Perry Issue: Gait Rehab

Can Strength Training Predictably Improve Gait Kinematics? A Pilot Study on the Effects of Hip and Knee Extensor Strengthening on Lower- Extremity Alignment in Cerebral Palsy Diane L. Damiano, Allison S. Arnold, Katherine M. Steele, Scott L. Delp

D.L. Damiano, PT, PhD, is Chief, Computer simulations have demonstrated that excessive hip and Functional and Applied Biome- Background. chanics Section, Rehabilitation knee flexion during gait, as frequently seen in ambulatory children with cerebral Medicine Department, NIH Clini- palsy (CP), can reduce the ability of muscles to provide antigravity support and cal Center, Bldg 10, Room increase the tendency of hip muscles to internally rotate the thigh. These findings 1–1469, Bethesda, MD 20892 suggest that therapies for improving upright posture during gait also may reduce (USA). Address all correspondence excessive internal rotation. to Dr Damiano at: damianod@ cc.nih.gov. Objective. The goal of this study was to determine whether strength training can A.S. Arnold, PhD, is Research As- diminish the degree of crouched, internally rotated gait in children with spastic sociate, Concord Field Station, Harvard University, Bedford, diplegic CP. Massachusetts. Design. This was a pilot prospective clinical trial. K.M. Steele, MS, is a doctoral stu- dent in the Department of Me- chanical Engineering, Stanford Methods. Eight children with CP participated in an 8-week progressive resistance University School of Engineering, exercise program, with 3-dimensional gait analysis and isokinetic testing performed Stanford, California. before and after the program. Secondary measures included passive range of motion, S.L. Delp, PhD, is Professor of Bio- the Ashworth Scale, and the PedsQL CP Module. To identify factors that may have engineering and Mechanical Engi- influenced outcome, individual and subgroup data were examined for patterns of neering, Stanford University change within and across variables. Schools of Engineering and Medicine. Results. Strength (force-generating capacity) increased significantly in the left hip [Damiano DL, Arnold AS, Steele extensors, with smaller, nonsignificant mean increases in the other 3 extensor muscle KM, Delp SL. Can strength train- groups, yet kinematic and functional outcomes were inconsistent. The first reported ing predictably improve gait kine- subject-specific computer simulations of crouch gait were created for one child who matics? A pilot study on the effects of hip and knee extensor strength- showed substantial benefit to examine the factors that may have contributed to this ening on lower-extremity align- outcome. ment in cerebral palsy. Phys Ther. 2010;90:269–279.] Limitations. The sample was small, with wide variability in outcomes. © 2010 American Physical Therapy Association Conclusions. Strength training may improve walking function and alignment in some patients for whom weakness is a major contributor to their gait deficits. However, in other patients, it may produce no change or even undesired outcomes. Given the variability of outcomes in this and other strengthening studies in CP, analytical approaches to determine the sources of variability are needed to better identify those individuals who are most likely to benefit from strengthening.

Post a Rapid Response or ﬁnd The Bottom Line: www.ptjournal.org

February 2010 Volume 90 Number 2 Physical Therapy f 269 Strength Training and Gait Kinematics

