Patella Alta in Ambulatory Children With Cerebral Palsy: Prevalence and Functional Significance.

BACKGROUND: Gait dysfunction associated with patella alta (PA) in subjects with cerebral palsy (CP) has been presumed but not objectively established clinically or through biomechanical modeling. It is hypothesized that PA is associated with increasing level of motor impairment, increasing age, obesity, and worse stance phase knee kinematics and kinetics in children with CP. METHODS: Retrospective case series of 297 subjects with CP studied in our Motion Analysis Center. Data analyzed included patient demographics (age, body mass index, CP classification), patella height (Koshino-Sugimoto Index), and knee kinematics and kinetics. RESULTS: PA was present in 180 of 297 subjects (61%), in 68 of the 146 (47%) with unilateral CP, and 112 of 151 subjects (74%) with bilateral CP. For unilateral CP, the prevalence of PA was not significantly different between Gross Motor Function Classification System (GMFCS) I and II (P=0.357). For bilateral CP, the prevalence of PA in GMFCS III was significantly greater than in GMFCS I and II (P=0.02). Regression analysis showed a significant trend between increasing age and PA in unilateral and bilateral groups (P<0.001 and 0.001, respectively). The prevalence of PA was not significantly different across body mass index categories for either unilateral or bilateral groups. There were only 2 of 10 significant correlations between PA and gait parameters for subjects with unilateral CP functioning at the GMFCS I and II levels. There were 8 of 12 significant correlations between PA and gait parameters for subjects with bilateral CP functioning at the GMFCS I, II, and III levels. CONCLUSIONS: PA is common in ambulatory children with CP across topographic types and motor functional levels. PA is well tolerated with respect to gait dysfunction in unilateral CP, but may contribute to crouch gait in bilateral CP. Gait dysfunction cannot be inferred from the radiographic assessment of patellar height, and radiographic evidence of PA by itself does not justify surgical correction with patellar tendon advancement or shortening. LEVEL OF EVIDENCE: Level III-prognostic, retrospective series.

Surgical management of persistent intoeing gait due to increased internal tibial torsion in children.

BACKGROUND: Intoeing gait is frequently seen in developing children, and in most cases it resolves with growth. However, persistent, extreme intoeing gait, due to increased internal tibial torsion, may disrupt gait function. At our institution, children with symptomatic intoeing gait are evaluated per a standardized protocol, which includes quantitative gait analysis. When the primary cause is increased internal tibial torsion, surgical correction by supramalleolar tibial rotational osteotomy is recommended. METHODS: The study design was a retrospective case series, with normative controls (31 children), of typically developing children with symptomatic intoeing gait who were treated by isolated supramalleolar tibial rotation osteotomy (28 children, with 45 treated extremities). Preoperative and 1-year postoperative physical examination, kinematic, kinetic, and pedobarographic data were compared. Patient-reported and parent-reported outcomes in functional and satisfaction domains were assessed by items on a 7-point questionnaire. RESULTS: Internal tibial torsion, foot progression angle, and knee rotation were normalized following tibial rotation osteotomy. Compensatory external hip rotation and external knee progression angle were significantly improved but not normalized following tibial rotation osteotomy. An increased coronal plane knee varus moment was significantly decreased following surgery. Increased sagittal and transverse plane knee moments were significantly decreased but not normalized following surgery. Significant improvements were observed with respect to tripping, falling, foot/ankle pain, and knee pain following surgery. CONCLUSIONS: Children with symptomatic intoeing gait because of increased internal tibial torsion have characteristic primary and compensatory kinematic gait deviations that result in increased loading about the knee during the stance phase of gait. Correction of the internal tibial torsion by rotation osteotomy improves, but does not normalize, all the kinematic and kinetic gait deviations associated with intoeing gait. The association between increased internal tibial torsion and degenerative arthritis of the knee in adults may be a consequence of longstanding increased loading of the knee joint due to the kinematic gait deviations seen with intoeing gait. LEVEL OF EVIDENCE: Therapeutic intervention, level III.

Addressing some misperceptions of the joint coordinate system.

The joint coordinate system convention proposed by Grood and Suntay has been widely adopted, but often misrepresented. Previous work has argued by logical explanation of the approach that the joint coordinate system is a set of Euler or Cardan angles. The current work offers both an explanation and mathematical proof that the joint coordinate system convention is equivalent to a Cardan angle sequence, thereby demonstrating the joint coordinate system angles are both sequence dependent and orthogonal.

In-toeing gait in children with clubfoot and the effect of tibial rotation osteotomy.

