Cerebral palsy (CP) is a term for a group of disorders characterized by nonprogressive damage to the developing brain, resulting in impaired motor function1,2 due to abnormal control of the central nervous system over skeletal musculature.3 This problem manifests in part as spasticity, which, in the lower leg, typically causes an equinus gait pattern. A decreased active dorsiflexion (DF), range of motion (ROM), and a lack of heel strike at initial contact are often the result.2,4
In addition to spasticity, the leg musculature of children with CP is also affected by marked weakness.5 The ankle dorsiflexors and plantarflexors as well as the hip extensors are affected5 and can benefit from strengthening.6–8 Although initial studies advocated the strengthening of nonspastic muscles only, for fear of exacerbating existing spasticity,9 more recent evidence suggests that strengthening these muscles may result in significant functional improvements.7,8,10 Both weakness and spasticity contribute to the higher energy expenditure during gait in children with CP compared with age-matched controls.11 This often leads to decreased levels of participation and the development of secondary complications due to a more sedentary lifestyle.12
One intervention gaining popularity in addressing weakness or loss of force generating ability around the ankle in CP is functional electrical stimulation (FES), which is characterized by the stimulation of intact peripheral nerves to activate their target muscles in a functional manner.13 FES can be distinguished from neuromuscular electrical stimulation (NMES) in that NMES is aimed at muscle strengthening and/or spasticity reduction and is not always applied functionally.14 NMES is applied at a higher intensity than FES, and the timing of the stimulation is set according to muscle fatigue.14 In comparison, FES is aimed at augmenting function, and thus only requires intensity settings allowing for the required action to take place.
By applying NMES to the agonist, muscle strengthening is brought about by increasing motor unit recruitment as well as by increasing contractile proteins, with resultant muscle hypertrophy.15 In addition, electrical stimulation applied to the antagonist can cause a decrease in tone in the agonist16 through reciprocal inhibition.17 These differing concepts suggest that stimulation of both the agonist and antagonist muscle groups can be effective.
Recent studies of FES in CP have targeted various muscle groups, including the tibialis anterior18–22 and triceps surae muscles18–20,22 as well as the gluteals,18 quadriceps18 and hamstring muscles.19 These studies used a wide spectrum of outcome measures and reported varying findings regarding muscle strength and spatiotemporal gait parameter improvements, functional gains, and carry-over effects.
Stimulation can be administered by either surface (S-FES) or percutaneous (P-FES) stimulation. With S-FES, electrodes are placed on the skin, whereas they are implanted into the target muscle for P-FES.23 In a comparison study, P-FES was perceived to be significantly more comfortable for patients24 and resulted in more marked increases in active ankle DF angles during gait,22 possibly because P-FES bypasses cutaneous receptors, allowing for a stronger muscle contraction. P-FES, however, is a costly procedure24 and may thus not be readily available to clinicians and patients. For that reason, S-FES remains the preferred mode of stimulation in the clinical setting.
Current available evidence on electrical stimulation highlights differing theories regarding the choice of target muscle(s) in the presence of spasticity.15–17 The objective of this review was, thus, to systematically assess and synthesize the existing evidence regarding the use of FES in children and adolescents with CP when applied to the lower limb. Key aspects included the assessment of the effect of FES on kinematic and spatiotemporal gait parameters as well as activities of daily living. The stimulation protocols and target muscles used for this purpose were also examined.
The following inclusion and exclusion criteria were applied to select studies for this review.
Intervention studies published in English were considered if they reported on primary data. All publications from the date of inception of the relevant database until December 2006 were included. Within each study, the stimulation had to have been applied to any lower leg muscle(s) during a functional activity by either S-FES or P-FES as a treatment program, and subjects had to have been younger the age of 18 years with a diagnosis of CP. As the reviewers wanted to assess the evidence on carry-over effects into a functional setting, the stimulation device was required to be removed or switched off at the time of testing.
Articles were excluded if the full-text articles were not locally obtainable, or if they scored below 8 of 15 during critical appraisal. Studies were also excluded if subjects presented with any additional neurologic conditions or a combination of movement disorders (eg spasticity with ataxia), or if they had undergone orthopedic surgery to the affected leg within 6 months before the start of the intervention.