he long-term goal of this re- dren, including regular physical ther- tributing to excessive hip internal ro- search collaboration is to de- apy and bracing, as well as more tation.9 For example, the capacity of Ttermine whether targeted invasive methods of spasticity reduc- the gluteus maximus muscle to gen- strengthening of the hip and knee tion (eg, intramuscular injections, or- erate external rotation of the hip is extensors, alone or in combination thopedic surgery). Unfortunately, compromised when the hip is with other treatments, can effec- the outcomes of strength training flexed. This finding suggests that im- tively diminish the excessive knee and of many of the other interven- provement of excessive hip flexion flexion, hip flexion, and hip internal tions for improving walking in these in people with crouch gait might im- rotation commonly observed in chil- individuals remain inconsistent. prove hip rotation. dren with spastic diplegia who walk with a crouched gait pattern. Cere- There are several reasons to hypoth- Despite this rationale, the potential bral palsy (CP) is the most prevalent esize that strengthening the hip and for strength training to improve the physical disability originating in knee extensors might improve the gait mechanics of people with spas- childhood, with the largest propor- gait kinematics of children with spastic diplegia remains unclear. tion of this patient population having tic diplegia. Perry’s pioneering work Strength training was briefly a part of spastic diplegia,1 characterized by in- demonstrated electromyographic early physical therapy management volvement primarily in the lower ex- (EMG) activity of the hip and knee of CP and has experienced a resur- tremities. Nearly all children with extensors during the early stance gence over the past 10 to 15 years. spastic diplegia will ambulate, al- phase of normal gait,3 and computer- This resurgence was preceded by de- though often at a later age and with based simulations of walking dynam- cades during which resistance train- greater hip and knee flexion than ics have revealed that the gluteus ing was contraindicated for people children without neuromotor dis- maximus and vastus muscles play an with CP because of clinical concerns abilities.2 Much therapeutic effort is important role in early stance by sup- that it would exacerbate spasticity. directed at promoting and maintain- porting body weight and controlling Some studies have demonstrated ing upright ambulation in these chil- hip and knee extension.4 (For a re- that short-term progressive resis- view of principles associated with tance exercise programs can safely human locomotion, refer to Kuo and increase strength (force-generating Available With Donelan in this issue.5) Simulations capacity) in targeted muscles in peo- This Article at analyzing the roles of these muscles ple with CP without increasing spas- ptjournal.apta.org during crouch gait have revealed ticity.10,11 Positive effects on gait and that the capacity of gluteus maximus gross motor function also have been • eTable 1: Participants Ordered and other muscles to extend the hip reported,12 although less consis- by Decreasing Change in Knee and knee may be diminished sub- tently than gains in strength.13 A re- Extensor Torque stantially by hip and knee flexion.6 cent meta-analysis, however, failed • eTable 2: Mean Results and P The presence of weakness in chil- to provide evidence that strength Values for Selected Gait Measures dren with spastic diplegia is now training is effective in improving Before and After Strengthening well documented, with even the strength, increasing gait speed, or • eTable 3: Mean Changes in most functional individuals demon- producing a clinically meaningful PedsQL Cerebral Palsy Module strating substantial generalized mus- change in gross motor function in Child Questionnaire Scores Before cle weakness.7,8 Together, these ob- people with CP.14 The potential for and After Strengthening servations suggest that the crouched strength training to improve the gait • Video: In honor of Dr Jacquelin postures of some children with CP kinematics of people with CP has Perry, view art by patients from may be exacerbated by lower- been evaluated in only a few studies, Rancho Los Amigos National extremity extensor weakness. Mus- and findings have been equivocal. Rehabilitation Center. culoskeletal modeling and anatomi- Most of these studies averaged data • Podcast: “Stepping Forward With cal studies also have shown that over a small number of subjects (ie, Gait Rehabilitation” symposium excessive hip flexion alters the bal- 11–21 subjects) and revealed only recorded at APTA Combined ance of the muscles that rotate the modest mean changes in the subjects’ Sections Meeting, San Diego. hip. With excessive hip flexion, the joint angles during walking.15–19 • Audio Abstracts Podcast moment arms (ie, the lever arms or mechanical advantage) of some hip The variability in outcomes reported This article was published ahead of print on December 18, 2009, at external rotators are diminished and in previous studies may be due to a ptjournal.apta.org. the moment arms of hip internal ro- range of factors, including method- tators are increased, potentially con- ological factors that may have lim-