In-toeing gait is common after treatment for clubfoot deformity and is often secondary to residual internal tibial torsion. The purpose of the current study was to characterize the gait pattern in children with an intoeing gait pattern associated with talipes equinovarus (TEV) deformity, identify secondary changes at the hip that occur with intoeing, and determine if these secondary effects resolve after correction of tibial torsion. Patients with a diagnosis of TEV deformity, in-toeing gait secondary to residual internal tibial torsion corrected with tibial rotation osteotomy (TRO) and complete preoperative and postoperative motion analysis studies obtained approximately 1 year apart, were included in the study. Nineteen children (19 left extremities) with a TRO at a mean age of 8.2 years met inclusion criteria. Clinical examination showed improvement in tibial torsion assessment by measure of the thigh foot axis and transmalleolar axis. Kinematically, an abnormal internal FPA was present in all cases preoperatively, was corrected to normal in 12 (63%), remained internal in 5 (26%), and was abnormally external in 2 (11%). External hip rotation was identified in 13 (68%) cases preoperatively. Hip rotation was normalized postoperatively in 7 (54%), and was unchanged in the remaining 6 (46%). TRO provides effective correction of excessive internal tibial torsion, resolution of kinematic internal knee rotation, and normalization of the internal foot progression angle in the majority of patients with TEV deformity. External hip rotation resolved in approximately 50% of cases. Overcorrection of the internal FPA is possible when secondary changes at the hip do not resolve.

Femoral anteversion assessment: Comparison of physical examination, gait analysis, and EOS biplanar radiography.

BACKGROUND: Multiple measurement methods are available to assess transverse plane alignment of the lower extremity. RESEARCH QUESTION: This study was performed to determine the extent of correlation between femoral anteversion assessment using simultaneous biplanar radiographs and three-dimensional modeling (EOS imaging), clinical hip rotation by physical examination, and dynamic hip rotation assessed by gait analysis. METHODS: Seventy-seven patients with cerebral palsy (GMFCS Level I and II) and 33 neurologically typical children with torsional abnormalities completed a comprehensive gait analysis with same day biplanar anterior-posterior and lateral radiographs and three-dimensional transverse plane assessment of femoral anteversion. Correlations were determined between physical exam of hip rotation, EOS imaging of femoral anteversion, and transverse plane hip kinematics for this retrospective review study. RESULTS: Linear regression analysis revealed a weak relationship between physical examination measures of hip rotation and biplanar radiographic assessment of femoral anteversion. Similarly, poor correlation was found between clinical evaluation of femoral anteversion and motion assessment of dynamic hip rotation. Correlations were better in neurologically typical children with torsional abnormalities compared to children with gait dysfunction secondary to cerebral palsy. SIGNIFICANCE: Dynamic hip rotation cannot be predicted by physical examination measures of hip range of motion or from three-dimensional assessment of femoral anteversion derived from biplanar radiographs.

Impact of ankle-foot orthoses on static foot alignment in children with cerebral palsy.

BACKGROUND: Children with cerebral palsy who are able to walk are often managed with an ankle-foot orthosis to assist with walking. Previous studies have shown kinematic, kinetic, and energetic benefits during gait with the addition of an ankle-foot orthosis, although the mechanism of this gait improvement is unknown. The ability of orthoses to correct foot malalignment in children with cerebral palsy is not known. The current study was performed to determine the impact of orthoses on static foot alignment in children with cerebral palsy. METHODS: A retrospective radiographic review was performed for 160 feet (102 patients). All patients had a diagnosis of cerebral palsy. Standing anteroposterior and lateral radiographs of the foot and ankle were made with the patient barefoot and while wearing the prescribed orthosis and were compared with use of the technique of quantitative segmental analysis of foot and ankle alignment. RESULTS: Analysis of the foot and ankle radiographs made with the patient barefoot and while wearing the brace revealed significant changes in all measurements of segmental alignment (p < 0.05). The magnitudes of these differences were small (<6 degrees or <10%) and would be considered clinically unimportant. The coupled malalignment of equinoplanovalgus (clinical flatfoot) showed radiographic correction of at least one segment (hindfoot, midfoot, or forefoot) to within the normal range in 24% to 44% of the feet. The coupled malalignment of equinocavovarus (clinical high arched foot) showed correction of at least one segment to within the normal range in 5% to 20% of feet. CONCLUSIONS: The present study demonstrates that the use of the ankle-foot orthoses failed to improve the static foot alignment in the majority of feet in children with cerebral palsy who were able to walk. LEVEL OF EVIDENCE: Therapeutic Level IV. See Instructions to Authors for a complete description of levels of evidence.