The primary reviewer conducted a comprehensive search, repeated 4 times between October and December 2006 (Fig. 1). Internet databases accessible via a local university library service—CINAHL, Pubmed, Journals@ Ovid, ScienceDirect: Medicine and Dentistry and Neurosciences, PEDro, ProQuest: Medical Library and Science Journals, PsycInfo—were searched. A wide variety of key words were used to accommodate for the different indexing terms of each database (Appendix). MeSH terminology and truncation was used as appropriate for each individual database. A hand search was also conducted by perusing the reference lists of the Poster and Platform Presentations for the 2004 and 2005 Combined Sections Meeting of Pediatric Physical Therapy, as well as the references cited during a national neurodevelopmental theory association congress held in 2005. In addition, searching the reference lists of all relevant articles was done to locate additional literature.
Quality of Evidence
Each study was critically appraised by 2 independent reviewers using the Critical Review Form for Quantitative Studies developed by the McMaster University Occupational Therapy Evidence-Based Practice Research Group.25 It contains 15 questions of equal weight, all to be answered by yes (allocated 1 point), no (allocated 0 points), or not discussed (allocated 0 points). The secondary reviewer was blinded to each article’s author and publication source, to prevent bias in scoring.
Information from all articles included in the review were summarized on an Excel spreadsheet including authors, date, research design, sample size, outcome measures and measurement instruments used, stimulated muscles, critical appraisal score, and level of evidence26 (Table 1). The reviewers intended to conduct a meta-analysis of collated data if at least 3 studies were comparable in terms of patient demographics, interventions, and use of outcome measures. As this was not possible, outcomes are discussed narratively.
Primary and secondary searching was conducted and the researchers are confident that the majority of evidence regarding the use of FES in the CP population was located, as 62% of articles found were duplicated at least once.
Description of Studies
There was full agreement between the 2 reviewers regarding the final inclusion of articles for the review, and final scores were awarded by discussion. Studies were excluded due to subjects being older than 18 years, stimulation not being applied functionally, stimulation not being used as a treatment program, stimulation not being removed or switched off at the time of outcome testing, and articles not locally obtainable.
Of the studies included in this review, 3 articles were case reports,18,19,22 whereas Durham et al21 made use of a single-subject research design and Comeaux et al20 used a crossover design (Table 1). In terms of Sackett’s hierarchy of evidence,26 2 articles ranked as evidence level 2C,20,21 whereas the other 3 studies ranked as levels 4.18,19,22
Although it was initially decided that articles scoring below 8 of 15 (ie below 50%) would be excluded from the review, the low number of articles scoring above this cutoff point necessitated the inclusion of 1 article scoring 7 of 15.19 The average score was 8.8 of 15 (SD = 1.9), ranging from 7 of 1519 to 12 of 1520 (Table 1). All studies clearly stated the purpose of their respective research and conducted a thorough review of the relevant literature. Although 3 of the 5 articles were case studies, the reviewers deemed this design appropriate, owing to the variability of the clinical presentation of CP.
None of the studies reported on the reliability of their outcome measures, and only Comeaux et al20 referred to the validity of their outcome measures. Carmick19 did not describe the intervention in sufficient detail, and no researchers reported on potential contamination of results by concomitant therapeutic treatment. Four articles18–21 reported on the clinical and/or statistical significance of their results, although Bertoti et al18 did not implement appropriate analysis methods to support their statements, and the link between the results and subsequent conclusions was unclear in a study by Carmick.19
The authors of one study19 did not describe the sample in sufficient detail. None of the reviewed studies provided information on where the subjects had been recruited from or motivated their choice of sample size. The average age of the subjects was 8.0 years (range: 1.6–15 years). Subjects with hemiplegia accounted for 53.9% of the study participants, 41.2% presented with diplegia, and subjects with quadriplegia and ataxia each accounted for 2.9% of the sample.
Different studies applied S-FES to the tibialis anterior,19,21 gastrocnemius,19,20 tibialis anterior and gastrocnemius alternatingly,19,20,22 and hamstring muscles.19 P-FES was applied to the gluteals and quadriceps,18 as well as the gastrocnemius18,22 and the tibialis anterior18,22 muscles (Table 1).
Comeaux et al20 included a 4-week withdrawal phase in their research design, whereas Durham et al21 allowed 12 weeks after completion of the intervention before final reassessment. Carmick19 commented on the carry-over effect at differing periods of time during the respective case studies (ranging from 2 days to 1 year postintervention). Pierce et al22 and Bertoti et al18 did not include a withdrawal phase in the research design, and assessed the dependent variables after each of the stimulation phases.