270 f Physical Therapy Volume 90 Number 2 February 2010 Strength Training and Gait Kinematics ited the effectiveness of the training tional Classification System (GMFCS) participants prior to the initial (eg, insufficient muscle loads, insuf- levels I to III. The children had to be assessment. ficient training durations) or, in more than 1 year postsurgery and 6 some cases, impairments other than months post–botulinum toxin injec- Eight children (3 male, 5 female) met weakness that may have limited the tions in the lower extremity. Addi- the inclusion criteria and completed subjects’ functional gains following tional inclusion criteria, based on the 8-week strengthening program strengthening (eg, difficulties with physical examination, included bilat- during the study period (eTab. 1 balance). We believe new ap- eral passive hip extension to neutral available at ptjournal.apta.org). The proaches are needed to rigorously with the other hip flexed to 90 de- children ranged in age from 5.5 to evaluate the biomechanical effects of grees, passive knee extension within 13.4 years. Five children were classi- resistance exercise programs, to 5 degrees of full extension while po- fied at GMFCS level III and used an help explain the reasons for incon- sitioned supine, and passive hip ex- assistive device to walk, 2 children sistent outcomes across subjects, ternal rotation of at least 20 degrees were classified at GMFCS level II, and and to identify individuals who are as assessed in a prone position with 1 child was classified at GMFCS level I. most likely to benefit from hip extended and knees flexed to 90 strengthening. degrees. These criteria were chosen Strength Training Program in an effort to exclude children Each child participated one-on-one This article presents the results from whose gait deviations were predom- with a therapist in a community- a pilot strength training study in inantly constrained by musculoskel- based physical therapy program, at- which we used a combination of etal contracture. Previous orthope- tending three 1-hour sessions per physical examination, gait analysis, dic surgery or neurosurgery were week for 8 weeks. Each therapist- and state-of-the-art computer simula- not considered a reason for exclu- guided session consisted of a pro- tion to evaluate the effects of an sion, with the exception of previous gressive resistance exercise program 8-week progressive resistance exer- rotational osteotomies. using free weights or weight ma- cise program on participants’ hip chines that targeted the gluteus and knee angles (kinematics) during Children were screened visually to maximus and quadriceps muscles bi- walking. We hypothesized that determine whether they exhibited a laterally, with at least 1 day of rest strengthening the hip and knee ex- crouched internal rotation gait pat- between sessions (examples shown tensor muscles would decrease the tern, and these kinematic criteria in Fig. 1). The program used a com- exaggerated hip and knee flexion were confirmed by gait analysis. In bination of open- and closed-chain and hip internal rotation of children particular, children had to satisfy the exercises for maximum transfer to with spastic diplegia during the following requirements bilaterally both the stance and swing phases of stance phase of gait. Secondary hy- prior to participating in the strength- gait, and these exercises were de- potheses were that strengthening ening portion of the study: (1) hip signed to work the muscles in the would improve temporal spatial gait and knee flexion greater than 1 stan- most extended portion of the range. parameters, such as walking speed dard deviation above the mean nor- Examples include the use of cuff and stride length, and lead to im- mative value at initial contact; (2) weights attached to the distal thigh proved physical functioning and re- excessive adduction and internal ro- during prone hip extension exer- lated quality of life. Our main objec- tation of the hip at mid stance; and cises with the knee kept flexed tive was to assess whether strength (3) less than 20 degrees of ankle dor- throughout the motion, the use of a training could improve lower- siflexion at mid stance, because ex- resisted leg press, and the use of a extremity alignment during walking cessive weakness of the plantar flex- weight machine that resisted knee and, if so, use simulation to provide ors or overlengthening of the calf extension in a reclined sitting biomechanical insights into the rea- muscles alone may contribute to position. sons for these improvements. crouch gait. The intensity and difficulty of the Method Participants were recruited from the program were adjusted individually. Participants neurology and neurosurgery cere- The amount of resistance applied The goal of this pilot study was to bral palsy clinics and the physical was based on the number of repeti- enroll 10 children with spastic diple- therapy clinic at St. Louis Children’s tions each child could perform be- gia, based on a power analysis using Hospital associated with Washington fore fatiguing; the target was 8 to 10 data from 2 previous studies,15,17 University. Informed consent and pe- repetitions, with the total number of within the age range of 5 to 17 years diatric assent were obtained from all repetitions consistent across partici- and diagnosed at Gross Motor Func- pants (30 repetitions per muscle

February 2010 Volume 90 Number 2 Physical Therapy f 271 Strength Training and Gait Kinematics

Figure 1. Two examples of hip extensor strengthening exercises performed during training using weight machines. Other exercises also were performed (see text). group). Passive stretching exercises cating normal tone [resistance to pilot nature of this study, no correc- at the hip and knee and a 5-minute passive stretch] and a score of 0 in- tion for multiple tests was applied. walk at a relaxed pace on a treadmill dicating less-than-normal tone). The or indoor track (depending on the 3-D gait analysis was performed to Subject-Specific Simulations of child’s ability and preference) were determine the child’s lower- Crouch Gait performed for warm-up before and extremity joint angles during walk- Additional data were collected on cool-down after the strengthening ing, with data from 5 trials collected children who were independent am- exercises. at both freely selected and “as fast as bulators with the aim of creating possible without running” speeds. subject-specific computer simula- Functional Assessment Fifteen retroreflective markers were tions of each child’s crouched gait. Each child received an assessment placed on the skin overlying specific To enable dynamic simulations, consisting of the following: (1) a anatomic locations on the pelvis and these data must include “clean” con- physical examination; (2) a 3- bilateral lower extremities. The 3-D secutive forceplate strikes during the dimensional (3-D) gait analysis; (3) locations of these markers were gait analysis. These data were avail- an isokinetic strength assessment, as tracked using an 8-camera Vicon sys- able, from both prestrengthening measured by maximum hip and knee tem,* and the data were processed and poststrengthening training as- extensor concentric torque at 30°/s; using Plug-In-Gait.† The isokinetic sessments, for one individual in this and (4) a validated self-report mea- strength assessment enabled com- sample. In this case, the gait assess- sure of physical functioning and re- parison of each child’s isokinetic ments included full-body 3-D kine- lated quality of life, as quantified by peak torque before and after matics, ground reaction forces and the parent-proxy version of the strengthening. The peak torque data moments from forceplates, and sur- PedsQL 3.0 Cerebral Palsy Module.20 were expressed as the actual value face EMG recordings from the tibialis The exercise program was started and also were divided by body anterior, gastrocnemius, rectus fem- within a week after the initial assess- weight and multiplied by 100 to fa- oris, medial hamstring, adductor lon- ment, and an identical assessment cilitate comparisons across children gus, vastus medialis, gluteus medius, was conducted within a week of of varying sizes. Paired t tests (PϽ.05) and gluteus maximus muscles. completing the program. The physi- were used to evaluate changes in se- cal examination was performed to lected assessment measures for the Computer simulations of the stance assess range of motion at the hip, children before and after strengthen- phase were generated that repro- knee, and ankle (to ensure that the ing. Due to the small sample and the duced this child’s 3-D gait kinematics child met the inclusion and exclu- and kinetics before and after strength- sion criteria) and to assess spasticity ening. This process provided one of in the hamstring and quadriceps the first subject-specific simulations of muscles. Spasticity was measured us- * Vicon, 7388 S Revere Pkwy, Suite 901, Cen- an individual with a crouched, internal tennial, CO 80112. ing the Ashworth Scale (ranging † Oxford Metrics, 14 Minns Business Park, rotation gait pattern and is the first from 0 to 4, with a score of 1 indi- West Way, Oxford OX2 0JB, United Kingdom. such model to evaluate changes in the

