Net V̇E was lower on day 3 than on day 1 (P = 0.04) and day 2 (P = 0.03), but only when subjects walked at 90% FWS (Fig. 1). When subjects walked at 90% FWS, net V̇E was higher than when they walked at 60% FWS (P = 0.01). There were no differences in net V̇E at 75% compared with 60 or 90% FWS (Fig. 1). Between-day reliability of net V̇E was high (R, Table 4). Intrasubject, between-day variability was lowest when subjects walked at 90% FWS and highest when they walked at 60% FWS (CV, Table 4). Repeated treadmill walks on different days did not affect net RR at any speed (Table 3). When subjects walked at 90% FWS net RR was higher than when they walked at 60 (P = 0.0005) and 75% FWS (P = 0.02). There was no significant difference when subjects walked at 75 and 60% FWS (Table 3). Between-day reliability of net RR was moderate (R, Table 4). Intrasubject, between-day variability was overall high, but lowest when subjects walked at 60% FWS and highest when they walked at 75% FWS (CV, Table 4).
When subjects walked at 90% FWS, net HR was lower on day 3 than on day 1 (P = 0.0004, Fig. 2), but day 2 values were not significantly different from those on day 1 or day 3 (Fig. 2). When they walked at 60% FWS, net HR was lower on day 3 than on day 2 (P = 0.03), whereas day 1 values were not significantly different from day 2 or day 3 values (Fig. 2). Net HR was higher when subjects walked at 90 than at 75 (P = 0.0003) or 60% FWS (P = 0.004). There was no significant difference between 60 and 75% FWS. Between-day reliability was high (R, Table 4). Intrasubject, between-day variability was also high, but lowest when subjects walked at 75% FWS and highest when they walked at 60% FWS (CV, Table 4).
To the best of the authors’ knowledge, this is the first study to examine the effect of repeated, multi-day, multi-speed treadmill sessions on physiologic variables in children and adolescents with mild CP. The authors’ hypothesized that there would be no between-day differences when subjects walked on the treadmill at a relatively slow speed but that at faster speeds their physiologic strain would be lower on day 2 and day 3 compared with day 1. When subjects walked at the fastest speed (90% FWS), there was a decrease from day 1 to day 3 in net V̇E (3.6 L·min−1;Fig. 1) and net HR (8 beat·min−1;Fig. 2), thereby partly confirming our hypothesis that between-day differences would be found at the faster speed but not at the slower speeds. With the exception of net HR (where there was only a trend for this same pattern at 60% FWS;Fig. 2), there were no other significant differences from day 1 to day 3 in any variable, at any speed.
The present study had adequate statistical power (0.80) to detect minimum between-day differences in net V̇E at any speed of about 3.5 L·min−1. Figure 1 shows there is no reduction in net V̇E over successive days that was missed due to low statistical power. Our failure to find a significant difference in net V̇E between 75 and 90% FWS independent of day, however, is likely due to low statistical power. There was adequate power to determine minimum between-speed differences of about 9.9 L·min−1, and the difference in this case was 8.3 L·min−1. We found a reduction in the net V̇E with repeated treadmill walks at a relatively high speed but not at the lower speeds. In the absence of a similar pattern in V̇O2, this suggests that at the higher exercise intensity (90% FWS), V̇E, and V̇O2 were no longer tightly coupled, perhaps because the subjects were working above their ventilatory threshold. Because the subjects did not perform a progressive exercise test to peak V̇O2, however, ventilatory threshold cannot be accurately determined for these subjects. We previously reported a similar finding with this population (15). The group mean net V̇E in the present study (Fig. 1) was similar to that of the subjects in this previous study (net V̇E at 90% FWS = 22 L·min−1).
