking with anterior and posterior walkers. Dev Med Child Neurol 39: 671 676. Melis EH, Torres-Moreno R, Barbeau H, Lemaire ED. (1999) Analysis of assisted-gait characteristics in persons with incomplete spinal cord injury. Spinal Cord 37: 430 439. Olney SJ, Costigan PA, Hedden DM. (1987) Mechanical energy patterns in gait of cerebral palsied children with hemiplegia. Phys Ther 67: 1348 1354. Palisano RJ, Hanna SE, Rosenbaum PL, Russell DJ, Walter SD, Wood EP, Raina PS, Galuppi BE. (2000) Validation of a model of gross motor function for children with cerebral palsy. Phys Ther 80: 974 985. Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galappi B. (1997) Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol 39: 214 223. Park ES, Park CI, Kim JY. (2001) Comparison of anterior and posterior walkers with respect to gait parameters and energy expenditure of children with spastic diplegic cerebral palsy. Yonsei Med J 42: 180 184. Rose J, Gamble JG, Burgos A, Medeiros J, Haskell WL. (1990) Energy expenditure index of walking for normal children and for children with cerebral palsy. Dev Med Child Neurol 32: 333 340. Rose J, Haskell WL, Gamble JG. (1993) A comparison of oxygen pulse and respiratory exchange ratio in cerebral palsied and nondisabled children. Arch Phys Med Rehabil 74: 702 705. Rosenbaum PL, Walter SD, Hanna SE, Palisano RJ, Russell DJ, Raina P, Wood E, Bartlett DJ, Galuppi BE. (2002) Prognosis for gross motor function in cerebral palsy: creation of motor development curves. JAMA 288: 1357 1363. Saunders JB, Inman VT, Eberhart HD. (1953) Major determinants in normal and pathological gait. J Bone Joint Surg 35: 543 558. Unnithan VB, Dowling JJ, Frost G, Bar-Or O. (1996) Role of cocontraction in the O 2 cost of walking in children with cerebral palsy. Med Sci Sports Exer 28: 1498 1504. Unnithan VB, Dowling JJ, Frost G, Bar-Or O. (1999) Role of mechanical power estimates in the O 2 cost of walking in children with cerebral palsy. Med Sci Sports Exer 31: 1703 1708. Waters RL. (1992) Energy expenditure. In: Perry J, editor. Gait Analysis: Normal and Pathological Function. Thorofare, NJ: Slack Inc. p 443 489. Waters RL, Mulroy S. (1999) The energy expenditure of normal and pathologic gait. Gait Post 9: 207 231. Waters RL, Lunsford BR, Perry J, Byrd R. (1988) Energy speed relationship of walking: standard tables. J Orthop Res 6: 215 222. Williams KR. (1985) The relationship between mechanical and physiological energy estimates. Med Sci Sports Exer 17: 317 325. Winter DA. (1983) Knee flexion during stance as a determinant of inefficient walking. Phys Ther 63: 331 333. Wood E, Rosenbaum P. (2000) The Gross Motor Function Classification System for cerebral palsy: a study of reliability and stability over time. Dev Med Child Neurol 42: 292 296. 38 Developmental Medicine & Child Neurology 2004, 46: 34 38