272 f Physical Therapy Volume 90 Number 2 February 2010 Strength Training and Gait Kinematics actions of muscles as a result of strength training. To generate the simulations, we ﬁrst created a computer model of the child that included 3-D representations of the bones, joints, and muscles scaled to the child’s an- thropometric dimensions.

Next, we used OpenSim biomechanics software21 to estimate the muscle activation patterns and muscle forces that, when applied to the model, produced joint angles and ground reaction forces that corre- sponded closely to the experimental measurements of the child’s gait kinematics and kinetics. This software uses an optimization algorithm, called Computed Muscle Control,22 to determine the excitation patterns for each of the 92 muscle compartments in the model. We used the child’s measured EMG data to verify that the simulated muscle excita- tions reﬂected the child’s activation patterns. We generated 2 simulations: one that reproduced the child’s gait dynamics before strengthening and another that reproduced the gait dynamics after strengthening.

Lastly, we analyzed the simulations to identify factors that enabled this individual to walk more upright after strengthening. In particular, we performed a perturbation analysis4 to assess the capacity of the gluteus maximus, vastus, and other muscles to support the body and extend the hip and knee during the stance phase. This analysis makes small adjust- ments (ie, perturbations) to the force in each muscle and measures the resulting changes in the hip and knee angles to determine the role that each muscle plays in extending or flexing the joints throughout the Figure 2. movement. Change in isokinetic peak torque divided by body weight and multiplied by 100 for each participant for the (A) right and left hip extensors and (B) right and left knee extensors. A positive change indicates an increase in strength. (C) Change in minimum knee flexion angle during stance. A positive change indicates an increase in knee flexion (ie, greater crouch).

February 2010 Volume 90 Number 2 Physical Therapy f 273 Strength Training and Gait Kinematics

Table. Mean (SD) Before and After Strengthening, Mean Difference and Standard Error (SE) of the Difference, 95% Conﬁdence Interval (CI) of the Difference, Percent Change, and P Values for Isokinetic Strength Tests (Nϭ8)a

Before After Strengthening Strengthening Mean 95% CI of the Percent Mean (SD) Mean (SD) Difference (SE) Difference Change P Peak concentric torque at 30°/s (ft-lb) Right hip extensors 12.7 (6.9) 17.2 (7.0) 4.5 (2.8) Ϫ2.21 to 11.0 35.4 .16 Left hip extensors 10.7 (4.9) 19.2 (8.6) 8.5 (2.5) 2.70 to 14.4 79.4 .01 Right knee extensors 13.9 (5.5) 16.4 (7.0) 2.6 (1.2) Ϫ0.20 to 5.34 18.0 .06 Left knee extensors 13.1 (6.3) 16.8 (9.3) 3.7 (1.6) Ϫ0.22 to 7.56 28.2 .06 a Values signiﬁcant at PϽ.05 shown in bold.

Results children responded less dramatically The 2 children who achieved the Effects of Strength Training on to the strengthening program, show- greatest gains in knee extensor Peak Isokinetic Torque ing either a negligible change or a strength (participant 1, classified at Most of the participants’ hip and small decrease in peak torque in one GMFCS level I, and participant 2, knee extensor muscles were stron- or both limbs and joints. classified at GMFCS level III) also ger after training, as measured by the showed the greatest improvements change in isokinetic peak torque. Effects of Strength Training on in knee extension during the stance The mean percentage gains in Gait Kinematics phase. However, for the other 6 chil- strength (ie, change in peak torque Some, but not all, of the children dren, knee flexion appeared to divided by initial peak torque value) walked with improved hip and knee worsen with increasing strength ranged from 18.0% for the right knee extension during stance following gains, so no general conclusions extensors to 79.4% for the left hip the exercise program (Fig. 2). The could be made (Fig. 3). extensors (Table); however, the per- changes in hip extension generally centage gains varied widely across were correlated to the changes in Effects of Strength Training on participants. The magnitude and knee extension (rϾ.75, PϽ.05). Sev- Gait Temporal-Spatial consistency of the strength changes eral of the children walked with im- Parameters generally were greater at the hip proved hip rotation; however, we Gait speed, stride length, and ca- than at the knee (eTab. 2 available at did not detect a correlation between dence were not significantly ptjournal.apta.org). Four children the changes in hip rotation and the changed, on average, in either the achieved appreciable increases in changes in either hip or knee self-selected or fast speed conditions peak torque in at least 3 of the 4 extension. following the strengthening program muscle groups (Fig. 2). The other 4 (eTab. 2 available at ptjournal. apta.org). However, changes in all 3 measures varied across subjects; for example, self-selected walking speed improved 18% for participant 7, but diminished 25% for participant 5. No significant correlations were found between change in strength and change in these temporal-spatial parameters.