The study had adequate statistical power (0.80) to detect minimum between-day differences in net HR at any speed of about 5 beat·min−1. Figure 2 shows that no new between-day patterns in HR were missed. Had statistical power been higher, it would have been more clearly shown that when subjects walked at 90% FWS, net HR decreased with each visit (day) and when they walked at 60% FWS, net HR was lower on day 3 compared with both day 1 and day 2. Our failure to find a significant difference in net HR between 60 and 75% FWS independent of testing day, however, is likely due to low statistical power. There was adequate power to determine minimum between-speed differences of about 11 beat·min−1, and the difference in this case was 8 beat·min−1. We found net HR was reduced with repeated treadmill walks on different days at 90% FWS (and with a trend for the same finding at 60% FWS). Although not specifically measured in this study, and in the absence of a similar pattern for the V̇O2 variables, it is possible that anxiety (due to walking relatively fast and relatively slow), which can increase HR (10,13), may have decreased over time as subjects became more familiar with treadmill walking without holding on to the handrails and with the testing environment and procedures. Although subjects were introduced to all equipment and procedures during the introductory visit, anxiety-related increases in HR cannot be ruled out. It is possible that anxiety during treadmill walking is relevant to HR with this population, who have difficulty walking, and not to healthy children where this phenomenon is not observed (8,30). Furthermore, the extent of walking impairment appears to affect HR more than V̇O2 in this population, which again suggests that factors other than exercise intensity elevate HR in those with CP. Compared with the present group (Fig. 2), while walking at the same relative intensity (90% FWS), children and adolescents with CP who have more difficulty walking (lower walking-related gross motor function scores and slower ground walking speeds than the present group) demonstrated 51% higher net HR but only 13% higher net V̇O2-kg (15). In the future, a measure of anxiety could prove useful to more clearly determine the cause of the between-day differences in HR. It is possible that more practice time walking on the treadmill is needed to reduce anxiety-related between-day differences in HR.
We had adequate statistical power (0.80) to detect minimum between-day differences in net V̇O2-kg at any speed of about 1.2 mL·kg−1·min−1, which was the difference between day 3 and day 2, independent of speed. With a larger sample, the 1.1 mL·kg−1·min−1 increase from day 1 to day 2, irrespective of speed would likely have also been significant. Our failure to find a significant difference in net V̇O2-kg between 60 and 75% FWS, independent of testing day is probably due to low statistical power, as there was adequate power to determine a minimum between-speed difference of about 4.4 mL·kg−1·min−1, and the difference in this case was 1.2 mL·kg−1·min−1. It is difficult to determine from the results of the present study why net V̇O2-kg was higher on day 2. Subjects may have walked differently on day 2 (and hence were less economical) than on day 1 or day 3, although a post hoc analysis of stride data (not shown) revealed no significant differences in stride length or rate among the three days. The between-day differences in net V̇O2-kg are small, however, compared with the differences among subjects, and thus between-day reliability is high (Table 4). A previous treadmill walking study with 6-yr-old healthy children (27) showed that, after 5 min of treadmill walking practice, within-day net V̇O2-kg was higher for the first than for the second and third trial. In this previous study, however, only one speed was used (1.34 m·s−1), which was about 30% faster than the group mean speed for 90% FWS (1.01 m·s−1). Thus, it is possible that we would have found speed-related differences in the pattern of the V̇O2 response to repeated treadmill walks had subjects been able to walk at a faster speed and therefore at a higher absolute exercise intensity. In other words, more practice time may be needed for very, very mild subjects with CP who, like young healthy children, have an immature gait pattern but are able to walk faster than the present subjects with CP. When 6- to 15-yr-old children with mild spastic hemiplegic CP walked at 0.67 m·s−1, the same speed as the group mean speed for our slowest speed, for example, there were no within-day differences in net V̇O2-kg, which also suggests that a clear reduction in net V̇O2 with repeated treadmill walks is not seen with slower speeds in CP. When subjects walked on the treadmill at 90% FWS, mean net V̇O2-kg for the group (Table 3) was similar to that reported in the literature (16–21 mL·kg−1·min−1) for children and adolescents with mild CP (15,29).
Post hoc sample size calculations for the other metabolic variables, V̇O2-stride and EE, showed the same results as for V̇O2-kg, i.e., the sample of eight subjects was sufficient to detect the main between-day difference (higher values on day 2 than on day 3 with no difference between day 1 and day 3;Table 3). As with V̇O2-kg, it is likely that there was insufficient power to detect the speed-related differences that were not significant (Table 3). Net V̇O2-stride showed the same pattern as net V̇O2-kg, possibly because there were no significant between-day differences in stride rate (not shown). We expected that stride rates would decrease over repeated walks on the treadmill as subjects became more familiar with treadmill walking and were able to take longer steps, but this was not the case. Our subjects, like healthy children (8), did not show any between-day differences in stride rate. EE showed the same pattern as V̇O2-kg, possibly because there were no significant interday differences in RER (not shown). We attempted to control fuel source somewhat by having the subjects fast (with the exception of water) for 2 h before coming to the lab (no caffeine for the preceding 3 h). We also tested each subject at the same time of day for the three visits. The last meal before testing, according to the information given to us by the subject or parent, was usually similar in content from testing day to testing day.