Stimulation Dosage and Parameters.
The stimulation parameters and exposure length to the intervention used in the reviewed studies varied between studies, ranging from 1 week22 to approximately 9 months19 (Table 2). Two of the reviewed studies made use of preset on:off stimulation times,19,20 whereas stimulation emission in the other 3 studies was triggered by ipsilateral footswitches21,22 or remote control.19
A wide variety of outcome measures were used to assess the efficacy of FES on impairment, functional, and participation levels (Table 3). However, in keeping with the aim of this review, only the most frequently used outcome measures relating to gait will be discussed at this point. Most studies reported on functional and/or participation changes in an anecdotal manner, which renders the discussion of these outcomes difficult in a group context.
Changes in Functional Levels
When the tibialis anterior and the gastrocnemius muscle were stimulated reciprocally, an increase in walking speed was recorded after the intervention (increase of 3.9 m/min and 3.5 m/min in 2 subjects)19 while Bertoti et al18 showed no change in walking speed after 28 and 40 weeks of treatment in 2 subjects. When the tibialis anterior muscle was stimulated alone,21 an increase of 0.6 m/min (normalized for height) was recorded by Durham et al.21 Pierce et al22 applied stimulation for 1 week per target muscle group and examined the immediate effect of the stimulation on walking speed. When the tibialis anterior and gastrocnemius muscles were simulated reciprocally, Pierce et al22 recorded slower walking speeds with the stimulating device on (with an average difference of 2.7 m/min) compared with no stimulation at all. When the tibialis anterior muscle was stimulated alone, subjects showed a slightly faster walking speed with the device switched on compared to with the device switched off (difference of 3.4 m/min). Pierce et al22 reported the only study which stimulated the triceps surae alone, and recorded a 1.12 m/min faster walking speed in 1 subject with the stimulation device switched on, whereas another subject showed a slower speed with the device switched on (1.8 m/min difference).
Step Length and Cadence.
This outcome measure was investigated in 4 of the 5 studies.18,19,21,22 Durham et al21 found no difference in step length in the affected leg after 3 months of tibialis anterior muscle stimulation, which contrasted with the 7.7 cm increase in step length recorded in Bertoti et al.18 When the tibialis anterior and gastrocnemius muscles were stimulated reciprocally, a decrease of 3.9 cm was recorded in 1 subject, compared with a 1.0 cm increase in another.19 Only Bertoti et al18 evaluated cadence, and found that a decrease in cadence accompanied a greater step length to produce an unchanged walking speed in two 6-year-old subjects. A lack of raw data prevented the reviewer from conducting cadence calculations from the study of Pierce et al,22 whereas both walking speed and step length (and thus cadence) remained unchanged in the study by Durham et al.21
Changes in Activities of Daily Living and Participation
Functional improvements were only reported anecdotally and ranged from a longer distance that a subject was able to gallop19 to increased participation in age-appropriate sporting activities, a decreased frequency of falling,18,19 and less need for external assistance during walking.18,19
The findings of this review seem to suggest that FES applied to the lower limb in children and adolescents with CP has a positive effect on gait and function. However, given the wide variety of intervention protocols and outcome measures used in the reviewed articles, analysis of the outcomes was difficult.
Three of the 5 articles included in the review were case reports. Validity and reliability of outcome measures used were poorly reported, and study samples were generally not well described. The average score of 8.8 of 15 points reflects a general lack of structured research in the CP population in this field. As the CP population is diverse and physical presentations vary greatly among individuals,27 it is difficult to conduct research with a high measure of external validity. If sampling were to conform to a strict set of criteria, the generalizeability of the outcomes would be low. Consequently, single-subject research designs28 or within-subject designs are recommended as more appropriate for this population. The advantage of the within-subject design is that the same subjects can be used repeatedly, with a washout period between intervention phases.29 This design allows the use of fewer subjects, and has a higher statistical power than if subjects are only used once.29 Single-subjects or (preferably) within-subjects study designs are recommended for future research to facilitate clinical decision making regarding the choice and timing of interventions for specific clinical presentations, especially in a heterogeneous population such as CP.
