Pathological Movements of the Pelvis and Trunk During Gait in Children With Cerebral Palsy: A Cross-Sectional Study With 3-Dimensional Kinematics and Lower Lumbar Spinal Loading

Pathological Movements of the Pelvis and Trunk During Gait in Children With Cerebral Palsy: A... Abstract Background Increased loading at the lumbar spine, particularly in the coronal plane, has been reported in children with cerebral palsy (CP). As pelvic and trunk movements associated with Trendelenburg and Duchenne type gait are most significant in the coronal plane, the potential exists for lower lumbar spinal loading to be negatively affected in children with CP and these types of movement patterns. Objective The objective of this study was to assess trunk and pelvic kinematics and lower lumbar spinal loading patterns in children with CP and Trendelenburg and Duchenne type gait. Design This was a cross-sectional study. Methods Three-dimensional kinematic (lower limb and thorax) and L5-S1 kinetic data were recorded. Children were divided according to clinical presentation of Trendelenburg or Duchenne type gait. Several discrete kinematic and kinetic parameters were assessed between groups. Results Three distinct pelvic and trunk movement patterns were identified for children with CP: Trendelenburg, Duchenne, and complex Trendelenburg–Duchenne. Peak L5-S1 lateral bending moments were increased by 62% in children with CP and Duchenne type gait. Children with CP and complex Trendelenburg-Duchenne gait demonstrated the largest deviations from normal, with increased peak ipsilateral and contralateral directed moments of 69% and 54%, respectively, compared with children with typical development. Limitations A test-retest reliability analysis or measure of minimal detectable change was not conducted as part of this study. Results suggest that measures of minimal detectable change may be high for some of the reported variables. In addition, the inverse dynamic approach determines only the net intersegmental reactive forces that reflect the effect of external loads. Previous studies have shown that spinal loads may be larger than the net intersegmental force. Conclusions Trendelenburg and Duchenne type movements were not always distinct, and a third type of movement, a combination of the two, was the most common in this study. Clinicians should be aware that children with CP and the Duchenne type or the complex Trendelenburg-Duchenne type of gait pattern experience abnormal loading that may have significant implications for the lower spine in the long term. Excessive movement of the trunk and pelvis are characteristic of walking in children with cerebral palsy (CP).1–5 In particular, Trendelenburg and Duchenne type gait patterns are often seen in this group.6 A Trendelenburg type gait pattern is characterized by a drop of the pelvis in the coronal plane on the unloaded side during stance.6–8 The trunk is tilted toward the supporting limb with respect to the pelvis while maintaining a neutral position with respect to the global reference frame.7 Hip adduction of the supporting limb is also increased.7 A Duchenne type gait pattern is characterized by a trunk lean toward the supporting limb with the pelvis level or elevated on the unloaded side.6,7,9 This type of compensatory movement of the trunk moves body weight toward the center of the hip resulting in a reduced hip abductor moment.6,10 Both movement patterns are often a compensatory mechanism for hip abductor weakness or hip dysfunction of the supporting limb.6,10,11 However, a Duchenne type gait pattern may have negative effects on the knee joint by increasing the lever arm around the knee.10 Additionally, it has been suggested that a Duchenne type gait pattern may result in spinal problems.6 Likewise, a Trendelenburg type gait pattern is thought to be harmful to the hip.6 However, while the effects of Trendelenburg and Duchenne type walking patterns on kinetics at the knee and hip have been considered,6,7,10 the effects of this movement further up the kinematic chain, particularly at the lower spine, are unknown. Our group have recently reported increased reactive forces and moments at the lower lumbar spine during gait in children with CP.12 As thorax coronal plane movement became more excessive, children with CP demonstrated increased peak L5-S1 reactive forces and moments of up to 63% and 90%, respectively.12 Consequently, as Trendelenburg and Duchenne type gait patterns are most significant in the coronal plane, we questioned whether loading at the lumbar spine would be adversely affected in children with CP when walking with these movement patterns? As the position of the trunk changes with respect to the pelvis during Trendelenburg and Duchenne type gait, mechanical loads at the spine must change. However, the extent to which this occurs is unclear. A thorough understanding of any potential negative effects at the spine is important because Duchenne type gait patterns were previously reported to be protective of the hip joint by increasing the cover of the femoral head.6 In addition, this type of walking pattern has also been recommended as a noninvasive intervention for hip pain in adults and as a conservative treatment in children with Legg-Calvé-Perthes disease.13,14 However, while it is not clear whether these recommendations are followed routinely in clinical practice, the potential for increased demands at the spine may make such recommendations counterproductive. Additionally, excessive spinal tilting as a consequence of these types of gait patterns may further increase the risk of low back pain in a population with CP, in whom incidents of low back pain were previously reported to be up to 63% higher than those in controls.15–17 Therefore, it is important that clinicians are fully aware of any adverse effects that may occur at the lower lumbar spine when presented with Trendelenburg or Duchenne type gait patterns. Following from this, the objective of this study was to assess trunk and pelvic kinematics and three-dimensional reactive forces and moments at the lower lumbar spine in children with CP who appeared on clinical presentation to have Trendelenburg or Duchenne type patterns of movement during gait. The cohort of children, both those with CP and those with typical development (TD), was that used in our previous study.12 Children with CP were regrouped and reanalyzed according to clinical presentation of Trendelenburg or Duchenne type gait. Methods Study Design This was a cross-sectional study for both children with CP and children with TD. The study is a secondary analysis of previously published data by our group.12 Participants were regrouped and the data reanalyzed according to newly defined groups. Participants A sample size calculation was performed as part of an earlier study.12 In that study, coronal plane trunk flexion was used as the primary outcome measure; the hypothesis being tested using coronal plane trunk flexion was that coronal plane trunk flexion outside 1 standard deviation from the normal group was pathological. In the present study, using coronal plane trunk flexion as the primary outcome measure, a sample size of 26 was required at a power of 0.95 and a significance of 0.05. This number was exceeded, and 52 children with CP participated in this study. Children with CP were recruited from a cohort attending the gait analysis laboratory over a period of 9 months (n = 52; 33 boys, 19 girls; mean age = 11.02 years [SD = 3.01]; 21 with hemiplegia, 31 with diplegia). Inclusion criteria were: diagnosis of hemiplegic or diplegic CP and ability to walk independently. Children were excluded if they had had surgery within 1 year of presenting to the gait laboratory. Children with TD were invited to participate after a general email to staff and colleagues at the host institution (n = 26; 15 boys, 11 girls; mean age = 10.15 years [SD = 3.17] years). Children with TD without any previous history of neurological, musculoskeletal, or orthopedic problems were included in the study. A participant information leaflet was provided to parents and guardians, who then gave written informed consent. Data Collection A full barefoot 3-dimensional kinematic and kinetic analysis was performed using the CODA cx1 active marker system (Charnwood Dynamics Ltd, Leicestershire, United Kingdom) and 2 Kistler 9281B (Kistler Instruments Ltd., 13 Murrell Green Business Park, London Road, RG27 9GR Hook, Hampshire, United Kingdom) and 2 AMTI Accugait force platforms (176 Waltham Street, Watertown, MA). Data were collected using Codamotion ODIN software (v1.06 Build 01 09) at capture rates of 100 Hz (kinematic) and 200 Hz (kinetic), respectively. Infrared light-emitting diodes (LEDs) were placed on the lower limbs in accordance with a modified Helen Hayes model.18 Thorax kinematic data were captured using a single cluster protocol shown to be valid and comparable to other protocols for measuring thorax kinematics.19 This protocol consisted of a rigid mount containing LEDs and was placed at spinal level T3.19 For L5-S1 joint location, a LED was placed at the joint space of L5-S1. A virtual point was created at a position corresponding to 5% of the distance of the line between the L5-S1 LED and the midpoint of the anterior superior iliac spine.12,20 L5-S1 reactive forces and moments were then realized at this point. Participants walked unassisted at a self-selected pace with at least 4 clean data trials recorded per participant. One representative walking trial was chosen and analyzed for each participant according to standard laboratory protocol. Coronal Plane Pattern Assessment In order to determine pattern type, coronal plane kinematic data of all participants were visually analyzed and compared to TD data. For the purposes of this study, normal data were defined as ±1 standard deviation about average. Values outside this range were considered abnormal. Three different movement types were observed. Type 1 was defined by a pelvic drop on the unloaded side during stance phase outside the ±1 standard deviation band with a laterally flexed trunk position toward the supporting limb in relation to the pelvis and an upright trunk position in relation to the global reference frame (referred to as a Trendelenburg type gait pattern)7,14 (Fig. 1b). Eight children (15.4%; 2 with hemiplegia, 6 with diplegia) demonstrated a type 1 gait pattern. Type 2 was defined by a level or elevated pelvis on the unloaded side in conjunction with a trunk lean toward the supporting limb outside the ±1 standard deviation band (referred to as a Duchenne type gait pattern)7,14 (Fig. 1c). Four children (7.7%; 1 with hemiplegia, 3 with diplegia) demonstrated a type 2 gait pattern. Finally, type 3 was defined by a distinctive pelvic drop on the unloaded side during stance in conjunction with an excessive lean of the trunk outside TD limits toward the supporting limb with respect to both the pelvis and the global reference frame (referred to as the complex Trendelenburg-Duchenne type gait pattern) (Fig. 1d). Twenty-two children (42.3%; 1 with hemiplegia, 21 with diplegia) demonstrated a type 3 pattern. Four children (7.7%; 3 with hemiplegia, 1 with diplegia) demonstrated trunk and pelvic kinematic patterns within normal ranges. The remaining 14 children (26.9%; 14 with hemiplegia) demonstrated patterns not consistent