Effects of Strength Training on Figure 3. Perceived Physical Function and Scatterplot of change in minimum knee ﬂexion angle (average for right and left legs) Related Quality of Life against (A) Ashworth Scale score and (B) change in knee extensor strength. Positive A secondary outcome measure was correlation between change in minimum knee ﬂexion angle during stance and Ash- the parent-proxy version of the worth Scale score of the hamstring muscles. Note that the Ashworth Scale score was not PedsQL 3.0 Cerebral Palsy Module. collected for one participant.

274 f Physical Therapy Volume 90 Number 2 February 2010 Strength Training and Gait Kinematics

Figure 4. Joint angles (with standard deviation bands), obtained from gait analysis, from the left side of one participant who responded well to the strengthening program (one for whom computer simulation was created). Note that the participant’s excessive knee flexion, hip flexion, and internal hip rotation were diminished after strengthening. Preϭbefore strengthening, Postϭafter strengthening, Flexϭflexion, Extϭextension, Dorsiϭdorsiflexion, Plantarϭplantar flexion.

These data showed trends toward a change of about 3 degrees. The pop- ity, the worse the response to small amount of improvement (eTab. liteal angle also worsened slightly, strengthening. 3 available at ptjournal.apta.org), but but the change was not significant none of the changes were statisti- on either side. No significant corre- Analysis of the Subject-Specific cally significant when averaged lation was found between changes in Simulations across the participants. passive and active knee extension. Participant 1 responded well to the Spasticity did not change signifi- strengthening program, demonstrat- Secondary Measures of cantly before and after the training, ing notable improvements in knee ex- Impairment but the degree of spasticity was di- tension, hip extension, and hip rota- Mean passive knee extension was rectly related to the change in knee tion throughout the gait cycle (Fig. 4). Ϫ2 degrees at the start of the pro- position (rϭ.55, PϽ.05). Figure 3 Analysis of the subject-specific gram. A small, but significant, de- shows the initial Ashworth Scale computer simulations and the associ- crease (increased tightness) in pas- score for each leg plotted against the ated clinical data provided insights sive knee extension was found as a change in minimum knee flexion, in- into why this child’s gait improved. result of the training (PϽ.05 for both dicating that the greater the spastic- right and left sides), with a mean

February 2010 Volume 90 Number 2 Physical Therapy f 275 Strength Training and Gait Kinematics

Figure 5. (Left) The 3-dimensional musculoskeletal model in poses corresponding to key events in the gait cycle. The musculoskeletal model used to create a simulation of one of the participants has 10 body segments and 92 muscle compartments. (Right) Capacity of the gluteus maximus and vastus muscles to accelerate the hip and knee into extension before (Pre) and after (Post) strengthening for the participant on whom the simulation was performed.

Prior to the strengthening program, without evidence of hip flexion. vastus muscles and increased the ca- participant 1 had lower than normal Thus, for this participant, an increase pacity of the gluteus maximus, vas- strength of the hip and knee exten- in strength was sufficient to improve tus, and other muscles to support sors, with the left side slightly his gait, and the resulting functional the body weight and to produce hip weaker than the right side, and he gains were not limited by other con- and knee extension. The decrease in walked in a moderate crouch gait founding impairments. hip flexion in stance also shifted the pattern. As reported in previous moment arms of the gluteus maxi- studies, this posture puts greater Our analysis of the child’s gait dy- mus, gluteus medius, and other hip than normal demands on the vastus namics supports this explanation. Af- muscles toward external rotation; muscles to support the body weight ter the strengthening program, he this may have contributed to this in- and diminishes the potential of the demonstrated a substantial increase dividual’s small improvement in hip gluteus maximus and vastus muscles (from 20% to 85% of peak isometric rotation. to extend the hip and knee.4,6,23 Al- torque) in the strength of his hip and though this child walked with exces- knee extensors, with the left extrem- Discussion sive knee flexion, analysis of the ity values now exceeding those of The goal of our pilot strengthening computer simulation revealed that the right extremity. The simulations program, guided by previous studies his hamstring muscles were operat- suggested that the force generated and insights from musculoskeletal ing at muscle-tendon lengths suffi- by the left gluteus maximus muscle models, was to improve the chil- cient for normal walking,24 so a sur- during early stance was increased, dren’s lower-extremity alignment gical lengthening of the hamstring restoring an important mechanism during walking. Unfortunately, the muscles was not needed to release that promotes hip and knee exten- program did not produce systematic excessively tight muscles. He had sion in normal gait. Furthermore, the decreases in excessive hip flexion or normal range of motion at the hip capacity for the gluteus maximus knee flexion during the stance and knee, with popliteal angles of and vastus muscles to accelerate the phase. Furthermore, because stance- 126 and 132 degrees on the right and hip and knee into extension was in- phase hip extension was improved left sides, respectively. He had un- creased (ie, angular acceleration per in only a few children, we were undergone a selective dorsal rhizotomy unit muscle force) (Fig. 5). Thus, af- able to test our hypothesis that im- at age 5 years and exhibited good ter strengthening, the child’s exces- proving hip extension would de- selective control, as demonstrated sive knee flexion was diminished, crease hip internal rotation. We by the ability to dorsiflex bilaterally which reduced the demands on the made an effort to exclude children