Indications for orthoses to improve gait in children with cerebral palsy.

Orthoses are frequently used to improve the gait of children with cerebral palsy. Optimal clinical decision-making for improving gait through orthotic management requires an understanding of the biomechanics of the foot and ankle during normal gait, the pathophysiology and pathomechanics of gait disruption in children with cerebral palsy, and the biomechanical characteristics of various orthoses. The clinician must seek to integrate his or her own goals with those of the child and family, the therapist, and the orthotist. Design, indications, and cost should be considered when choosing an orthosis. The physician can construct a paradigm for clinical decision-making, focusing on the evaluation of ankle/foot alignment, range of motion, and assessment of dynamic gait deviations. This paradigm will guide the clinician in the use of orthoses to improve gait in children with cerebral palsy. For optimal orthotic management, the physician must clearly identify the gait deviation and functional deficits to be addressed with the orthosis. The outcome of the orthotic intervention should be documented as objectively as possible.

Biomechanical changes associated with femoral derotational osteotomy.

Torsional deformities of the femur in children may occur as a result of either idiopathic or neuromuscular disorders and may be corrected with derotational osteotomies. Regardless of the underlying etiology, neither the effects of the torsional pathologies nor the alterations resulting from corrective osteotomies are well understood. A study of children with isolated femoral anteversion undergoing a single corrective procedure may assist in understanding the biomechanics of the pathology and the efficacy of surgical correction. A multicenter retrospective study included 25 subjects with idiopathic femoral anteversion who underwent femoral derotational osteotomy and had completed pre and postoperative gait analyses. Both changes with surgery and comparisons to typically developing controls were analyzed. Reduced gait pathology and expected improvements in hip rotation and foot progression were found with derotational osteotomy. Overall gait pathology and pathological differences in pelvic tilt, hip flexion moment and knee adduction moment were found comparing anteversion subjects with typically developing subjects. Following surgery, only hip rotation was significantly and clinically different from typically developing subjects, changing from relatively inward to outward. Idiopathic femoral anteversion creates multifaceted and significant alterations to normal gait and should not be considered solely a cosmetic issue. Additionally, the efficacy of derotational osteotomy is illustrated and may be more broadly applied to other conditions where pathologic femoral anteversion is present.

Optimization of walking ability of children with cerebral palsy.

A new paradigm based on an appreciation of the biomechanics of normal and pathologic gait and a better understanding of muscle-tendon unit anatomy and physiology has emerged for orthopaedic clinical decision making to optimize the ambulatory abilities of children with cerebral palsy. This quantitative, biomechanically based approach has been accepted as a research and teaching tool and as an instrument of outcome assessment; however, controversy remains concerning the expense of using this approach and about its accuracy and repeatability. This paradigm is used within a diagnostic matrix consisting of five data sources. Members of the clinical and technical teams from the motion analysis laboratory interpret data from the clinical history, physical examination, diagnostic imaging, quantitative gait analysis, and examination under anesthesia. The certainty of intervention selection is proportional to the consistency of the data within the diagnostic matrix. When inconsistencies in the data exist, input from both the clinical and technical teams is needed to resolve discrepancies. Working within the framework of the diagnostic matrix, it is possible to identify the indications used in the selection and recommendation of musculoskeletal surgical interventions to optimize gait in children with cerebral palsy. It is important to examine indications and controversies for surgical intervention related to iliopsoas recession, femoral rotational osteotomy, medial hamstring lengthening, rectus femoris transfer, and gastrocnemius recession.

Kinematics and kinetics of normal and planovalgus feet during walking.

Planovalgus deformity is prevalent in cerebral palsy patients, but very few studies have quantitatively reported differences between planovalgus and normal foot function. Intersegmental foot kinetics have not been reported in this population. In this study, a three segment (hindfoot, forefoot, hallux) kinematic and kinetic model was applied to typically developing (n=10 subjects, 20 feet) and planovalgus (n=10 subjects, 18 feet) pediatric subjects by two clinicians for each subject. Intra-clinician and inter-clinician repeatability of kinematic variables have been previously reported. Variability of kinetic outcomes (joint moments and power) is reported and found to be equally repeatable in typically developing and planovalgus groups. Kinematic differences in the planovalgus foot including excessive ankle eversion (valgus) and plantarflexion, reduced ankle flexion range of motion, and increased midfoot joint dorsiflexion and pronation reflected the reported pathology. Contrary to clinical expectations no significant difference was observed in midfoot flexion or ankle eversion ranges of motion. Kinetic differences in planovalgus feet compared to typically developing feet included reduced ankle plantarflexion moment, ankle power and midfoot joint power.