with Trendelenburg or Duchenne type gait. Children who demonstrated patterns within normal ranges or not consistent with Trendelenburg or Duchenne type gait were excluded from further evaluation. Children were then grouped according to movement pattern types 1 to 3. Figure 1. View largeDownload slide Movement patterns in the coronal plane. (a) Normal trunk (T) and pelvis (P) positions. (b) Trendelenburg type pattern. Note the upright position of the trunk in the global frame and the significant drop of the pelvis on the contralateral side. (c) Duchenne type pattern. Note the lateral lean of the trunk in the global frame and the level (or slightly raised on the contralateral side) position of the pelvis. (d) Complex Trendelenburg-Duchenne pattern. Note the lateral lean of the trunk in the global frame and the drop of the pelvis on the contralateral side. Figure 1. View largeDownload slide Movement patterns in the coronal plane. (a) Normal trunk (T) and pelvis (P) positions. (b) Trendelenburg type pattern. Note the upright position of the trunk in the global frame and the significant drop of the pelvis on the contralateral side. (c) Duchenne type pattern. Note the lateral lean of the trunk in the global frame and the level (or slightly raised on the contralateral side) position of the pelvis. (d) Complex Trendelenburg-Duchenne pattern. Note the lateral lean of the trunk in the global frame and the drop of the pelvis on the contralateral side. Data Analysis A number of discrete kinematic and kinetic parameters were assessed between groups. The 4 kinematic parameters were as follows: maximum pelvic obliquity in stance, maximum trunk side flexion in stance (with respect to the global or laboratory reference frame), maximum trunk side flexion stance (with respect to the pelvis), and maximum hip adduction in stance. Trunk side flexion was measured with respect to the pelvis and with respect to the global or laboratory reference frame. The 6 kinetic parameters were as follows: hip abductor moment peak 1 (N·m/kg), hip abductor moment peak 2 (N·m/kg), peak L5-S1 medial/lateral force directed toward the ipsilateral side (N/kg), peak L5-S1 medial/lateral force directed toward the contralateral side (N/kg), peak L5-S1 lateral bending moment directed toward the ipsilateral side (N·m/kg), and peak L5-S1 lateral bending moment directed toward the contralateral side (N·m/kg). The analyzed side was the side to which the trunk leaned during stance. This was then referred to as “ipsilateral” for the duration of the gait cycle. The word “contralateral” refers to the opposite side for the same period of the gait cycle. Data were checked for distribution using the Shapiro-Wilk normality test. Differences between movement types and children with TD were assessed using a 1-way analysis of variance with Bonferroni post hoc tests for comparisons between groups. Dunnett tests were also used to compare each movement type with the TD group. For data that did not follow a normal distribution, differences were assessed using a Kruskal-Wallis test and post hoc Mann-Whitney U tests. All statistical analyses were performed using IBM SPSS Statistics (v23.0.0.2). The level of significance was set at 0.05. Additionally, ensemble average kinematic and kinetic profiles were visually analyzed for deviations between groups. Role of the Funding Source Funding was provided in part by the Scientific and Research Trust of the Central Remedial Clinic. The funder had no role in the study design or in the collection, analysis, or interpretation of the data. Results Kinematic Classification After coronal plane classification, 34 children with CP were assessed. Eight children (7 boys and 1 girl; mean age = 10.9 years [SD = 2.9]; mean height = 1.44 m [SD = 0.12]; mean weight = 38.04 kg [SD = 12.5]) had type 1 gait. Four children (3 boys and 1 girl; mean age = 12.0 years [SD = 2.9]; mean height = 1.52 m [SD = 0.18]; mean weight = 41.03 kg [SD = 20.2]) had type 2 gait. Twenty-two children (14 boys and 8 girls; mean age = 10.2 years [SD = 3.0]; mean height = 1.41 m [SD = 0.2]; mean weight = 37.2 kg [SD = 14.6]) had type 3 gait. Type 1 (Trendelenburg Gait Pattern) Children with a type 1 gait pattern demonstrated an increased peak pelvic obliquity in stance by 4.8 degrees compared with that in children with TD (Fig. 2; Tabs. 1 and 2). Peak trunk side flexion (with respect to the pelvis) was increased by 7.5 degrees, while peak trunk side flexion (with respect to the global or laboratory reference frame) remained within normal limits throughout the gait cycle. Peak hip adduction in stance was increased by 5.1 degrees compared with that in children with TD. Hip abductor moment remained close to that in children TD, with no statistically significant differences (Tabs. 1 and 2). Figure 2. View largeDownload slide Kinematic and kinetic ensemble average profiles for the Trendelenburg (type 1) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Figure 2. View largeDownload slide Kinematic and kinetic ensemble average profiles for the Trendelenburg (type 1) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Table 1. Values for Children With Typical Development (TD) and Children With Cerebral Palsy (CP) by Movement Typea Parameter  Children With TD  Children With CP  Concurrent 
P Value 
Determined by ANOVA  Post Hoc Comparisonc  Type 1 
(n = 8)  Type 2 
(n = 4)  Type 3 
(n = 22)  Peak pelvic obliquity stance (°)  5.42 (2.3)  10.2 (3.2)  2.27 (0.86)  9.87 (2.7)  < .001  A; B  Peak hip adduction in stance (°)b  6.90 (19.7)  12.0 (10.4)  6.27 (7.3)  9.83 (24.1)  .02  A; B  Peak trunk side flexion (with respect to global or laboratory reference frame) (°)  2.29 (4.0)  4.40 (2.2)  9.25 (2.5)  10.87 (3.7)  < .001  B; C  Peak trunk side flexion (with respect to pelvis) (°)  6.47 (4.6)  13.9 (4.1)  12.0 (2.9)  19.35 (5.8)  < .001    Peak L5-S1 moment toward contralateral (N·m/kg)b  −0.29 (0.41)  −0.37 (0.41)  −0.47 (0.7)  −0.49 (0.8)  < .001  A; B  Peak L5-S1 moment toward ipsilateral (N·m/kg)b  0.24 (0.56)  0.33 (0.33)  0.32 (0.35)  0.37 (1.12)  .01  B  Hip abductor moment peak 1 (N·m/kg)  0.53 (0.14)  0.56 (0.14)  0.58 (0.64)  0.41 (0.17)  .11  B  Hip abductor moment peak 2 (N·m/kg)  0.46 (0.14)  0.34 (0.24)  0.50 (0.26)  0.25 (0.23)  .00  B  Peak L5-S1 force toward contralateral (N/kg)b  −0.53 (1.42)  −0.88 (0.92)  −0.79 (0.44)  −1.1 (2.03)  .00  A; B  Peak L5-S1 force toward ipsilateral (N/kg)b  0.55 (1.48)  0.32 (0.88)  0.70 (0.29)  0.61 (1.65)  .55    Parameter  Children With TD  Children With CP  Concurrent 
P Value 
Determined by ANOVA  Post Hoc Comparisonc  Type 1 
(n = 8)  Type 2 
(n = 4)  Type 3 
(n = 22)  Peak pelvic obliquity stance (°)  5.42 (2.3)  10.2 (3.2)  2.27 (0.86)  9.87 (2.7)  < .001  A; B  Peak hip adduction in stance (°)b  6.90 (19.7)  12.0 (10.4)  6.27 (7.3)  9.83 (24.1)  .02  A; B  Peak trunk side flexion (with respect to global or laboratory reference frame) (°)  2.29 (4.0)  4.40 (2.2)  9.25 (2.5)  10.87 (3.7)  < .001  B; C  Peak trunk side flexion (with respect to pelvis) (°)  6.47 (4.6)  13.9 (4.1)  12.0 (2.9)  19.35 (5.8)  < .001    Peak L5-S1 moment toward contralateral (N·m/kg)b  −0.29 (0.41)  −0.37 (0.41)  −0.47 (0.7)  −0.49 (0.8)  < .001  A; B  Peak L5-S1 moment toward ipsilateral (N·m/kg)b  0.24 (0.56)  0.33 (0.33)  0.32 (0.35)  0.37 (1.12)  .01  B  Hip abductor moment peak 1 (N·m/kg)  0.53 (0.14)  0.56 (0.14)  0.58 (0.64)  0.41 (0.17)  .11  B  Hip abductor moment peak 2 (N·m/kg)  0.46 (0.14)  0.34 (0.24)  0.50 (0.26)  0.25 (0.23)  .00  B  Peak L5-S1 force toward contralateral (N/kg)b  −0.53 (1.42)  −0.88 (0.92)  −0.79 (0.44)  −1.1 (2.03)  .00  A; B  Peak L5-S1 force toward ipsilateral (N/kg)b  0.55 (1.48)  0.32 (0.88)  0.70 (0.29)  0.61 (1.65)  .55    aData are reported as mean (standard deviation) unless otherwise indicated. Descriptive statistics are reported for type 2. ANOVA = analysis of variance, type 1 = Trendelenburg only, type 2 = Duchenne only, type 3 = Trendelenburg-Duchenne. bMedian (range) for nonnormally distributed data, with distribution of all data assessed using a Shapiro-Wilk normality test. cResults of post hoc tests are indicated as follows: A = type 1-TD, B = type 3-TD, C = type 1-type 3. A breakdown of post hoc tests is reported in Table 2. View Large Table 2. Results of Post Hoc Analysis for Differences and 95% CIs Between Groupsa Parameter  Difference Between Type 1 
and TD  95% CI  Con-current 
P Value  Difference Between Type 3 and TD  95% CI  Con-current 
P Value  Difference Between Type 1 and 
Type 3  95% CI  Con-current 
P Value  Peak pelvic obliquity stance (°)  4.78  2.28 to 7.28  .00  4.46  2.66 to 6.25  .00  0.33  −2.53 to 3.18  1.00  Peak hip adduction in stance (°)b  5.10  −1.46 to 5.69  .00  2.93  6.02 to 11.1  .03  2.17    .30  Peak trunk side flexion (with respect to global or laboratory reference frame) (°)  2.11    .38  8.58  9.38 to 16.4  .00  6.47  2.39 to 10.55  .00  Peak trunk side flexion (with respect to pelvis) (°)  7.48  2.58 to 12.38  .00  12.89    .00  −5.40  −11.0 to 0.19  .06  Peak L5-S1 moment toward contralateral (N·m/kg)b  −0.08    .14  −0.20    .00  0.12    .08  Peak L5-S1 moment toward ipsilateral (N·m/kg)b  0.09    .33  0.13    .01  0.04    .37  Hip abductor moment peak 1 (N·m/kg)  0.03  −0.17 to 0.24  .96  −0.12  −0.27 to 0.03  .13  0.16  −0.08 to 0.39  .45  Hip abductor moment peak 2 (N·m/kg)  −0.12  −0.31 to 0.08  .35  −0.21  −0.35 to –0.07  .00  0.09  −0.13 to 0.31  1.00  Peak L5-S1 force toward contralateral (N/kg)b  −0.35    .03  −0.57    .00  0.22    .48  Peak L5-S1 force toward ipsilateral (N/kg)b  −0.23    .38  0.06    .53  0.29    .92  Parameter  Difference Between Type 1 
and TD  95% CI  Con-current 
P Value  Difference Between Type 3 and TD  95% CI  Con-current 
P Value  Difference Between Type 1 and 
Type 3  95% CI  Con-current 