Between-day differences in all physiologic variables were relatively small compared with the intersubject differences, and thus between-day reliability of all variables, irrespective of speed, was high, with the exception of RR, where reliability was moderate (Table 4). In other words, day-to-day intrasubject differences in these measures (with the exception of RR) have minimal effect on the ability of these measures to reliably differentiate between subjects. Our between-day reliability (Table 4) for net V̇O2-kg was greater than the within-day reliability reported in the literature (0.78) for children and adolescents with spastic hemiplegic CP (11). This is perhaps because our subjects received more walking practice (12–15 min compared with 5 min). It is also possible that the subjects themselves were more similar to each other in the previous study (all had hemiplegic CP) and, thus, compared with the present study, between-trial differences in V̇O2 were relatively greater than the between-subject differences.
Intrasubject, between-day variability was less for the metabolic variables at 90% FWS than at the slower speeds (Table 4). In a clinical setting, where V̇O2 during treadmill walking might be tested for an individual before and after an intervention, it would perhaps be prudent to test such individuals at close to their fastest treadmill walking speed, rather than at slower speeds, especially if the expected difference in V̇O2 due to the intervention might be small. Net V̇E showed higher intrasubject, between-day variability than the V̇O2-related measures (Table 4). It may not be the most suitable physiologic outcome measure in a clinical setting. Although net HR and net RR both showed higher intrasubject, between-day variability than the other measures, this is mostly a “mathematical artifact” due to low values for both net HR and net RR. When absolute HR and RR are considered, the CV is much lower (about 4% for HR and 9–10% for RR, across the three speeds). Our mean intrasubject variability for net V̇O2-kg during treadmill walking was similar to that reported in the literature for this population (8.4 ± 8.5%) (11). Our mean intrasubject variability for absolute V̇O2-kg, 5.7–8.1%, is lower than what is reported for over ground walking (4) (17% when speed is not controlled for, 13% when speed is controlled for and V̇O2 is calculated per meter walked). The higher variability for over ground walking may reflect the greater motor impairment of the subjects in that study in that some of them used braces and walking aids. It is also possible that with speed (and the width of the walking track) more precisely controlled with treadmill walking, subjects walk more similarly from trial to trial on the treadmill than over ground. If this is true, then a treadmill walking protocol might be more appropriate than an overground one for assessment of an intervention, especially if small, but clinically relevant differences are expected. Further research is needed to determine whether physiologic responses during treadmill walking are more reliable and less variable than during over ground walking in those with mild CP.
Only 81% of individual trials reached the steady state criteria in this study, which is similar to that (82%) reported for healthy children (8). It is difficult to say, but possible, that interindividual variability was increased due to difficulties with achieving steady state in 19% of the trials. Previous research, however, has shown that group V̇O2 values are not significantly different between minutes 2, 3, or 4 during treadmill walking in this population (29). It is also possible that patterns in the data were obscured due to the subjects having resting measures taken only on day 1. We elected to have only one resting measurement session for each subject to decrease the time commitment for subjects (this study was part of a larger study and subjects were making a total of seven visits to the lab, three after this phase of the study was complete). We also elected to have only one resting measurement session to decrease the between-day variability in the net physiologic measures due to the increased susceptibility of anxiety-related differences in resting measures. Because our RER measures were not different among the days, it is likely that there were no between-day differences in V̇O2 due to fuel or diet differences and thus the resting metabolic measures, if not affected by factors such as anxiety, would not have differed greatly among the testing days.