276 f Physical Therapy Volume 90 Number 2 February 2010 Strength Training and Gait Kinematics whose gait deviations may have been Previous studies have shown that re- ger and colleagues16 evaluated the caused primarily by other factors, sistance training, in some cases, can effects of an 8-week, generalized such as weak or over-lengthened produce positive effects on gross lower-extremity strengthening pro- plantar flexors, muscle contractures, motor functional abilities, including gram for 21 adolescents with CP and or bone deformities. Nonetheless, re- self-selected and maximum walking reported a significant improvement sponses to the strength training speed.10,13 In the present study, al- in their gait kinematics of 5.1 de- were variable, similar to previous though there was a nonsignificant grees at mid stance only by summing studies.15,16,18 The large variability in trend toward improvement on the up changes at the hip, knee, and an- response, combined with the small questionnaire that measured the chil- kle. Eek and colleagues18 trained the sample size, likely affected our abil- dren’s perception of physical perfor- 4 weakest muscles per subject in 16 ity to detect significant mean group mance and related quality of life, we independently ambulatory children responses to strengthening. Al- did not detect a systematic improve- with CP and observed improvements though the conclusions we can draw ment in walking speed or other gait in stride length and hip and ankle from analysis of only 8 participants parameters. This finding is not sur- kinetics, but no systematic change in are limited, examination of these re- prising, given the limited response to the gait kinematics. Lee and col- sults is informative for clinical prac- strength training for some of the chil- leagues19 conducted a generalized tice and for guiding future research. dren, as we suspect that improve- functional lower-extremity training ments in strength may be needed to program in 16 children, ages 4 to 12 Consistent with previous studies, we stimulate changes in these functional years, and found no changes in the showed that gains in hip and knee measures. gait kinematics. extensor strength are possible in ambulatory children with CP, although Only a few studies have evaluated Previous studies10,12,15–19 generally some children achieved larger gains, the effects of increased strength on have reported mean changes in gait and responded differently to those gait kinematics in people with CP, measures and have provided little in- gains, than others. It is difficult to and findings have been equivocal. sight into subject-specific factors explain why some children did not For example, Damiano and col- that may have influenced the func- make appreciable strength gains, leagues15 examined the effectiveness tional or kinematic outcomes. One given the fairly long (8-week) dura- of a 6-week isotonic quadriceps mus- potential precipitating factor often tion and the design of this program, cle strengthening program for im- mentioned clinically is growth. As whereby a therapist closely moni- proving crouch gait in 14 children children grow, their strength must tored each individual and progres- with spastic diplegia. Knee extensor keep pace, or else their posture is sively increased the amount of resis- strength improved significantly, as compromised, as often is apparent in tance used during training. Possible did stride length at free and fast normal adolescence, suggesting that explanations include neurological speeds. However, the only signifi- crouch in individuals with CP may factors, such as a primary agonist in- cant change in the gait kinematics be exacerbated by growth. How- sufficiency that was not amenable to was a more extended knee position, ever, Wren and colleagues25 recently training, or pre-existing muscle adap- on average, at initial contact. Six of evaluated the influence of age, tations that may have limited the ca- the children also achieved greater among other factors, on different pacity of some muscles to change in knee extension at mid stance, but 2 types of malalignment in nearly response to loading. At present, children developed a slightly wors- 1,000 patients with CP and found these explanations remain hypothet- ened crouch gait (ie, knee flexion that rotational malalignment was as- ical and warrant further investiga- increased approximately 5°) (unpub- sociated with advancing age, but in- tion. The children’s strength lished observation). terestingly, crouch did not emerge as changes were less consistent and being significantly affected by age. It smaller at the knee than at the hip, In a subsequent strength study, is likely that the natural history of which may be related to increases in Damiano and Abel17 targeted the 2 crouch is altered by surgical proce- hamstring muscle strength. Greater weakest muscles per individual and dures or botulinum toxin injections strength in the hamstring muscles found significant changes in walking on the hamstrings and other mus- would contribute to increased exten- speed and function in a sample of 11 cles, which may have affected their sion torque at the hip, but could be children with hemiplegia or diplegia, analyses. a potential source of greater antago- but no consistent improvement in nist restraint when producing exten- gait kinematics, which was not sur- Given the variability of responses ob- sion torque at the knee. prising given the variability in the served here, the reasons why some target muscles across subjects. Un- children in the present study walked