P Value  Peak pelvic obliquity stance (°)  4.78  2.28 to 7.28  .00  4.46  2.66 to 6.25  .00  0.33  −2.53 to 3.18  1.00  Peak hip adduction in stance (°)b  5.10  −1.46 to 5.69  .00  2.93  6.02 to 11.1  .03  2.17    .30  Peak trunk side flexion (with respect to global or laboratory reference frame) (°)  2.11    .38  8.58  9.38 to 16.4  .00  6.47  2.39 to 10.55  .00  Peak trunk side flexion (with respect to pelvis) (°)  7.48  2.58 to 12.38  .00  12.89    .00  −5.40  −11.0 to 0.19  .06  Peak L5-S1 moment toward contralateral (N·m/kg)b  −0.08    .14  −0.20    .00  0.12    .08  Peak L5-S1 moment toward ipsilateral (N·m/kg)b  0.09    .33  0.13    .01  0.04    .37  Hip abductor moment peak 1 (N·m/kg)  0.03  −0.17 to 0.24  .96  −0.12  −0.27 to 0.03  .13  0.16  −0.08 to 0.39  .45  Hip abductor moment peak 2 (N·m/kg)  −0.12  −0.31 to 0.08  .35  −0.21  −0.35 to –0.07  .00  0.09  −0.13 to 0.31  1.00  Peak L5-S1 force toward contralateral (N/kg)b  −0.35    .03  −0.57    .00  0.22    .48  Peak L5-S1 force toward ipsilateral (N/kg)b  −0.23    .38  0.06    .53  0.29    .92  aTD = typical development, type 1 = Trendelenburg only, type 3 = Trendelenburg-Duchenne. bNonnormally distributed data, for which 95% CIs were not reported. View Large L5-S1 ensemble average kinetic profiles were similar for children with a type 1 gait pattern and children with TD (Fig. 2). However, children with a type 1 gait pattern demonstrated an increased peak L5-S1 force directed toward the contralateral limb at initial swing phase (increase of 0.35 N/kg or ≈66%). No other statistically significant differences were present for medial lateral force or lateral bending moment for children with a type 1 gait pattern compared with that in children with TD (Tabs. 1 and 2). Type 2 (Duchenne Gait Pattern) The number of children demonstrating a type 2 gait pattern was too small for statistical analysis (n = 4). As an alternative, descriptive statistics are provided and ensemble average profiles commented on. For children with a type 2 gait pattern, pelvic obliquity remained relatively flat and close to a neutral position throughout the gait cycle (Fig. 3; Tab. 1). Peak trunk side flexion (with respect to the pelvis) demonstrated an increase of 5.5 degrees compared with that in children with TD. In addition, peak trunk side flexion (with respect to the global or laboratory reference frame) was also increased compared with that in children with TD, by 7.0 degrees (Tab. 1). Hip abductor moment was slightly reduced at peak points but remained just within normal limits (Fig. 3). Figure 3. View largeDownload slide Kinematic and kinetic ensemble average profiles for the Duchenne (type 2) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Figure 3. View largeDownload slide Kinematic and kinetic ensemble average profiles for the Duchenne (type 2) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Kinetic ensemble average profiles demonstrated an increased lateral bending moment toward the contralateral side outside normal limits during stance for children with a type 2 gait pattern compared with that in children with TD (Fig. 3). This peak moment was up to 62% greater than that in children with TD (Tab. 1). Lateral bending moment toward the ipsilateral side during swing remained within normal limits (Fig. 3; Tab. 1). Medial lateral force also remained within normal limits throughout the gait cycle (Fig. 3). Type 3 (Complex Trendelenburg-Duchenne Gait Pattern) For children with a type 3 gait pattern, peak pelvic obliquity was significantly increased by 4.5 degrees compared with that in children with TD (Fig. 4; Tabs. 1 and 2). A similar increase was demonstrated when compared with the type 2 gait pattern (Tab. 1). Peak trunk side flexion, with respect to both the pelvis and the global or laboratory reference frame, was significantly increased by 12.9 and 8.6 degrees, respectively, compared with the values in children with TD (Tabs. 1 and 2). Peak hip adduction in stance demonstrated a statistically significant increase of 2.9 degrees, while hip abductor moment peak 1 remained within normal limits. However, hip abductor peak 2 was significantly decreased for children with a type 3 gait pattern compared with that for children with TD (reduction of 0.21 N·m/kg or ≈54%) (Tabs. 1 and 2). Figure 4. View largeDownload slide Kinematic and kinetic ensemble average profiles for the complex Trendelenburg-Duchenne (type 3) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Figure 4. View largeDownload slide Kinematic and kinetic ensemble average profiles for the complex Trendelenburg-Duchenne (type 3) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Several significant deviations were evident in the kinetic ensemble average profiles for children with a type 3 gait pattern compared with that in children with TD (Fig. 4). Like children with a type 1 gait pattern, children with a type 3 gait pattern demonstrated an increased peak L5-S1 force directed toward the contralateral limb at initial swing phase. However, in this case, peak force was approximately double that in children with TD (increase of 0.57 N/kg or ≈107%) (Tabs. 1 and 2). Both peak L5-S1 ipsilateral and contralateral directed moments were also larger for children with CP than for children with TD (increased by 69% and 54%, respectively) (Tabs. 1 and 2). Discussion Pelvic, trunk, and hip kinematics during type 1 (Trendelenburg gait) in this study were consistent with the literature when describing this type of movement pattern.7 Hip abductor moment remained within normal limits as expected.7 Consequently, as hip abductor moment represents the predominant factor in hip joint loading,7,21 loading at the hip can be judged to be close to normal. The resulting effects at the lumbar spine were small. Only mild differences were present between the CP and TD groups. Most notably, at the point of ipsilateral toe off, medial/lateral force directed toward the contralateral side was increased. At this point, the pelvis was raised on the contralateral side, the trunk leaned toward the contralateral pelvis and a resultant increased force was demonstrated. When ensemble average profiles were considered, an increased lateral bending moment was evident at this point (Fig. 2). However, this increased moment was not statistically significant and remained just within normal limits suggesting the impact on the spine and corresponding trunk musculature was small or even negligible. Consequently, it would be unlikely that a type 1 gait pattern would negatively impact the health of the spine over time in children with CP. During type 2 (Duchenne type gait), the pelvis maintained an almost neutral position while the trunk leaned toward the supporting limb, well outside normal limits. Again, this was consistent with the literature for this type of movement pattern.7 One reported feature of a type 2 pattern is a reduction in loading at the hip.7 As previously mentioned, the hip abductor moment has been suggested to represent the predominant factor in hip joint loading.7,21 Although strength data were not collected in this study, there was evidence of a reduced hip abductor moment, particularly at peak moment values (Fig. 3). Although moment values remained just within normal limits at these points, it would suggest that pathology of the hip abductors may have played a role in this case. Corresponding reactive forces at the lower lumbar spine were slightly raised for children with the type 2 gait pattern, although they remained within normal SD bands. However, an increased lateral bending moment during stance, with 2 distinct peaks occurring at contralateral toe off and heel strike, was evident. This increased moment toward the contralateral side was not altogether surprising. L5-S1 lateral bending moments during CP gait have been previously reported to increase with excessive trunk movement in the coronal plane.12 This would suggest that a type 2 pattern, while it may have the potential to reduce demands at the hip for some participants, could result in greater demands placed on the lumbar spine as the trunk moves into a suitably compensatory position. This type of movement pattern was previously reported as a therapy for unloading the hip joint during gait.13,14 However, our results suggest there may be a negative impact at the lower lumbar spine. Although it is not clear whether this recommended therapy has been used routinely in clinical practice, the promotion of this type of movement pattern as a form of therapy would need to be considered with caution. A type 3 (complex Trendelenburg-Duchenne) gait pattern involving both contralateral pelvic drop and excessive trunk lean was the most common pattern amongst the cohort in this study with just over 40% of participants in this category. To our knowledge, this is the first study to specifically describe trunk and pelvic kinematics associated with this type 3 movement. Reference has been made to a high proportion of people with diplegic CP with both contralateral pelvic drop and ipsilateral trunk lean in a study examining hip deformities in CP.6 However, this was reported by means of observational gait analysis and did not further split the groups according to specific Trendelenburg-Duchenne presentation. We refer to this walking pattern as complex Trendelenburg-Duchenne. It is not strictly accurate to refer to it as a combination of Trendelenburg and Duchenne, as in Trendelenburg type gait (type 1), the contralateral pelvis dips, while in Duchenne type gait (type 2), it remains level or elevated.6,7 With this in mind, another descriptive name should be applied, such as complex Trendelenburg-Duchenne gait (type 3). When the kinematic presentation of this type 3 pattern was considered, a similar pelvic position to the type 1 pattern was demonstrated. However, as a consequence of the pelvic obliquity, trunk side flexion with respect to the pelvis was therefore increased compared with the type 2 pattern where the pelvis remained flat. Peak trunk side flexion in the global frame was also marginally increased compared with the type 2 pattern, suggesting that children with CP and a type 3 gait pattern may have had slightly more involvement of the trunk than those with a type 2 gait pattern. For all 3 movement types the resultant effects at the lower spine were most evident for type 3. Similar to the type 1 pattern, there was an increased force toward the contralateral side at initial swing. However, this force was greater in magnitude compared with type 1. As pelvic position was similar, the increase in force must therefore be attributed to the increased side flexion of the trunk. Peak lateral bending moments were also increased compared with TD and marginally increased compared with those in children with the type 2 pattern. Consequently, it would appear that greater levels of loading were placed on the lower lumbar spine as the trunk and pelvis maintained suitably compensatory positions during type 3 gait compared with TD. Due to only a small increase in L5-S1 moments compared with the type 2 pattern, our results suggest that the contralateral drop of the pelvis (Trendelenburg component) has a relatively small impact on demands at the lower spine. Instead, it is primarily the excessive trunk lean (Duchenne type component) that contributes mainly to increased lower lumbar spinal loading in children who had CP and a type 3 pattern during gait. It is difficult to identify exactly what the impact of pathological reactive forces and moments will be at the spine over time. Although low back pain has been reported in adults with CP,15 this has not been recorded in children with CP. However, radiographic studies have highlighted changes in the lumbar spine of children with CP.22 The Wolff law states that bone tissue will respond by deposition of new bone in response to cyclic high stresses,23 and it has been suggested that pathological changes in the mechanics of motion of the spine can result in the formation of osteophytes along the junction of the vertebral bodies and intervertebral disks.24 Based on this, it is a reasonable assumption that altered trunk movement will have some impact on the health of the spine over time. However, this assumption cannot be concluded from the results of this study. No study to date has specifically examined lumbar stresses and strains specifically related to altered trunk motion during CP gait and the long-term consequences of this altered movement. As a starting point, this investigation provides a foundation and the methods for further exploration. A longitudinal study of low back pain in CP from childhood into adulthood, where reactive forces and moments at the spine are assessed at regular intervals, and an investigation of bone-on-bone forces possibly by means of a forward dynamics approach, may shed further light as to the impact of Trendelenburg and Duchenne type movements over time during CP gait. Based