It is difficult to determine from our results or from the literature what the optimum protocol is to ensure that subjects with CP are habituated to treadmill walking. It appears that as little as 5 min of treadmill walking practice is needed to obtain reliable V̇O2 data if the walking speed is slow (11). Based on our results, after 12–15 min of walking practice, reliable and stable metabolic data may be collected at various speeds. Both comfortable and fast ground walking speeds (Table 2) were, however, on average faster for the group than any of the treadmill walking speeds. The fastest treadmill speed (90% FWS) was closest to the comfortable ground walking speed (mean intrasubject difference = 0.25 ± 0.47 m·s−1). It has previously been shown that the comfortable walking speed on the treadmill is slower than the comfortable walking speed on the ground for children with mild spastic hemiplegic CP (9). The subjects in the previous study received a similar amount of treadmill walking practice to the present subjects. Thus, perhaps more practice would result in increases in the comfortable and fastest walking speeds on the treadmill and therefore decrease the differences between the comfortable and fast walking speeds on the ground and treadmill.
In conclusion, this is the first study to examine in children and adolescents with mild spastic CP, the effect on physiologic responses of repeated treadmill walks on different days, and whether the effects are different at different speeds. Irrespective of speed, in this sample, net metabolic responses on day 1 did not differ from those on day 3. Net V̇E and net HR were both lower on day 3 compared with day 1, but only when subjects walked at the fastest speed (90% FWS). Because metabolic responses did not show the same pattern as these cardiorespiratory responses, it is possible that the reduction over time in net HR was due to a reduction in anxiety rather than an improvement in walking economy per se and the reductions in net V̇E, to an uncoupling of V̇O2 and V̇E. Thus, these two variables may not be the most appropriate to measure if researchers wish to use a faster walking speed and only 12–15 min of treadmill walking practice. Researchers interested in mechanisms and interventions related to walking economy in subjects with mild spastic CP may be able to collect reliable net physiologic and stable net metabolic variables, especially those related to V̇O2, after one treadmill walking practice session. More research is needed to determine whether physiologic responses during treadmill walking are more reliable and less variable than during over ground walking in those with mild CP. More research is also needed to determine the optimal treadmill walking habituation protocol in this population and whether more treadmill walking practice would result in greater similarity between comfortable and fast walking speeds on the ground and on the treadmill. A larger sample, of subjects with mild CP that included subjects who walk faster and slower than the present group, would also increase the generalizability of our findings, which at present likely relate best to those whose walking speeds are similar to the present group.
We would like to thank the volunteers and their families for their participation in this study. We would also like to thank N. Huybrechts, B. Smith, and B. Timmons for assistance with data collection and M. L. Schmuck for assistance with data analysis.
This study was supported by a grant from the Bloorview Children’s Hospital Foundation.
1. B ar- O r, O. Pediatric Sports Medicine for the Practitioner: From Physiologic Principals to Clinic Applications.
New York: Springer Verlag, 1983, pp. 343–348.
2. Bohannon, R. W., and M. B. Smith. Interrater reliability of a modified Ashworth scale of muscle spasticity. Phys. Ther. 67: 206–207, 1987.
3. Booth, C. M., M. J. Cortina-Borja, and T. N. Theologis. Collagen accumulation in muscles of children with cerebral palsy and correlation with severity of spasticity. Dev. Med. Child Neurol. 43: 314–320, 2001.
4. Bowen, T. R., N. Lennon, P. Castagno, F. Miller, and J. Richards. Variability of energy-consumption measures in children with cerebral palsy. J. Pediatr. Orthop. 18: 738–742, 1998.
5. Campbell, J., and J. Ball. Energetics of walking in cerebral palsy. Orthop. Clin. North Am. 9: 374–377, 1978.
6. Dahlbäck, G. O., and R. Norlin. The effect of corrective surgery on energy expenditure during ambulation in children with cerebral palsy. Eur. J. Appl. Physiol. 54: 67–70, 1985.
7. Damiano, D. L., and M. F. Abel. Relation of gait analysis to gross motor function in cerebral palsy. Dev. Med. Child Neurol. 38: 389–396, 1996.
8. Frost, G., O. Bar-Or, J. Dowling, and C. White. Habituation of children to treadmill walking and running: metabolic and kinematic criteria. Pediatr. Exerc. Sci. 7: 162–175, 1995.
9. Jeng, S.-F., K. G. Holt, L. Fetters, and C. Certo. Self-optimization in nondisabled children and children with cerebral palsy. J. Mot. Behav. 28: 15–27, 1996.
10. Katz, K., R. Fogelman, J. Attias, E. Baron, and M. Soudry. Anxiety reaction in children during removal of their plaster cast with a saw. J. Bone Joint Surg. Br. 83: 388–390, 2001.