February 2010 Volume 90 Number 2 Physical Therapy f 277 Strength Training and Gait Kinematics with exaggerated knee flexion or 8-week strengthening program—in they hold promise for identifying the internal rotation following the pro- this case, targeting the hip abductors impairments that impede locomotor gram remain unclear. In the sub- and lateral rotators in 15 female run- function, thereby refining and im- group analysis comparing indepen- ners who were healthy—can alter proving treatment approaches in CP. dent ambulators with those who kinematics in a clinically significant used an assistive device, the children and predictable way.26 Based on our Dr Damiano was the principal investigator who were more functional showed a own observations and data from the and project director of the clinical trial. Dr small improvement in comparison literature, we remain cautiously op- Arnold and Dr Delp provided concept/idea/ with an appreciable worsening in timistic about the potential for research design and project management. the children who required an assis- strength training to improve walking Dr Arnold, Ms Steele, and Dr Delp provided tive device to walk; however, this in some patients. One hypothesis writing and data analysis. Ms Steele provided consultation (including review of difference did not reach significance, that emerges from these results is manuscript before submission). and the subgroup sample sizes were that strengthening may be more ben- small. We did find decreased passive eficial for children with milder in- The authors acknowledge the United Cere- bral Palsy Research & Education Foundation knee extension as a result of the pro- volvement, which seems reasonable for funding this project. This research also gram, as well as a relationship be- because those patients are likely to was supported, in part, by the intramural tween greater hamstring muscle have less spasticity and fewer motor program at the National Institutes of Health. spasticity and poorer kinematic out- control deficits that could limit their The authors thank Julie Anderson for her as- comes. We believe that unavoidable response to strengthening. It also is sistance in all aspects of the study and Tom Nuzum and Dave Reddy for conducting the strengthening of the hamstring mus- possible that programs need to be of training sessions for all participants. cles during the hip extensor training longer duration or coupled with may be responsible for the decrease other interventions to address all of Human subjects approval was obtained from the Washington University Human Studies in active and passive knee extension, the factors that potentially contrib- Committee prior to study initiation. despite the greater extensor ute to excessive hip flexion, knee strength, with those children with flexion, and hip internal rotation dur- This article was received February 23, 2009, and was accepted September 21, 2009. greater hamstring muscle spasticity ing walking. at greater risk. More stringent mus- DOI: 10.2522/ptj.20090062 cle length or spasticity criteria, as Conclusion well as closer monitoring of ham- The effect of strength training on References string muscle length during the inter- gait kinematics in people with CP 1 Hirtz D, Thurman DJ, Gwinn-Hardy K, vention, may be warranted in future remains unpredictable at the level of et al. How common are the “common” neurologic disorders? Neurology. 2007; studies. Interestingly, the child who the individual patient, and the cur- 68:326–337. had the most positive outcome was rent approach of mean group analy- 2 Sussman M, ed. The Diplegic Child Evalu- an independent ambulator and had sis of heterogeneous samples will ation and Management. Rosemont, IL: American Academy of Orthopaedic Sur- undergone a selective dorsal rhizot- not resolve this dilemma. We con- geons; 1992. omy several years prior to the study. tend that studies with larger samples 3 Perry J Determinants of muscle function in that can more adequately examine the spastic lower extremity. Clin Orthop Relat Res. 1993;(288):10–26. In their review of the role of strength patient and intervention factors that 4 Arnold AS, Anderson FC, Pandy MG, Delp training for improving gait in ambu- could influence outcome, large-scale SL. Muscular contributions to hip and latory children and adolescents with regression analyses, or subject- knee extension during the single limb stance phase of normal gait: a framework CP, Mockford and Caulton13 noted specific computer simulations of for investigating the causes of crouch gait. that the wide range of gait-related walking are needed to help explain J Biomech. 2005;38:2181–2189. outcomes within and across studies the variability in outcomes and to 5 Kuo AD, Donelan JM. Dynamic principles of gait and their clinical implications. Phys occurs because individuals have identify individuals who are most Ther. 2010;90:157–174. shown such a wide variety of abnor- likely to benefit from strengthening. 6 Hicks JL, Schwartz MH, Arnold AS, Delp malities and compensations; there- More research also is needed to de- SL. Crouched postures reduce the capacity of muscles to extend the hip and knee fore, they concluded that no general termine how strength training might during the single-limb stance phase of gait. conclusions can be drawn. Indeed, complement or interact with other J Biomech. 2008;41:960–967. the heterogeneity of the motor dis- treatments, especially those inter- 7 Damiano DL, Vaughan, CL, Abel MF. Mus- cle response to heavy resistance exercise order in people with CP makes it ventions that tend to further weaken in spastic cerebral palsy. Dev Med Child difficult to predict outcomes of any patients. Subject-specific musculo- Neurol. 1995;37:731–739. intervention in this population. One skeletal models and simulations of 8 Wiley ME, Damiano DL. Lower extremity strength profiles in spastic cerebral palsy. study in the adult orthopedic litera- walking are currently labor- and Dev Med Child Neurol. 1998;40:100–107. ture supports the premise that an technology-intensive efforts, but