A multi-segment foot model based on anatomically registered technical coordinate systems: method repeatability and sensitivity in pediatric planovalgus feet.

Several multisegment foot models have been proposed and some have been used to study foot pathologies. These models have been tested and validated on typically developed populations; however application of such models to feet with significant deformities presents an additional set of challenges. For the first time, in this study, a multisegment foot model is tested for repeatability in a population of children with symptomatic abnormal feet. The results from this population are compared to the same metrics collected from an age matched (8-14 years) typically developing population. The modified Shriners Hospitals for Children, Greenville (mSHCG) foot model was applied to ten typically developing children and eleven children with planovalgus feet by two clinicians. Five subjects in each group were retested by both clinicians after 4-6 weeks. Both intra-clinician and inter-clinician repeatability were evaluated using static and dynamic measures. A plaster mold method was used to quantify variability arising from marker placement error. Dynamic variability was measured by examining trial differences from the same subjects when multiple clinicians carried out the data collection multiple times. For hindfoot and forefoot angles, static and dynamic variability in both groups was found to be less than 4° and 6° respectively. The mSHCG model strategy of minimal reliance on anatomical markers for dynamic measures and inherent flexibility enabled by separate anatomical and technical coordinate systems resulted in a model equally repeatable in typically developing and planovalgus populations.

A multi-segment foot model based on anatomically registered technical coordinate systems: method repeatability in pediatric feet.

Several multi-segment foot models to measure the motion of intrinsic joints of the foot have been reported. Use of these models in clinical decision making is limited due to lack of rigorous validation including inter-clinician, and inter-lab variability measures. A model with thoroughly quantified variability may significantly improve the confidence in the results of such foot models. This study proposes a new clinical foot model with the underlying strategy of using separate anatomic and technical marker configurations and coordinate systems. Anatomical landmark and coordinate system identification is determined during a static subject calibration. Technical markers are located at optimal sites for dynamic motion tracking. The model is comprised of the tibia and three foot segments (hindfoot, forefoot and hallux) and inter-segmental joint angles are computed in three planes. Data collection was carried out on pediatric subjects at two sites (Site 1: n=10 subjects by two clinicians and Site 2: five subjects by one clinician). A plaster mold method was used to quantify static intra-clinician and inter-clinician marker placement variability by allowing direct comparisons of marker data between sessions for each subject. Intra-clinician and inter-clinician joint angle variability were less than 4°. For dynamic walking kinematics, intra-clinician, inter-clinician and inter-laboratory variability were less than 6° for the ankle and forefoot, but slightly higher for the hallux. Inter-trial variability accounted for 2-4° of the total dynamic variability. Results indicate the proposed foot model reduces the effects of marker placement variability on computed foot kinematics during walking compared to similar measures in previous models.

Ankle dorsiflexor function after plantar flexor surgery in children with cerebral palsy.

BACKGROUND: Surgical lengthening is used to address both overactivity and shortening of the spastic agonist muscle in children with cerebral palsy. It has been presumed that the function of the antagonist muscle will improve when the spastic agonist muscle has been surgically lengthened. The purposes of the current study were to use quantitative gait analysis to determine the prevalence of the ankle dorsiflexor muscles (antagonist) dysfunction during the swing phase of the gait cycle and to analyze how this function is affected following surgical lengthening of the ankle plantar flexor muscles (agonist). METHODS: The study design was a retrospective, cohort series of fifty-three children with cerebral palsy who underwent gait analysis before and after surgical lengthening of the gastrocnemius-soleus muscle group. Data from the physical examination, gait study kinematics, and dynamic electromyography in swing phase were analyzed. RESULTS: The mean age at the time of the initial gait analysis was eight years and eleven months. Significant improvements were noted in ankle dorsiflexion passive range of motion (p < 0.001), ankle dorsiflexor selective control (p = 0.002), ankle dorsiflexor strength (p = 0.001), and peak and mean ankle dorsiflexion in swing phase (p < 0.001 for each) following ankle plantar flexor lengthening surgery. Active ankle dorsiflexor function in swing phase was present in 79% of the extremities prior to ankle plantar flexor surgery. Swing phase dorsiflexor function was present in 96% of the extremities following surgery, with ten extremities improving from absent to present. CONCLUSIONS: The kinematic data support the clinical impression that ankle dorsiflexion during swing phase is improved following ankle plantar flexor lengthening surgery in children with cerebral palsy. In the majority of patients, this was a consequence of the correction of a fixed equinus contracture of the ankle plantar flexors that was constraining preexisting ankle dorsiflexor function. Weakness of all of the muscles is common, and surgical lengthening should only be considered for the correction of recalcitrant muscle contractures. Improved function of the antagonist muscle should be anticipated and optimized by appropriately focused strength training and other modalities during rehabilitation.