According to Bleck2 and Campbell,30 early physiotherapy intervention delivers the most promising outcomes, as movement patterns are not yet as fixed and secondary deformities not as prevalent. The average age of subjects in this review was 8.0 years, which falls into the age bracket deemed suitable for early intervention.2,30 The present review, however, revealed mixed outcomes in terms of age. A 6.7-year-old subject in Carmick’s19 study showed an increase of 13.9 m/min after receiving stimulation therapy, whereas a subject of similar age showed no change in terms of walking speed in another study.18 Subjects of around 10 years of age showed lesser improvements in terms of walking speed.19,22 The greatest effect size was recorded in the 6.7-year-old subject,19 which seems to suggest that younger subjects respond more favorably to stimulation treatment. However, the lengths of the stimulation protocols varied significantly across these studies, and could account for different treatment effects.
Younger subjects tended to present with an increased step length after intervention,18 whereas older subjects21 showed no change in this outcome variable. As children with diplegic CP generally rely on increases in cadence to increase their walking speed with a less pronounced increase in step length,31 data from at least 2 of these parameters are necessary to deduce clinical efficacy of FES. However, changes in walking speed, cadence, or step length after an intervention needs to be quantified in terms of the concomitant changes in the other variables to provide clinically relevant data. In addition, improvements in all or some of these variables may occur at the cost of energy economy during gait.31 Reporting on changes in spatiotemporal gait parameters in conjunction with gait economy measures is important, as improvements in some of these gait parameters may lose their clinical relevance when viewed in conjunction with energy economy during gait.
Although studies stimulating the same target muscle(s) differed from each other in terms of stimulation parameters and subject demographics, grouping the results according to target muscle allows some interpretation of optimal treatment approaches. Stimulation of the tibialis anterior muscle seems to have limited success21 and carry-over effect,19 or could even result in a deterioration of the gait pattern.19 A higher incidence of heel strike (ie larger DF angles at initial contact)19,20,22 with a greater carry-over effect was evident when the tibialis anterior muscle was stimulated alternately with the plantarflexors. Similar improvements were also found when the triceps surae muscle group was stimulated alone.19 Increased active DF ROM, a higher incidence of heel strike at initial contact, and greater functional independence maintained for 1 year after the end of the intervention program have been reported with this application protocol.19 Future research specifically regarding different stimulation parameters is needed to allow conclusive deductions on optimal treatment protocols.
Although the main aim of FES is to facilitate function rather than to increase individual muscle strength, strengthening may have occurred due to repetitive activity in muscles that were previously inactive or weak. FES may, thus, contribute to some strengthening and result in subsequent increases in ROM as secondary effects, thereby contributing to the carry-over effects reported in some of the reviewed studies.19–21 Wiley and Damiano5 have shown that both the dorsiflexors and plantarflexors are weak in CP, and thus both need strengthening. Targeting only 1 of these muscles can, thus, create a muscle imbalance, which could account for some of the deterioration in gait reported by Carmick.19 Furthermore, the dorsiflexors are stronger than their antagonists in CP.5,32 Stimulation of this muscle group alone would thus further strengthen the “stronger” of the 2 muscles without addressing the weak plantarflexors.
In comparison, stimulating the plantarflexors alone showed positive outcomes.19 As the triceps surae complex plays a vital role in stabilizing the knee and ankle during gait as well as decelerating anterior movement of the tibia on the foot during stance,33 stimulation of this muscle group alone could account for the positive intervention results described above.19 Olney et al34 have also shown that the positive work done by the plantarflexors during gait in CP is a third less than in normal subjects, and positive intervention outcomes with triceps surae muscle stimulation alone could, thus, also be attributed to improved work output by this muscle. Only 1 study targeted the hamstring muscle alone,19 which immediately resulted in improved gait as well as movement patterns during galloping, maintained for at least 4 weeks after the intervention. More rigorous research investigating proximal muscle strength needs to be conducted to support these findings.
In contrast to what has previously been a hesitancy to stimulate spastic muscles, the functional improvements reported on the reviewed studies seem to suggest that FES does not increase spasticity. Although FES is not aimed at improving muscle strength, it may contribute to force production due to the overload principle, a point made by Bührs.35 As most of the results discussed above were obtained during case studies18,19,22 and thus have low levels of evidence, extrapolations to the greater CP population should not be made without further research.