278 f Physical Therapy Volume 90 Number 2 February 2010 Strength Training and Gait Kinematics

9 Delp SL, Hess WE, Hungerford DS, Jones 15 Damiano DL, Kelly LE, Vaughn CL. Effects 22 Thelen DG, Anderson FC, Delp SL. Gener- LC. Variation of rotation moment arms of quadriceps femoris muscle strengthen- ating dynamic simulations of movement with hip flexion. J Biomech. 1999;32: ing on crouch gait in children with spastic using computed muscle control. J Bio- 493–501. diplegia. Phys Ther. 1995;75:658–667. mech. 2003;36:321–328. 10 Andersson C, Grooten W, Hellsten M, 16 Unger M, Faure M, Frieg A. Strength train- 23 Liu MQ, Anderson FC, Pandy MG, Delp SL. et al. Adults with cerebral palsy: walking ing in adolescent learners with cerebral Muscles that support the body also modu- ability after progressive strength training. palsy. Clin Rehabil. 2006;20:469–477. late forward progression during walking. Dev Med Child Neurol. 2003;45:220–228. J Biomech. 2006;39:2623–2630. 17 Damiano DL, Abel MF. Functional out- 11 Fowler EG, Ho TW, Nwigwe AI, Dorey FJ. comes of strength training in spastic cere- 24 Arnold AS, Asakawa DJ, Delp SL. Do the The effect of quadriceps femoris muscle bral palsy. Arch Phys Med Rehabil. 1998; hamstrings and the adductors contribute strengthening exercises on spasticity in 79:119–125. to excessive internal rotation of the hip in children with cerebral palsy. Phys Ther. persons with cerebral palsy? Gait Posture. 18 Eek MN, Tranberg R, Zu¨gner R, et al. Mus- 2001;81:1215–1223. cle strength training to improve gait func- 2000;11:181–190. 12 Dodd KJ, Taylor NF, Damiano DL. A sys- tion in children with cerebral palsy. Dev 25 Wren TA, Rethlefsen S, Kay RM. Preva- tematic review of the effectiveness of Med Child Neurol. 2008;50:759–764. lence of specific gait abnormalities in chil- strength-training programs for people dren with cerebral palsy: influence of ce- Arch Phys Med Reha- 19 Lee JH, Sung IY, Yoo JY. Therapeutic ef- with cerebral palsy. fects of strengthening on gait function in rebral palsy subtype, age, and previous bil. 2002;83:1157–1164. Disabil Rehab surgery. J Pediatr Orthop. 2005;25: cerebral palsy. . 2008;30: 79–83. 13 Mockford M, Caulton JM. Systematic re- 1439–1444. view of progressive strength training in 26 Snyder KR, Earl JE, O’Connor KM, Eber- children and adolescents with cerebral 20 Varni JW, Burwinkle TM, Berrin SJ, et al. sole KT. Resistance training is accompa- palsy who are ambulatory. Pediatr Phys The PedsQL in pediatric cerebral palsy: nied by increases in hip strength and Ther. 2008;20:318–333. reliability, validity, and sensitivity of the changes in lower extremity biomechanics Generic Core Scales and Cerebral Palsy during running. Clin Biomech (Bristol, 14 Scianni A, Butler JM, Ada L, Teixeira- Module. Dev Med Child Neurol. 2006;48: Avon) Salmela LF. Muscle strengthening is not 442–449. . 2009;24:26–34. effective in children and adolescents with cerebral palsy: a systematic review. Aust J 21 Delp SL, Anderson FC, Arnold AS, et al. Physiother OpenSim: open-source software to create . 2009;55:81–87. and analyze dynamic simulations of movement. IEEE Trans Biomed Eng. 2007;55: 1940–1950.

February 2010 Volume 90 Number 2 Physical Therapy f 279