on the results of this study, and taking into account the previously reported moderate to strong relationship between thorax movement and lower lumbar spinal loading,12 it is suggested that therapy is aimed at reducing thorax movement in children with a type 2 (Duchenne) or type 3 (complex Trendelenburg-Duchenne) pattern. However, altered thorax movement is usually a compensation for a lower limb deficiency (eg, weak hip abductors or dysfunction of the affected hip).10 As a result, the optimal thorax lean for the best tradeoff between minimizing the effects of lower limb problems and reducing reactive forces and moments at the spine is unknown. Hip abductor muscle strengthening may be beneficial to reduce excessive mediolateral thorax sway and thus reduce reactive forces and moments at the spine. Alternatively, adjustable walking poles have been shown to result in less lateral trunk lean.25 However, as discussed, more investigation is warranted as to the long-term effects of increased reactive forces and moments at the spine during gait in people with CP. Limitations A number of limitations exist with this study. First, data were collected at a self-selected walking speed. Walking speed has been shown to impact on both kinematics and kinetics.26 However, no attempt was made to control walking speed as the aim was to report patterns characteristic of CP gait. Second, the number of children with the type 2 pattern was small, and results relating to this pattern need to be interpreted with caution. However, as it was the purpose of the study to also examine the effects of a type 2 pattern, descriptive statistics and ensemble average profiles were discussed. Third, a test-retest reliability analysis or measure of minimal detectable change (MDC) was not conducted as part of this study. Results suggest that measures of MDC may be high for some of the reported variables and, while a preliminary assessment of reliability demonstrated good results, a further assessment of reliability is planned. Finally, the inverse dynamic approach only provides the net inter-segmental reactive forces that reflect the effect of external loads. Previous studies have shown that spinal loads can be larger than the net inter-segmental force.27,28 As excessive cocontraction of muscles can occur during CP gait,29 future work incorporating studies of trunk electromyography may be required. Conclusions The majority of children with CP in our cohort had a type 3 (complex Trendelenburg-Duchenne) walking pattern. To our knowledge, this is the first study to specifically describe the associated trunk and pelvic kinematics of this pattern, and we believe that the phrase complex Trendelenburg-Duchenne appears to be suitable for describing this pattern. Additionally, children who had CP and this type 3 pattern demonstrated the greatest increase in kinetics at the lumbar spine compared with TD controls. Trunk position was a critical factor with only a small contribution as a result of the contralateral drop of the pelvis. For children with CP and a type 2 (Duchenne) pattern, kinetics at the lumbar spine were increased. However, although the position of the trunk was also important, numbers in this group were low and only descriptive statistics reported. Finally, the type 1 (Trendelenburg) gait pattern had the least impact at the lower lumbar spine with only small differences in lumbar spinal kinetics compared with TD controls. Consequently, clinicians need to be aware of the resultant effects at the lower spine when presented with these types of walking patterns. Specifically, promotion of a type 2 or type 3 pattern as a noninvasive rehabilitative or conservative treatment should be considered with caution due to the potential for negative effects at the lumbar spine. Author Contributions and Acknowledgments Concept/idea/research design: D. Kiernan, R. O’Sullivan, A. Malone, T. O’Brien, C.K. Simms Writing: D. Kiernan, R. O’Sullivan, A. Malone, C.K. Simms Data collection: D. Kiernan, A. Malone Data analysis: D. Kiernan, R. O’Sullivan, A. Malone, C.K. Simms Project management: D. Kiernan Fund procurement: D. Kiernan Providing participants: D. Kiernan, R. O’Sullivan Consultation (including review of manuscript before submitting): D. Kiernan, R. O’Sullivan, A. Malone, T. O’Brien, C.K. Simms Each of the authors has read and concurs with the content in the final manuscript. The material within has not been and will not be submitted for publication elsewhere except as an abstract. The material within constitutes part of a PhD thesis written by the corresponding author. Ethics Approval Ethics approval was granted by the Central Remedial Clinic's ethics committee. Funding Funding was provided in part by the Scientific and Research Trust of the Central Remedial Clinic. Disclosures The authors completed the ICJME Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest. References 1 Attias M, Bonnefoy-Mazure A, Lempereur M, Lascombes P, De Coulon G Armand S. Trunk movements during gait in cerebral palsy. Clin Biomech (Bristol, Avon) . 2015; 30: 28– 32. Google Scholar CrossRef Search ADS PubMed  2 Heyrman L, Feys H, Molenaers G et al.   Three-dimensional head and trunk movement characteristics during gait in children with spastic diplegia. Gait Posture . 2013; 38: 770– 776. Google Scholar CrossRef Search ADS PubMed  3 Romkes J, Peeters W, Oosterom A, Molenaar S, Bakels I, Brunner R. Evaluating upper body movements during gait in healthy children and children with diplegic cerebral palsy. J Pediatr Orthop B . 2007; 16: 175– 180. Google Scholar CrossRef Search ADS PubMed  4 O’Sullivan R, Walsh M, Jenkinson A, O’Brien T. Factors associated with pelvic retraction during gait in cerebral palsy. Gait Posture . 2007; 25: 425– 431. Google Scholar CrossRef Search ADS PubMed  5 Salazar-Torres JJ, McDowell BC, Kerr C, Cosgrove AP. Pelvic kinematics and their relationship to gait type in hemiplegic cerebral palsy. Gait Posture . 2011; 33: 620– 624. Google Scholar CrossRef Search ADS PubMed  6 Metaxiotis D, Accles W, Siebel A, Doederlein L. Hip deformities in walking patients with cerebral palsy. Gait Posture . 2000; 11: 86– 91. Google Scholar CrossRef Search ADS PubMed  7 Westhoff B, Petermann A, Hirsch MA, Willers R, Krauspe R. Computerized gait analysis in Legg Calve Perthes disease: analysis of the frontal plane. Gait Posture . 2006; 24: 196– 202. Google Scholar CrossRef Search ADS PubMed  8 Trendelenburg F. Ueber den Gang bei angeborener Huftgelenksluxation [On the gait of people with congenital dislocation of the hip]. Dtsch Med Wochenschr . 1895; 2: 21– 24. Google Scholar CrossRef Search ADS   9 Duchenne G. Selections from the clinical works of Dr. Duchenne (de Boulogne). Poore GV, ed. London, England: The New Sydenham Society; 1883. 10 Stief F, Bohm H, Ebert C, Doderlein L, Meurer A. Effect of compensatory trunk movements on knee and hip joint loading during gait in children with different orthopedic pathologies. Gait Posture . 2014; 39: 859– 864. Google Scholar CrossRef Search ADS PubMed  11 Krautwurst BK, Wolf SI, Heitzmann DW, Gantz S, Braatz F, Dreher T. The influence of hip abductor weakness on frontal plane motion of the trunk and pelvis in patients with cerebral palsy. Res Dev Disabil . 2013; 34: 1198– 1203. Google Scholar CrossRef Search ADS PubMed  12 Kiernan D, Malone A, O’Brien T, Simms CK. Children with cerebral palsy experience greater levels of loading at the low back during gait compared to healthy controls. Gait Posture . 2016; 48: 249– 255. Google Scholar CrossRef Search ADS PubMed  13 Schröter J, Güth V, Overbeck M, Rosenbaum D, Winkelmann W. The ‘Entlastungsgang’: a hip unloading gait as a new conservative therapy for hip pain in the adult. Gait Posture . 1999; 9: 151– 157. Google Scholar CrossRef Search ADS PubMed  14 Svehlik M, Kraus T, Steinwender G, Zwick EB, Linhart WE. Pathological gait in children with Legg-Calve-Perthes disease and proposal for gait modification to decrease the hip joint loading. Int Orthop . 2012; 36: 1235– 1241. Google Scholar CrossRef Search ADS PubMed  15 Jahnsen R, Villien L, Aamodt G, Stanghelle JK, Holm I. Musculoskeletal pain in adults with cerebral palsy compared with the general population. J Rehabil Med . 2004; 36: 78– 84. Google Scholar CrossRef Search ADS PubMed  16 Bottos M, Feliciangeli A, Sciuto L, Gericke C, Vianello A. Functional status of adults with cerebral palsy and implications for treatment of children. Dev Med Child Neurol . 2001; 43: 516– 528. Google Scholar CrossRef Search ADS PubMed  17 Schwartz L, Engel JM, Jensen MP. Pain in persons with cerebral palsy. Arch Phys Med Rehabil . 1999; 80: 1243– 1246. Google Scholar CrossRef Search ADS PubMed  18 Kiernan D, Hosking J, O’Brien T. Is adult gait less susceptible than paediatric gait to hip joint centre regression equation error? Gait Posture . 2016; 45: 133– 136. Google Scholar CrossRef Search ADS PubMed  19 Kiernan D, Malone A, O’Brien T, Simms CK. A 3-dimensional rigid cluster thorax model for kinematic measurements during gait. J Biomech . 2014; 47: 1499– 1505. Google Scholar CrossRef Search ADS PubMed  20 Seay J, Selbie WS, Hamill J. In vivo lumbo-sacral forces and moments during constant speed running at different stride lengths. J Sports Sci . 2008; 26: 1519– 1529. Google Scholar CrossRef Search ADS PubMed  21 Brinckmann P, Frobin W, Hierholzer E. Stress on the articular surface of the hip joint in healthy adults and persons with idiopathic osteoarthrosis of the hip joint. J Biomech . 1981; 14: 149– 156. Google Scholar CrossRef Search ADS PubMed  22 Harada T, Ebara S, Anwar MM et al.   The lumbar spine in spastic diplegia: a radiographic study. J Bone Joint Surg Br . 1993; 75: 534– 537. Google Scholar PubMed  23 Wolff J. Das Gesetz der Transformation der Knoche . Berlin, Germany: Hirschwald; 1892. 24 Bogduk N, Twomey LT. Clinical Anatomy of the Lumbar Spine . London, United Kingdom: Churchill Livingstone; 
 1991. 25 Bechard DJ, Birmingham TB, Zecevic AA et al.   The effect of walking poles on the knee adduction moment in patients with varus gonarthrosis. Osteoarthritis Cartilage . 2012; 20: 1500– 1506. Google Scholar CrossRef Search ADS PubMed  26 Schwartz MH, Rozumalski A, Trost JP. The effect of walking speed on the gait of typically developing children. J Biomech . 2008; 41: 1639– 1650. Google Scholar CrossRef Search ADS PubMed  27 Granata KP, Marras WS. The influence of trunk muscle coactivity on dynamic spinal loads. Spine (Phila Pa 1976) . 1995; 20: 913– 919. Google Scholar CrossRef Search ADS PubMed  28 Paul JP. Paper 8: forces transmitted by joints in the human body. Proceedings of the Institution of Mechanical Engineers, Conference Proceedings . 1966; 181: 8– 15. Google Scholar CrossRef Search ADS   29 Gage JR, Stout JL. Gait analysis: kinematics, kinetics, electromyography, oxygen consumption and pedobarography. In: Gage JR, Schwartz MH, Koop SE, Novacheck TF, eds. The Identification and Treatment of Gait Problems in Cerebral Palsy . London, United Kingdom: Mac Keith Press; 2009: 260– 284. © 2017 American Physical Therapy Association http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Physical Therapy Oxford University Press

Pathological Movements of the Pelvis and Trunk During Gait in Children With Cerebral Palsy: A Cross-Sectional Study With 3-Dimensional Kinematics and Lower Lumbar Spinal Loading

Loading next page...