The efficacy of the floor-reaction ankle-foot orthosis in children with cerebral palsy.

BACKGROUND: The floor-reaction ankle-foot orthosis is commonly prescribed for children with cerebral palsy who walk with excessive ankle dorsiflexion and excessive knee flexion during the stance phase of gait. The purposes of this study were to evaluate the efficacy of this orthosis objectively and to identify clinical parameters that may compromise its function. METHODS: All children with cerebral palsy who had comprehensive gait analyses in both barefoot and braced walking conditions during a single visit to our Motion Analysis Laboratory between January 2001 and August 2007 were identified. Kinematic study parameters included mean sagittal dynamic range of motion of the ankle in stance, peak ankle dorsiflexion in stance, peak knee extension in midstance, and mean foot progression angle in stance. The minimum sagittal knee moment in midstance was also examined in this study for subjects who walked without assistive devices. Range-of-motion and skeletal alignment data obtained from the physical examination record of each subject included knee flexion contracture, popliteal angle, hip flexion contracture, and thigh-foot angle. RESULTS: Twenty-seven children had quantitative gait analyses (barefoot and with the orthoses in the same visit). The mean sagittal plane dynamic range of motion of the ankle in stance was reduced from 23 degrees +/- 9 degrees when walking barefoot to 10 degrees +/- 3 degrees when the orthosis was worn (p < 0.001), and the mean peak knee extension in midstance improved from 29 degrees +/- 14 degrees of flexion to 18 degrees +/- 14 degrees of flexion (p = 0.013). Strong negative linear correlations were found between the magnitude of knee and hip flexion contractures on physical examination and the amount of peak knee extension in midstance (r = -0.784 and r = -0.705, respectively). A strong positive correlation was found between the mean minimum sagittal knee moment in midstance and the amount of peak knee extension in midstance (r = 0.820). Our investigation did not provide evidence of a correlation between peak knee extension in midstance and any of the following parameters in the orthosis: clinical examination measurements of the thigh-foot angle (r = 0.120), the popliteal angle (r = -0.300), or the mean foot progression angle in the stance phase of gait (r = -0.188). CONCLUSIONS: The floor-reaction ankle-foot orthosis is effective in restricting sagittal plane ankle motion during the stance phase of gait in patients with cerebral palsy. As a result, improvements in knee extension and the sagittal plane knee extensor moment in stance phase are achieved. The best outcomes with this orthosis, as determined by peak knee extension in midstance, were seen in the subjects with knee and hip flexion contracture of < or =10 degrees . Knee and hip flexion contractures of > or =15 degrees were found to limit the efficacy of the orthosis in controlling knee extension in midstance. Such contractures should be considered as contraindications to the prescription of this orthosis or should be addressed (surgically or otherwise) prior to the application of a floor-reaction ankle-foot orthosis in these patients.

Dynamic pedobarography for children: use of the center of pressure progression.

BACKGROUND: Although pedobarography has been widely used in quantitative clinical gait analysis for children, the collection, processing, analysis, and interpretation of the data vary widely. In most cases in children, foot dysfunction during gait is primarily a consequence of skeletal segmental malalignment, which can be characterized by the location and duration of the center of pressure progression (COPP) relative to the foot. This study determines the validity and reliability of a technique using the COPP and establishes a normative database for the COPP in children. STUDY DESIGN: Prospective case series to evaluate a diagnostic test. METHODS: Simultaneous pedobarograph and kinematic data collection was performed on 23 children (46 feet) who were neurologically healthy. The validity of the COPP technique was determined by comparing the pedobarograph- and kinematic-based determinations of the orientation of the longitudinal (or long) axis of the foot, an essential component of the COPP approach. Intrarater and interrater reliability for the pedobarograph-based technique were determined by comparing repeated measures of the long axis of the foot from 4 analysts. Normative data for the location and duration of the COPP were generated from this cohort of neurologically healthy children. RESULTS: The mean difference for the long axis of the foot between the pedobarograph- and kinematic-based methods was 2.3 degrees (P < 0.001). The mean difference between first and second determinations of the long axis of the foot by the same analyst was 1.0 degrees (P < 0.001; correlation coefficient, 0.975). The mean difference between the 4 analysts' determinations of the long axis of the foot was 1.9 degrees (P < 0.001; correlation coefficient, 0.969). The normal COPP is located under the heel segment for 23.7% of stance, under the midfoot segment for 28.7% of stance, and under the forefoot segment for 47.5% of stance. CONCLUSIONS: This study establishes clinically acceptable validity and reliability for the pedobarograph COPP technique and determines the location and duration of the COPP in a cohort of neurologically healthy children. CLINICAL RELEVANCE: This standardized approach to the determination of foot loading patterns, based upon normative data, should facilitate the characterization of abnormal foot loading patterns, clinical decision making, and the assessment of outcome after a variety of interventions.