Comeaux et al20 and Bertoti et al18 applied electrical stimulation during gait and gait subtasks by means of FES and NMES, respectively. However, the outcomes used in these studies were not assessed separately after FES and NMES, but were assessed at the end of the entire stimulation program. It is, thus, not possible to attribute the effects of the intervention to either NMES or FES, as both were used. Future studies should, thus, clearly distinguish between FES and NMES and base their choice of intervention on the required aim of the intervention functional improvement versus muscle strengthening versus spasticity reduction. Outcomes should also be assessed separately for these 2 intervention modes.
As a guideline, frequencies of around 30 Hz are used during FES to elicit a smooth contraction,36 as a balance between muscle fatigue and an effectively maintained contraction occurs in this range. Similar settings were used in the reviewed studies (Table 2). De Vahl36 further states that most stimulation devices accommodate a pulse width range of 0.2 to 0.4 μsec, and the settings of all reviewed studies fall into this range. It is interesting to note that both Bertoti et al18 and Pierce et al22 made use of an unusually low amplitude of 20 mA. The higher the amplitude, the more nerve fibers are recruited, and thus a more effective muscle contraction is achieved.14 Although these researchers reported similar functional18 and kinematic18,22 improvements as other studies using S-FES, they implemented P-FES, which could account for comparable treatment effects to those of other studies, while using lower intensity settings.
Although an extensive number of outcome measures were implemented in the reviewed studies, they were mostly aimed at assessing impairment and function (Table 3). Improvements in activities of daily living were only assessed anecdotally, which rendered structural analysis of data impossible. As impairment-orientated improvements cannot be assumed to effect functional gains,37 future investigations should incorporate participation-targeted outcome measures.38
CONCLUSION AND IMPLICATIONS FOR RESEARCH AND CLINICAL PRACTICE
This review has revealed that a wide variety of FES application methods and stimulation protocols exist. Although treatment results were generally not assessed in a standardized manner, available evidence seems to suggest that FES application to either the triceps surae muscle alone or in conjunction with the tibialis anterior muscle is favorable. Because of the lack of evidence regarding the effects of proximal muscle stimulation, further studies need to be conducted to substantiate available data. Most studies offered a low level of evidence and also did not clearly distinguish FES from NMES. Future research needs to ensure that the mode of application (FES or NMES) coincides with the aims of the intervention to provide results attributable to that intervention. Additionally, the majority of outcome measures in the reviewed studies were impairment focused. Future studies, thus, need to strive toward assessing the effect of intervention on activity and participation.
1. Rosenbaum P, Paneth N, Goldstein M, et al. A report: the definition and classification of cerebral palsy, April 2006. Dev Med Child Neurol
2. Bleck EE. Orthopaedic Management in Cerebral Palsy
. Oxford: Blackwell Scientific; 1987.
3. Skinner SR. Direct measurement of spasticity. In: Sussman MD, ed. The Diplegic Child
. Rosemont, IL: American Academy of Orthopedic Surgeons; 1992:31–44.
4. Gage JR. Gait Analysis in Cerebral Palsy
. New York, NY: Cambridge University Press; 1991.
5. Wiley ME, Damiano DL. Lower-extremity strength profiles in spastic cerebral palsy. Dev Med Child Neurol
6. Damiano DL, Vaughan CL, Abel MF. Muscle response to heavy resistance exercise in children with spastic cerebral palsy. Dev Med Child Neurol
7. Dodd JD, Taylor NF, Graham HK. A randomized clinical trial of strength training in young people with cerebral palsy. Dev Med Child Neurol
8. Unger M, Faure M, Frieg A. Strength training in adolescent learners with cerebral palsy: a randomized controlled trial. Clin Rehabil
9. Alfieri V. Electrical treatment of spasticity. Scand J Rehabil
10. Dodd KJ, Taylor NF, Damiano DL. A systematic review of the effectiveness of strength-training programs for people with cerebral palsy. Arch Phys Med Rehabil
11. Rose J, Gamble GJ, Burgos A, et al. Energy expenditure index of walking for normal children and for children with cerebral palsy. Dev Med Child Neurol