 
/lp/ou_press/pathological-movements-of-the-pelvis-and-trunk-during-gait-in-children-Yp3lpnK0h8
Publisher
American Physical Therapy Association
Copyright
© 2017 American Physical Therapy Association
ISSN
0031-9023
eISSN
1538-6724
D.O.I.
10.1093/ptj/pzx113
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Increased loading at the lumbar spine, particularly in the coronal plane, has been reported in children with cerebral palsy (CP). As pelvic and trunk movements associated with Trendelenburg and Duchenne type gait are most significant in the coronal plane, the potential exists for lower lumbar spinal loading to be negatively affected in children with CP and these types of movement patterns. Objective The objective of this study was to assess trunk and pelvic kinematics and lower lumbar spinal loading patterns in children with CP and Trendelenburg and Duchenne type gait. Design This was a cross-sectional study. Methods Three-dimensional kinematic (lower limb and thorax) and L5-S1 kinetic data were recorded. Children were divided according to clinical presentation of Trendelenburg or Duchenne type gait. Several discrete kinematic and kinetic parameters were assessed between groups. Results Three distinct pelvic and trunk movement patterns were identified for children with CP: Trendelenburg, Duchenne, and complex Trendelenburg–Duchenne. Peak L5-S1 lateral bending moments were increased by 62% in children with CP and Duchenne type gait. Children with CP and complex Trendelenburg-Duchenne gait demonstrated the largest deviations from normal, with increased peak ipsilateral and contralateral directed moments of 69% and 54%, respectively, compared with children with typical development. Limitations A test-retest reliability analysis or measure of minimal detectable change was not conducted as part of this study. Results suggest that measures of minimal detectable change may be high for some of the reported variables. In addition, the inverse dynamic approach determines only the net intersegmental reactive forces that reflect the effect of external loads. Previous studies have shown that spinal loads may be larger than the net intersegmental force. Conclusions Trendelenburg and Duchenne type movements were not always distinct, and a third type of movement, a combination of the two, was the most common in this study. Clinicians should be aware that children with CP and the Duchenne type or the complex Trendelenburg-Duchenne type of gait pattern experience abnormal loading that may have significant implications for the lower spine in the long term. Excessive movement of the trunk and pelvis are characteristic of walking in children with cerebral palsy (CP).1–5 In particular, Trendelenburg and Duchenne type gait patterns are often seen in this group.6 A Trendelenburg type gait pattern is characterized by a drop of the pelvis in the coronal plane on the unloaded side during stance.6–8 The trunk is tilted toward the supporting limb with respect to the pelvis while maintaining a neutral position with respect to the global reference frame.7 Hip adduction of the supporting limb is also increased.7 A Duchenne type gait pattern is characterized by a trunk lean toward the supporting limb with the pelvis level or elevated on the unloaded side.6,7,9 This type of compensatory movement of the trunk moves body weight toward the center of the hip resulting in a reduced hip abductor moment.6,10 Both movement patterns are often a compensatory mechanism for hip abductor weakness or hip dysfunction of the supporting limb.6,10,11 However, a Duchenne type gait pattern may have negative effects on the knee joint by increasing the lever arm around the knee.10 Additionally, it has been suggested that a Duchenne type gait pattern may result in spinal problems.6 Likewise, a Trendelenburg type gait pattern is thought to be harmful to the hip.6 However, while the effects of Trendelenburg and Duchenne type walking patterns on kinetics at the knee and hip have been considered,6,7,10 the effects of this movement further up the kinematic chain, particularly at the lower spine, are unknown. Our group have recently reported increased reactive forces and moments at the lower lumbar spine during gait in children with CP.12 As thorax coronal plane movement became more excessive, children with CP demonstrated increased peak L5-S1 reactive forces and moments of up to 63% and 90%, respectively.12 Consequently, as Trendelenburg and Duchenne type gait patterns are most significant in the coronal plane, we questioned whether loading at the lumbar spine would be adversely affected in children with CP when walking with these movement patterns? As the position of the trunk changes with respect to the pelvis during Trendelenburg and Duchenne type gait, mechanical loads at the spine must change. However, the extent to which this occurs is unclear. A thorough understanding of any potential negative effects at the spine is important because Duchenne type gait patterns were previously reported to be protective of the hip joint by increasing the cover of the femoral head.6 In addition, this type of walking pattern has also been recommended as a noninvasive intervention for hip pain in adults and as a conservative treatment in children with Legg-Calvé-Perthes disease.13,14 However, while it is not clear whether these recommendations are followed routinely in clinical practice, the potential for increased demands at the spine may make such recommendations counterproductive. Additionally, excessive spinal tilting as a consequence of these types of gait patterns may further increase the risk of low back pain in a population with CP, in whom incidents of low back pain were previously reported to be up to 63% higher than those in controls.15–17 Therefore, it is important that clinicians are fully aware of any adverse effects that may occur at the lower lumbar spine when presented with Trendelenburg or Duchenne type gait patterns. Following from this, the objective of this study was to assess trunk and pelvic kinematics and three-dimensional reactive forces and moments at the lower lumbar spine in children with CP who appeared on clinical presentation to have Trendelenburg or Duchenne type patterns of movement during gait. The cohort of children, both those with CP and those with typical development (TD), was that used in our previous study.12 Children with CP were regrouped and reanalyzed according to clinical presentation of Trendelenburg or Duchenne type gait. Methods Study Design This was a cross-sectional study for both children with CP and children with TD. The study is a secondary analysis of previously published data by our group.12 Participants were regrouped and the data reanalyzed according to newly defined groups. Participants A sample size calculation was performed as part of an earlier study.12 In that study, coronal plane trunk flexion was used as the primary outcome measure; the hypothesis being tested using coronal plane trunk flexion was that coronal plane trunk flexion outside 1 standard deviation from the normal group was pathological. In the present study, using coronal plane trunk flexion as the primary outcome measure, a sample size of 26 was required at a power of 0.95 and a significance of 0.05. This number was exceeded, and 52 children with CP participated in this study. Children with CP were recruited from a cohort attending the gait analysis laboratory over a period of 9 months (n = 52; 33 boys, 19 girls; mean age = 11.02 years [SD = 3.01]; 21 with hemiplegia, 31 with diplegia). Inclusion criteria were: diagnosis of hemiplegic or diplegic CP and ability to walk independently. Children were excluded if they had had surgery within 1 year of presenting to the gait laboratory. Children with TD were invited to participate after a general email to staff and colleagues at the host institution (n = 26; 15 boys, 11 girls; mean age = 10.15 years [SD = 3.17] years). Children with TD without any previous history of neurological, musculoskeletal, or orthopedic problems were included in the study. A participant information leaflet was provided to parents and guardians, who then gave written informed consent. Data Collection A full barefoot 3-dimensional kinematic and kinetic analysis was performed using the CODA cx1 active marker system (Charnwood Dynamics Ltd, Leicestershire, United Kingdom) and 2 Kistler 9281B (Kistler Instruments Ltd., 13 Murrell Green Business Park, London Road, RG27 9GR Hook, Hampshire, United Kingdom) and 2 AMTI Accugait force platforms (176 Waltham Street, Watertown, MA). Data were collected using Codamotion ODIN software (v1.06 Build 01 09) at capture rates of 100 Hz (kinematic) and 200 Hz (kinetic), respectively. Infrared light-emitting diodes (LEDs) were placed on the lower limbs in accordance with a modified Helen Hayes model.18 Thorax kinematic data were captured using a single cluster protocol shown to be valid and comparable to other protocols for measuring thorax kinematics.19 This protocol consisted of a rigid mount containing LEDs and was placed at spinal level T3.19 For L5-S1 joint location, a LED was placed at the joint space of L5-S1. A virtual point was created at a position corresponding to 5% of the distance of the line between the L5-S1 LED and the midpoint of the anterior superior iliac spine.12,20 L5-S1 reactive forces and moments were then realized at this point. Participants walked unassisted at a self-selected pace with at least 4 clean data trials recorded per participant. One representative walking trial was chosen and analyzed for each participant according to standard laboratory protocol. Coronal Plane Pattern Assessment In order to determine pattern type, coronal plane kinematic data of all participants were visually analyzed and compared to TD data. For the purposes of this study, normal data were defined as ±1 standard deviation about average. Values outside this range were considered abnormal. Three different movement types were observed. Type 1 was defined by a pelvic drop on the unloaded side during stance phase outside the ±1 standard deviation band with a laterally flexed trunk position toward the supporting limb in relation to the pelvis and an upright trunk position in relation to the global reference frame (referred to as a Trendelenburg type gait pattern)7,14 (Fig. 1b). Eight children (15.4%; 2 with hemiplegia, 6 with diplegia) demonstrated a type 1 gait pattern. Type 2 was defined by a level or elevated pelvis on the unloaded side in conjunction with a trunk lean toward the supporting limb outside the ±1 standard deviation band (referred to as a Duchenne type gait pattern)7,14 (Fig. 1c). Four children (7.7%; 1 with hemiplegia, 3 with diplegia) demonstrated a type 2 gait pattern. Finally, type 3 was defined by a distinctive pelvic drop on the unloaded side during stance in conjunction with an excessive lean of the trunk outside TD limits toward the supporting limb with respect to both the pelvis and the global reference frame (referred to as the complex Trendelenburg-Duchenne type gait pattern) (Fig. 1d). Twenty-two children (42.3%; 1 with hemiplegia, 21 with diplegia) demonstrated a type 3 pattern. Four children (7.7%; 3 with hemiplegia, 1 with diplegia) demonstrated trunk and pelvic kinematic patterns within normal ranges. The remaining 14 children (26.9%; 14 with hemiplegia) demonstrated patterns not consistent with Trendelenburg or Duchenne type gait. Children who demonstrated patterns within normal ranges or not consistent with Trendelenburg or Duchenne type gait were excluded from further evaluation. Children were then grouped according to movement pattern types 1 to 3. Figure 1. View largeDownload slide Movement patterns in the coronal plane. (a) Normal trunk (T) and pelvis (P) positions. (b) Trendelenburg type pattern. Note the upright position of the trunk in the global frame and the significant drop of the pelvis on the contralateral side. (c) Duchenne type pattern. Note the lateral lean of the trunk in the global frame and the level (or slightly raised on the contralateral side) position of the pelvis. (d) Complex Trendelenburg-Duchenne pattern. Note the lateral lean of the trunk in the global frame and the drop of the pelvis on the contralateral side. Figure 1. View largeDownload slide Movement patterns in the coronal plane. (a) Normal trunk (T) and pelvis (P) positions. (b) Trendelenburg type pattern. Note the upright position of the trunk in the global frame and the significant drop of the pelvis on the contralateral side. (c) Duchenne type pattern. Note the lateral lean of the trunk in the global frame and the level (or slightly raised on the contralateral side) position of the pelvis. (d) Complex Trendelenburg-Duchenne pattern. Note the lateral lean of the trunk in the global frame and the drop of the pelvis on the contralateral side. Data Analysis A number of discrete kinematic and kinetic parameters were assessed between groups. The 4 kinematic parameters were as follows: maximum pelvic obliquity in stance, maximum trunk side flexion in stance (with respect to the global or laboratory reference frame), maximum trunk side flexion stance (with respect to the pelvis), and maximum hip adduction in stance. Trunk side flexion was measured with respect to the pelvis and with respect to the global or laboratory reference frame. The 6 kinetic parameters were as follows: hip abductor moment peak 1 (N·m/kg), hip abductor moment peak 2 (N·m/kg), peak L5-S1 medial/lateral force directed toward the ipsilateral side (N/kg), peak L5-S1 medial/lateral force directed toward the contralateral side (N/kg), peak L5-S1 lateral bending moment directed toward the ipsilateral side (N·m/kg), and peak L5-S1 lateral bending moment directed toward the contralateral side (N·m/kg). The analyzed side was the side to which the trunk leaned during stance. This was then referred to as “ipsilateral” for the duration of the gait cycle. The word “contralateral” refers to the opposite side for the same period of the gait cycle. Data were checked for distribution using the Shapiro-Wilk normality test. Differences between movement types and children with TD were assessed using a 1-way analysis of variance with Bonferroni post hoc tests for comparisons between groups. Dunnett tests were also used to compare each movement type with the TD group. For data that did not follow a normal distribution, differences were assessed using a Kruskal-Wallis test and post hoc Mann-Whitney U tests. All statistical analyses were performed using IBM SPSS Statistics (v23.0.0.2). The level of significance was set at 0.05. Additionally, ensemble average kinematic and kinetic profiles were visually analyzed for deviations between groups. Role of the Funding Source Funding was provided in part by the Scientific and Research Trust of the Central Remedial Clinic. The funder had no role in the study design or in the collection, analysis, or interpretation of the data. Results Kinematic Classification After coronal plane classification, 34 children with CP were assessed. Eight children (7 boys and 1 girl; mean age = 10.9 years [SD = 2.9]; mean height = 1.44 m [SD = 0.12]; mean weight = 38.04 kg [SD = 12.5]) had type 1 gait. Four children (3 boys and 1 girl; mean age = 12.0 years [SD = 2.9]; mean height = 1.52 m [SD = 0.18]; mean weight = 41.03 kg [SD = 20.2]) had type 2 gait. Twenty-two children (14 boys and 8 girls; mean age = 10.2 years [SD = 3.0]; mean height = 1.41 m [SD = 0.2]; mean weight = 37.2 kg [SD = 14.6]) had type 3 gait. Type 1 (Trendelenburg Gait Pattern) Children with a type 1 gait pattern demonstrated an increased peak pelvic obliquity in stance by 4.8 degrees compared with that in children with TD (Fig. 2; Tabs. 1 and 2). Peak trunk side flexion (with respect to the pelvis) was increased by 7.5 degrees, while peak trunk side flexion (with respect to the global or laboratory reference frame) remained within normal limits throughout the gait cycle. Peak hip adduction in stance was increased by 5.1 degrees compared with that in children with TD. Hip abductor moment remained close to that in children TD, with no statistically significant differences (Tabs. 1 and 2). Figure 2. View largeDownload slide Kinematic and kinetic ensemble average profiles for the Trendelenburg (type 1) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Figure 2. View largeDownload slide Kinematic and kinetic ensemble average profiles for the Trendelenburg (type 1) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Table 1. Values for Children With Typical Development (TD) and Children With Cerebral Palsy (CP) by Movement Typea Parameter  Children With TD  Children With CP  Concurrent 
P Value 
Determined by ANOVA  Post Hoc Comparisonc  Type 1 
(n = 8)  Type 2 
(n = 4)  Type 3 
(n = 22)  Peak pelvic obliquity stance (°)  5.42 (2.3)  10.2 (3.2)  2.27 (0.86)  9.87 (2.7)  < .001  A; B  Peak hip adduction in stance (°)b  6.90 (19.7)  12.0 (10.4)  6.27 (7.3)  9.83 (24.1)  .02  A; B  Peak trunk side flexion (with respect to global or laboratory reference frame) (°)  2.29 (4.0)  4.40 (2.2)  9.25 (2.5)  10.87 (3.7)  < .001  B; C  Peak trunk side flexion (with respect to pelvis) (°)  6.47 (4.6)  13.9 (4.1)  12.0 (2.9)  19.35 (5.8)  < .001    Peak L5-S1 moment toward contralateral (N·m/kg)b  −0.29 (0.41)  −0.37 (0.41)  −0.47 (0.7)  −0.49 (0.8)  < .001  A; B  Peak L5-S1 moment toward ipsilateral (N·m/kg)b  0.24 (0.56)  0.33 (0.33)  0.32 (0.35)  0.37 (1.12)  .01  B  Hip abductor moment peak 1 (N·m/kg)  0.53 (0.14)  0.56 (0.14)  0.58 (0.64)  0.41 (0.17)  .11  B  Hip abductor moment peak 2 (N·m/kg)  0.46 (0.14)  0.34 (0.24)  0.50 (0.26)  0.25 (0.23)  .00  B  Peak L5-S1 force toward contralateral (N/kg)b  −0.53 (1.42)  −0.88 (0.92)  −0.79 (0.44)  −1.1 (2.03)  .00  A; B  Peak L5-S1 force toward ipsilateral (N/kg)b  0.55 (1.48)  0.32 (0.88)  0.70 (0.29)  0.61 (1.65)  .55    Parameter  Children With TD  Children With CP  Concurrent 
P Value 
Determined by ANOVA  Post Hoc Comparisonc  Type 1 
(n = 8)  Type 2 
(n = 4)  Type 3 
(n = 22)  Peak pelvic obliquity stance (°)  5.42 (2.3)  10.2 (3.2)  2.27 (0.86)  9.87 (2.7)  < .001  A; B  Peak hip adduction in stance (°)b  6.90 (19.7)  12.0 (10.4)  6.27 (7.3)  9.83 (24.1)  .02  A; B  Peak trunk side flexion (with respect to global or laboratory reference frame) (°)  2.29 (4.0)  4.40 (2.2)  9.25 (2.5)  10.87 (3.7)  < .001  B; C  Peak trunk side flexion (with respect to pelvis) (°)  6.47 (4.6)  13.9 (4.1)  12.0 (2.9)  19.35 (5.8)  < .001    Peak L5-S1 moment toward contralateral (N·m/kg)b  −0.29 (0.41)  −0.37 (0.41)  −0.47 (0.7)  −0.49 (0.8)  < .001  A; B  Peak L5-S1 moment toward ipsilateral (N·m/kg)b  0.24 (0.56)  0.33 (0.33)  0.32 (0.35)  0.37 (1.12)  .01  B  Hip abductor moment peak 1 (N·m/kg)  0.53 (0.14)  0.56 (0.14)  0.58 (0.64)  0.41 (0.17)  .11  B  Hip abductor moment peak 2 (N·m/kg)  0.46 (0.14)  0.34 (0.24)  0.50 (0.26)  0.25 (0.23)  .00  B  Peak L5-S1 force toward contralateral (N/kg)b  −0.53 (1.42)  −0.88 (0.92)  −0.79 (0.44)  −1.1 (2.03)  .00  A; B  Peak L5-S1 force toward ipsilateral (N/kg)b  0.55 (1.48)  0.32 (0.88)  0.70 (0.29)  0.61 (1.65)  .55    aData are reported as mean (standard deviation) unless otherwise indicated. Descriptive statistics are reported for type 2. ANOVA = analysis of variance, type 1 = Trendelenburg only, type 2 = Duchenne only, type 3 = Trendelenburg-Duchenne. bMedian (range) for nonnormally distributed data, with distribution of all data assessed using a Shapiro-Wilk normality test. cResults of post hoc tests are indicated as follows: A = type 1-TD, B = type 3-TD, C = type 1-type 3. A breakdown of post hoc tests is reported in Table 2. View Large Table 2. Results of Post Hoc Analysis for Differences and 95% CIs Between Groupsa Parameter  Difference Between Type 1 
and TD  95% CI  Con-current 
P Value  Difference Between Type 3 and TD  95% CI  Con-current 
P Value  Difference Between Type 1 and 
Type 3  95% CI  Con-current 
P Value  Peak pelvic obliquity stance (°)  4.78  2.28 to 7.28  .00  4.46  2.66 to 6.25  .00  0.33  −2.53 to 3.18  1.00  Peak hip adduction in stance (°)b  5.10  −1.46 to 5.69  .00  2.93  6.02 to 11.1  .03  2.17    .30  Peak trunk side flexion (with respect to global or laboratory reference frame) (°)  2.11    .38  8.58  9.38 to 16.4  .00  6.47  2.39 to 10.55  .00  Peak trunk side flexion (with respect to pelvis) (°)  7.48  2.58 to 12.38  .00  12.89    .00  −5.40  −11.0 to 0.19  .06  Peak L5-S1 moment toward contralateral (N·m/kg)b  −0.08    .14  −0.20    .00  0.12    .08  Peak L5-S1 moment toward ipsilateral (N·m/kg)b  0.09    .33  0.13    .01  0.04    .37  Hip abductor moment peak 1 (N·m/kg)  0.03  −0.17 to 0.24  .96  −0.12  −0.27 to 0.03  .13  0.16  −0.08 to 0.39  .45  Hip abductor moment peak 2 (N·m/kg)  −0.12  −0.31 to 0.08  .35  −0.21  −0.35 to –0.07  .00  0.09  −0.13 to 0.31  1.00  Peak L5-S1 force toward contralateral (N/kg)b  −0.35    .03  −0.57    .00  0.22    .48  Peak L5-S1 force toward ipsilateral (N/kg)b  −0.23    .38  0.06    .53  0.29    .92  Parameter  Difference Between Type 1 