Idiopathic toe walking: a kinematic and kinetic profile.

PURPOSE: The differential diagnosis in children who walk on their toes includes mild spastic diplegia and idiopathic toe walking (ITW). A diagnosis of ITW is often one of exclusion. To better characterize the diagnosis of ITW, quantitative gait analysis was utilized in a series of patients with an established diagnosis of ITW. STUDY DESIGN: Patients with an established diagnosis of ITW were analyzed by quantitative gait analysis. Data were recorded as each subject walked in a self-selected toe-walking pattern. The subject was then asked to ambulate making every effort to walk in a normal heel-toe reciprocating fashion. Data were collected to determine if this group of idiopathic toe walkers was able to normalize their gait. Datasets were compared with each other and with historical normal controls. RESULTS: Fifty-one neurologically normal children (102 extremities) with ITW were studied in the Motion Analysis Laboratory at a mean age of 9.3 years. In the self-selected trials, significant deviations in both kinematics and kinetics at the level of the ankle were identified. Disruption of all 3 ankle rockers and a plantar flexion bias of the ankle throughout the gait cycle were most commonly seen. When asked to attempt a normal heel-toe gait, 17% of the children were able to normalize both stance and swing variables. In addition, 70% were able to normalize some but not all of the stance and swing variables. CONCLUSION: Quantitative gait analysis is an effective tool for differentiating mild cerebral palsy from ITW. Kinematic and kinetic distinctions between the diagnoses are evident at the knee and ankle. The ability to normalize on demand at least some of the kinematic and kinetic variables associated with toe walking is seen in most children with ITW.

Prevalence of obesity in ambulatory children with cerebral palsy.

BACKGROUND: According to the most recent data, an estimated 17.1% of children in the United States are obese. We found no published studies documenting the prevalence of obesity in ambulatory children with cerebral palsy. The purpose of this study was to document the prevalence of obesity in ambulatory children with cerebral palsy and examine the trend in this measure over the last decade. METHODS: A retrospective review was performed to analyze the age, gender, height, weight, physical classification of the cerebral palsy, and functional level as determined with the Gross Motor Function Classification System (GMFCS) of all children with cerebral palsy who had a gait analysis performed in the Motion Analysis Laboratory of our institution between January 1994 and December 2004. This information was used to determine the prevalence of obesity (a body mass index in or above the 95th percentile of the sex-specific body mass index-for-age growth chart) in this population and its relationship to age, gender, the physical classification of the cerebral palsy, and the GMFCS level. RESULTS: When the data were grouped into three time periods (1994 to 1997, 1998 to 2002, and 2003 to 2004), a significant increase in obesity over time was noted (p = 0.017). The prevalences increased from 7.7% to 14% to 16.5% in the respective time periods. The prevalence increased over time in both males and females, those with hemiplegia and those with diplegia, and those with level-I function and those with level-II function according to the GMFCS. The association between obesity and time was significant in the female (p = 0.015), hemiplegic (p = 0.049), less than eight-year-old (p = 0.020), and GMFCS level-II (p = 0.003) groups. We found that the time period was independently associated with obesity when we controlled for age, type of cerebral palsy, and GMFCS level (p = 0.014). Children with a lesser degree of involvement (GMFCS level II) had twice the odds of becoming obese than did children with greater involvement (GMFCS level III). CONCLUSIONS: The prevalence of obesity in ambulatory children with cerebral palsy has risen over the last decade from 7.7% to 16.5%, an increase that is similar to that seen in the general pediatric population in the United States. This finding may have a major impact on the general health and functional abilities of these children as they reach adult life.