12. Brunstrom JE. Clinical considerations in cerebral palsy and spasticity. J Child Neurol
14. Bührs D. FES. E-mail to A. Seifart (online). 16 September 2007. Available email: email@example.com.
15. Reed B. The physiology of neuromuscular electrical stimulation. Pediatr Phys Ther
16. Liberson WT. Experiment concerning reciprocal inhibition of antagonists elicited by electrical stimulation of agonists in a normal individual. Am J Phys Med
17. Apkarian JA, Naumann S. Stretch reflex inhibition using electrical stimulation in normal subjects and subjects with spasticity. J Biomed Eng
18. Bertoti DB, Stanger M, Betz RR, et al. Percutaneous intramuscular functional electrical stimulation as an intervention choice for children with cerebral palsy. Pediatr Phys Ther
19. Carmick J. Clinical use of neuromuscular electrical stimulation for children with cerebral palsy, part 1: lower extremity. Phys Ther
20. Comeaux PA, Patterson ND, Rubin MO, et al. Effect of neuromuscular electrical stimulation during gait in children with cerebral palsy. Pediatr Phys Ther
21. Durham S, Eve L, Stevens C, et al. Effect of functional electrical stimulation on asymmetries in gait of children with hemiplegic cerebral palsy. Physiotherapy
22. Pierce SR, Laughton CA, Smith BT, et al. Direct effect of percutaneous electrical stimulation during gait in children with hemiplegic cerebral palsy: a report of 2 cases. Arch Phys Med Rehabil
23. Orlin MN, Pierce SR, Stackhouse BT, et al. Immediate effect of percutaneous intramuscular stimulation during gait in children with cerebral palsy: a feasibility study. Dev Med Child Neurol
24. Chae J, Hart R. Comparison of discomfort associated with surface and percutaneous intramuscular electrical stimulation for persons with chronic hemiplegia. Am J Phys Med Rehabil
25. Law M, Stewart D, Pollock N, et al. Critical review form—quantitative studies. McMaster University Occupational Therapy Evidence-Based Practice Research Group. 1998.
26. Sackett DL. Evidence based medicine: what it is and what it isn’t. BMJ
27. Stanger M, Oresic S. Rehabilitation approaches for children with cerebral palsy: overview. J Child Neurol
28. Martin JE, Epstein LH. Evaluating treatment effectiveness in cerebral palsy. Phys Ther
29. Schuster DP, Powers WJ. Experimental study designs. In: Schuster DP, Powers WJ, eds. Translational and Experimental Clinical Research
. Baltimore, MD: Lippincott William and Wilkins; 2005:88–90.
30. Campbell SK. Quantifying the effects of intervention for movement disorders resulting from cerebral palsy. J Child Neurol
31. Abel MF, Damiano DL. Strategies for increasing walking speed in diplegic cerebral palsy. J Pediatr Orthop
32. Dietz V, Berger W, Quintern J. Electrophysiological studies of gait in spasticity and rigidity: evidence that altered mechanical properties of muscle contribute to hypertonia. Brain
33. Sutherland DH, Cooper L, Daniel D. The role of the ankle plantar flexors in normal walking. J Bone Joint Surg Am
. 1980;62:354– 363.
34. Olney SJ, MacPhail HEA, Hedden DM, et al. Work and power in hemiplegic cerebral palsy gait. Phys Ther
. 1990;70:431– 438.
35. Bührs D. Functional electrical stimulation: the use of functional electrical stimulation in rehabilitation. Course notes (Professional FES South Africa); 2007.
36. De Vahl J. Neuromuscular electrical stimulation (NMES) in rehabilitation. In: Gersh MR, ed. Electrotherapy in Rehabilitation
. Philadelphia: Davies; 1992:218–268.
37. Van der Linden ML, Hazlewood ME, Aitchinson AM, et al. Electrical stimulation of gluteus maximus in children with cerebral palsy: effects on gait characteristics and muscle strength. Dev Med Child Neurol
38. World Health Organization. International Classification of Functioning, Disability, and Health. Short Version. Geneva. 2001.
Search Strategies Implemented
Strategy 1: cerebral palsy AND electrical stimulation OR electric stimulation OR electric stimulation therapy OR surface stimulation OR percutaneous stimulation.
Strategy 2: strategy 1 AND gait OR lower limb OR leg OR gastrocnemius OR tibialis anterior OR triceps surae OR gluteus OR hamstrings OR quadriceps OR energy economy OR energy expenditure. Cited Here...
© 2009 Lippincott Williams & Wilkins, Inc.