11. Keefer, D. J., K. Apperson, S. Mcgreal, W. Tseh, J. L. Caputo, and D. W. Morgan. Within-day stability of walking oxygen uptake in children with cerebral palsy. Med. Sci. Sports 34: S291, 2002.
12. Leonard, C. T., H. Hirschfeld, and H. Forssberg. The development of independent walking in children with cerebral palsy. Dev. Med. Child Neurol. 33: 567–577, 1991.
13. Lumley, M. A., B. G. Melamed, and L. A. Abeles. Predicting children’s presurgical anxiety and subsequent behavior changes. J. Pediatr. Psychol. 18: 481–497, 1993.
14. L usk, G. The Elements of the Science of Nutrition.
Philadelphia: WB Saunders, 1928, pp. 61–74.
15. Maltais, D., O. Bar-Or, V. Galea, and M. Pierrynowski. Use of orthoses lowers the O2
cost of walking in children with spastic cerebral palsy. Med. Sci. Sports Exerc. 33: 320–325, 2001.
16. Matsudo, S. M., and K. R. Matsudo. Physician assessment of sexual maturation in Brazilian boys and girls: concordance and reproducibility. Am. J. Hum. Biol. 6: 451–455, 1994.
17. Mcdowell, B. C., V. Hewitt, A. Nurse, T. Weston, and R. Baker. The variability of goniometric measurements in ambulatory children with spastic cerebral palsy. Gait Posture 12: 114–121, 2000.
18. Minear, W. L. A classification of cerebral palsy. Pediatrics 18: 841–841, 1956.
19. Mutch, L., E. Alberman, B. Hagberg, K. Kodama, and M. V. Perat. Cerebral palsy epidemiology: where are we now and where are we going? Dev. Med. Child Neurol. 34: 547–551, 1992.
20. Palisano, R., P. Rosenbaum, S. Walter, D. Russell, E. Wood, and B. Galuppi. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev. Med. Child Neurol. 39: 214–223, 1997.
21. Rose, J., J. G. Gamble, J. Medeiros, A. Burgos, and W. L. Haskell. Energy cost of walking in normal children and in those with cerebral palsy: comparison of heart rate and oxygen uptake. J. Pediatr. Orthop. 9: 276–279, 1989.
22. Russell, D., P. Rosenbaum, C. Gowland, et al. Gross Motor Function Measure Manual.
Hamilton, Canada: Neurodevelopmental Clinical Research Unit, McMaster University, 1993, pp. 1–112.
23. Stanley, F., E. Blair, and E. Alberman. Cerebral Palsies: Epidemiology and Causal Pathways. London: MacKeith Press, 2000, pp.14–39.
24. Suputtitada, A. Managing spasticity in pediatric cerebral palsy using a very low dose of botulinum toxin type A: preliminary report. Am. J. Phys. Med. Rehabil. 79: 320–326, 2000.
25. Sutherland, D. H., R. Olshen, L. Cooper, and S. L. Woo. The development of mature gait. J. Bone Joint Surg. (Am.) 62: 336–353, 1980.
26. Tanner, J. M. Growth in Adolescence. Oxford: Blackwell Scientific, 1962, pp. 32–37.
27. Tseh, W., J. L. Caputo, I. S. Craig, D. J. Keefer, P. E. Martin, and D. W. Morgan. Metabolic accommodation of young children to treadmill walking. Gait Posture 12: 139–142, 2000.
28. Unnithan, V., J. Dowling, G. Frost, and O. Bar-Or. Role of mechanical power estimates in the O2
cost of walking in children with cerebral palsy. Med. Sci. Sports Exerc. 31: 1703–1706, 1999.
29. Unnithan, V. B., J. J. Dowling, G. Frost, and O. Bar-Or. Role of cocontraction in the O2
cost of walking in children with cerebral palsy. Med. Sci. Sports Exerc. 28: 1498–1504, 1996.
30. Unnithan, V. B., L. A. Murray, J. A. Timmons, D. Buchanan, and J. Y. Paton. Reproducibility of cardiorespiratory measurements during submaximal and maximal running in children. Br. J. Sports Med. 29: 66–71, 1995.
Keywords:©2003The American College of Sports Medicine
RELIABILITY; ECONOMY; LOCOMOTION; OXYGEN UPTAKE