and TD  95% CI  Con-current 
P Value  Difference Between Type 3 and TD  95% CI  Con-current 
P Value  Difference Between Type 1 and 
Type 3  95% CI  Con-current 
P Value  Peak pelvic obliquity stance (°)  4.78  2.28 to 7.28  .00  4.46  2.66 to 6.25  .00  0.33  −2.53 to 3.18  1.00  Peak hip adduction in stance (°)b  5.10  −1.46 to 5.69  .00  2.93  6.02 to 11.1  .03  2.17    .30  Peak trunk side flexion (with respect to global or laboratory reference frame) (°)  2.11    .38  8.58  9.38 to 16.4  .00  6.47  2.39 to 10.55  .00  Peak trunk side flexion (with respect to pelvis) (°)  7.48  2.58 to 12.38  .00  12.89    .00  −5.40  −11.0 to 0.19  .06  Peak L5-S1 moment toward contralateral (N·m/kg)b  −0.08    .14  −0.20    .00  0.12    .08  Peak L5-S1 moment toward ipsilateral (N·m/kg)b  0.09    .33  0.13    .01  0.04    .37  Hip abductor moment peak 1 (N·m/kg)  0.03  −0.17 to 0.24  .96  −0.12  −0.27 to 0.03  .13  0.16  −0.08 to 0.39  .45  Hip abductor moment peak 2 (N·m/kg)  −0.12  −0.31 to 0.08  .35  −0.21  −0.35 to –0.07  .00  0.09  −0.13 to 0.31  1.00  Peak L5-S1 force toward contralateral (N/kg)b  −0.35    .03  −0.57    .00  0.22    .48  Peak L5-S1 force toward ipsilateral (N/kg)b  −0.23    .38  0.06    .53  0.29    .92  aTD = typical development, type 1 = Trendelenburg only, type 3 = Trendelenburg-Duchenne. bNonnormally distributed data, for which 95% CIs were not reported. View Large L5-S1 ensemble average kinetic profiles were similar for children with a type 1 gait pattern and children with TD (Fig. 2). However, children with a type 1 gait pattern demonstrated an increased peak L5-S1 force directed toward the contralateral limb at initial swing phase (increase of 0.35 N/kg or ≈66%). No other statistically significant differences were present for medial lateral force or lateral bending moment for children with a type 1 gait pattern compared with that in children with TD (Tabs. 1 and 2). Type 2 (Duchenne Gait Pattern) The number of children demonstrating a type 2 gait pattern was too small for statistical analysis (n = 4). As an alternative, descriptive statistics are provided and ensemble average profiles commented on. For children with a type 2 gait pattern, pelvic obliquity remained relatively flat and close to a neutral position throughout the gait cycle (Fig. 3; Tab. 1). Peak trunk side flexion (with respect to the pelvis) demonstrated an increase of 5.5 degrees compared with that in children with TD. In addition, peak trunk side flexion (with respect to the global or laboratory reference frame) was also increased compared with that in children with TD, by 7.0 degrees (Tab. 1). Hip abductor moment was slightly reduced at peak points but remained just within normal limits (Fig. 3). Figure 3. View largeDownload slide Kinematic and kinetic ensemble average profiles for the Duchenne (type 2) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Figure 3. View largeDownload slide Kinematic and kinetic ensemble average profiles for the Duchenne (type 2) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Kinetic ensemble average profiles demonstrated an increased lateral bending moment toward the contralateral side outside normal limits during stance for children with a type 2 gait pattern compared with that in children with TD (Fig. 3). This peak moment was up to 62% greater than that in children with TD (Tab. 1). Lateral bending moment toward the ipsilateral side during swing remained within normal limits (Fig. 3; Tab. 1). Medial lateral force also remained within normal limits throughout the gait cycle (Fig. 3). Type 3 (Complex Trendelenburg-Duchenne Gait Pattern) For children with a type 3 gait pattern, peak pelvic obliquity was significantly increased by 4.5 degrees compared with that in children with TD (Fig. 4; Tabs. 1 and 2). A similar increase was demonstrated when compared with the type 2 gait pattern (Tab. 1). Peak trunk side flexion, with respect to both the pelvis and the global or laboratory reference frame, was significantly increased by 12.9 and 8.6 degrees, respectively, compared with the values in children with TD (Tabs. 1 and 2). Peak hip adduction in stance demonstrated a statistically significant increase of 2.9 degrees, while hip abductor moment peak 1 remained within normal limits. However, hip abductor peak 2 was significantly decreased for children with a type 3 gait pattern compared with that for children with TD (reduction of 0.21 N·m/kg or ≈54%) (Tabs. 1 and 2). Figure 4. View largeDownload slide Kinematic and kinetic ensemble average profiles for the complex Trendelenburg-Duchenne (type 3) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Figure 4. View largeDownload slide Kinematic and kinetic ensemble average profiles for the complex Trendelenburg-Duchenne (type 3) gait pattern. The black line indicates the analyzed side, and the gray shading indicates the ± 1 standard deviation around the average for typical development. Abd = abduction, Ad = adduction, Flex = flexion, Ipsi = ipsilateral, Lab = global or laboratory reference frame, Lat or Lat. = lateral, Med = medial, Nm = newton-meter, Pel = pelvis, Tnk = trunk, w.r.t = with respect to, %GC = percentage gait cycle. Several significant deviations were evident in the kinetic ensemble average profiles for children with a type 3 gait pattern compared with that in children with TD (Fig. 4). Like children with a type 1 gait pattern, children with a type 3 gait pattern demonstrated an increased peak L5-S1 force directed toward the contralateral limb at initial swing phase. However, in this case, peak force was approximately double that in children with TD (increase of 0.57 N/kg or ≈107%) (Tabs. 1 and 2). Both peak L5-S1 ipsilateral and contralateral directed moments were also larger for children with CP than for children with TD (increased by 69% and 54%, respectively) (Tabs. 1 and 2). Discussion Pelvic, trunk, and hip kinematics during type 1 (Trendelenburg gait) in this study were consistent with the literature when describing this type of movement pattern.7 Hip abductor moment remained within normal limits as expected.7 Consequently, as hip abductor moment represents the predominant factor in hip joint loading,7,21 loading at the hip can be judged to be close to normal. The resulting effects at the lumbar spine were small. Only mild differences were present between the CP and TD groups. Most notably, at the point of ipsilateral toe off, medial/lateral force directed toward the contralateral side was increased. At this point, the pelvis was raised on the contralateral side, the trunk leaned toward the contralateral pelvis and a resultant increased force was demonstrated. When ensemble average profiles were considered, an increased lateral bending moment was evident at this point (Fig. 2). However, this increased moment was not statistically significant and remained just within normal limits suggesting the impact on the spine and corresponding trunk musculature was small or even negligible. Consequently, it would be unlikely that a type 1 gait pattern would negatively impact the health of the spine over time in children with CP. During type 2 (Duchenne type gait), the pelvis maintained an almost neutral position while the trunk leaned toward the supporting limb, well outside normal limits. Again, this was consistent with the literature for this type of movement pattern.7 One reported feature of a type 2 pattern is a reduction in loading at the hip.7 As previously mentioned, the hip abductor moment has been suggested to represent the predominant factor in hip joint loading.7,21 Although strength data were not collected in this study, there was evidence of a reduced hip abductor moment, particularly at peak moment values (Fig. 3). Although moment values remained just within normal limits at these points, it would suggest that pathology of the hip abductors may have played a role in this case. Corresponding reactive forces at the lower lumbar spine were slightly raised for children with the type 2 gait pattern, although they remained within normal SD bands. However, an increased lateral bending moment during stance, with 2 distinct peaks occurring at contralateral toe off and heel strike, was evident. This increased moment toward the contralateral side was not altogether surprising. L5-S1 lateral bending moments during CP gait have been previously reported to increase with excessive trunk movement in the coronal plane.12 This would suggest that a type 2 pattern, while it may have the potential to reduce demands at the hip for some participants, could result in greater demands placed on the lumbar spine as the trunk moves into a suitably compensatory position. This type of movement pattern was previously reported as a therapy for unloading the hip joint during gait.13,14 However, our results suggest there may be a negative impact at the lower lumbar spine. Although it is not clear whether this recommended therapy has been used routinely in clinical practice, the promotion of this type of movement pattern as a form of therapy would need to be considered with caution. A type 3 (complex Trendelenburg-Duchenne) gait pattern involving both contralateral pelvic drop and excessive trunk lean was the most common pattern amongst the cohort in this study with just over 40% of participants in this category. To our knowledge, this is the first study to specifically describe trunk and pelvic kinematics associated with this type 3 movement. Reference has been made to a high proportion of people with diplegic CP with both contralateral pelvic drop and ipsilateral trunk lean in a study examining hip deformities in CP.6 However, this was reported by means of observational gait analysis and did not further split the groups according to specific Trendelenburg-Duchenne presentation. We refer to this walking pattern as complex Trendelenburg-Duchenne. It is not strictly accurate to refer to it as a combination of Trendelenburg and Duchenne, as in Trendelenburg type gait (type 1), the contralateral pelvis dips, while in Duchenne type gait (type 2), it remains level or elevated.6,7 With this in mind, another descriptive name should be applied, such as complex Trendelenburg-Duchenne gait (type 3). When the kinematic presentation of this type 3 pattern was considered, a similar pelvic position to the type 1 pattern was demonstrated. However, as a consequence of the pelvic obliquity, trunk side flexion with respect to the pelvis was therefore increased compared with the type 2 pattern where the pelvis remained flat. Peak trunk side flexion in the global frame was also marginally increased compared with the type 2 pattern, suggesting that children with CP and a type 3 gait pattern may have had slightly more involvement of the trunk than those with a type 2 gait pattern. For all 3 movement types the resultant effects at the lower spine were most evident for type 3. Similar to the type 1 pattern, there was an increased force toward the contralateral side at initial swing. However, this force was greater in magnitude compared with type 1. As pelvic position was similar, the increase in force must therefore be attributed to the increased side flexion of the trunk. Peak lateral bending moments were also increased compared with TD and marginally increased compared with those in children with the type 2 pattern. Consequently, it would appear that greater levels of loading were placed on the lower lumbar spine as the trunk and pelvis maintained suitably compensatory positions during type 3 gait compared with TD. Due to only a small increase in L5-S1 moments compared with the type 2 pattern, our results suggest that the contralateral drop of the pelvis (Trendelenburg component) has a relatively small impact on demands at the lower spine. Instead, it is primarily the excessive trunk lean (Duchenne type component) that contributes mainly to increased lower lumbar spinal loading in children who had CP and a type 3 pattern during gait. It is difficult to identify exactly what the impact of pathological reactive forces and moments will be at the spine over time. Although low back pain has been reported in adults with CP,15 this has not been recorded in children with CP. However, radiographic studies have highlighted changes in the lumbar spine of children with CP.22 The Wolff law states that bone tissue will respond by deposition of new bone in response to cyclic high stresses,23 and it has been suggested that pathological changes in the mechanics of motion of the spine can result in the formation of osteophytes along the junction of the vertebral bodies and intervertebral disks.24 Based on this, it is a reasonable assumption that altered trunk movement will have some impact on the health of the spine over time. However, this assumption cannot be concluded from the results of this study. No study to date has specifically examined lumbar stresses and strains specifically related to altered trunk motion during CP gait and the long-term consequences of this altered movement. As a starting point, this investigation provides a foundation and the methods for further exploration. A longitudinal study of low back pain in CP from childhood into adulthood, where reactive forces and moments at the spine are assessed at regular intervals, and an investigation of bone-on-bone forces possibly by means of a forward dynamics approach, may shed further light as to the impact of Trendelenburg and Duchenne type movements over time during CP gait. Based on the results of this study, and taking into account the previously reported moderate to strong relationship between thorax movement and lower lumbar spinal loading,12 it is suggested that therapy is aimed at reducing thorax movement in children with a type 2 (Duchenne) or type 3 (complex Trendelenburg-Duchenne) pattern. However, altered thorax movement is usually a compensation for a lower limb deficiency (eg, weak hip abductors or dysfunction of the affected hip).10 As a result, the optimal thorax lean for the best tradeoff between minimizing the effects of lower limb problems and reducing reactive forces and moments at the spine is unknown. Hip abductor muscle strengthening may be beneficial to reduce excessive mediolateral thorax sway and thus reduce reactive forces and moments at the spine. Alternatively, adjustable walking poles have been shown to result in less lateral trunk lean.25 However, as discussed, more investigation is warranted as to the long-term effects of increased reactive forces and moments at the spine during gait in people with CP. Limitations A number of limitations exist with this study. First, data were collected at a self-selected walking speed. Walking speed has been shown to impact on both kinematics and kinetics.26 However, no attempt was made to control walking speed as the aim was to report patterns characteristic of CP gait. Second, the number of children with the type 2 pattern was small, and results relating to this pattern need to be interpreted with caution. However, as it was the purpose of the study to also examine the effects of a type 2 pattern, descriptive statistics and ensemble average profiles were discussed. Third, a test-retest reliability analysis or measure of minimal detectable change (MDC) was not conducted as part of this study. Results suggest that measures of MDC may be high for some of the reported variables and, while a preliminary assessment of reliability demonstrated good results, a further assessment of reliability is planned. Finally, the inverse dynamic approach only provides the net inter-segmental reactive forces that reflect the effect of external loads. Previous studies have shown that spinal loads can be larger than the net inter-segmental force.27,28 As excessive cocontraction of muscles can occur during CP gait,29 future work incorporating studies of trunk electromyography may be required. Conclusions The majority of children with CP in our cohort had a type 3 (complex Trendelenburg-Duchenne) walking pattern. To our knowledge, this is the first study to specifically describe the associated trunk and pelvic kinematics of this pattern, and we believe that the phrase complex Trendelenburg-Duchenne appears to be suitable for describing this pattern. Additionally, children who had CP and this type 3 pattern demonstrated the greatest increase in kinetics at the lumbar spine compared with TD controls. Trunk position was a critical factor with only a small contribution as a result of the contralateral drop of the pelvis. For children with CP and a type 2 (Duchenne) pattern, kinetics at the lumbar spine were increased. However, although the position of the trunk was also important, numbers in this group were low and only descriptive statistics reported. Finally, the type 1 (Trendelenburg) gait pattern had the least impact at the lower lumbar spine with only small differences in lumbar spinal kinetics compared with TD controls. Consequently, clinicians need to be aware of the resultant effects at the lower spine when presented with these types of walking patterns. Specifically, promotion of a type 2 or type 3 pattern as a noninvasive rehabilitative or conservative treatment should be considered with caution due to the potential for negative effects at the lumbar spine. Author Contributions and Acknowledgments Concept/idea/research design: D. Kiernan, R. O’Sullivan, A. Malone, T. O’Brien, C.K. Simms Writing: D. Kiernan, R. O’Sullivan, A. Malone, C.K. Simms Data collection: D. Kiernan, A. Malone Data analysis: D. Kiernan, R. O’Sullivan, A. Malone, C.K. Simms Project management: D. Kiernan Fund procurement: D. Kiernan Providing participants: D. Kiernan, R. O’Sullivan Consultation (including review of manuscript before submitting): D. Kiernan, R. O’Sullivan, A. Malone, T. O’Brien, C.K. Simms Each of the authors has read and concurs with the content in the final manuscript. The material within has not been and will not be submitted for publication elsewhere except as an abstract. The material within constitutes part of a PhD thesis written by the corresponding author. Ethics Approval Ethics approval was granted by the Central Remedial Clinic's ethics committee. Funding Funding was provided in part by the Scientific and Research Trust of the Central Remedial Clinic. Disclosures The authors completed the ICJME Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest. References 1 Attias M, Bonnefoy-Mazure A, Lempereur M, Lascombes P, De Coulon G Armand S. Trunk movements during gait in cerebral palsy. Clin Biomech (Bristol, Avon) . 2015; 30: 28– 32. Google Scholar CrossRef Search ADS PubMed  2 Heyrman L, Feys H, Molenaers G et al.   Three-dimensional head and trunk movement characteristics during gait in children with spastic diplegia. Gait Posture . 2013; 38: 770– 776. Google Scholar CrossRef Search ADS PubMed  3 Romkes J, Peeters W, Oosterom A, Molenaar S, Bakels I, Brunner R. Evaluating upper body movements during gait in healthy children and children with diplegic cerebral palsy. J Pediatr Orthop B . 2007; 16: 175– 180. Google Scholar CrossRef Search ADS PubMed  4 O’Sullivan R, Walsh M, Jenkinson A, O’Brien T. Factors associated with pelvic retraction during gait in cerebral palsy. Gait Posture . 2007; 25: 425– 431. Google Scholar CrossRef Search ADS PubMed  5 Salazar-Torres JJ, McDowell BC, Kerr C, Cosgrove AP. Pelvic kinematics and their relationship to gait type in hemiplegic cerebral palsy. Gait Posture . 2011; 33: 620– 624. Google Scholar CrossRef Search ADS PubMed  6 Metaxiotis D, Accles W, Siebel A, Doederlein L. Hip deformities in walking patients with cerebral palsy. Gait Posture . 2000; 11: 86– 91. Google Scholar CrossRef Search ADS PubMed  7 Westhoff B, Petermann A, Hirsch MA, Willers R, Krauspe R. Computerized gait analysis in Legg Calve Perthes disease: analysis of the frontal plane. Gait Posture . 2006; 24: 196– 202. Google Scholar CrossRef Search ADS PubMed  8 Trendelenburg F. Ueber den Gang bei angeborener Huftgelenksluxation [On the gait of people with congenital dislocation of the hip]. Dtsch Med Wochenschr . 1895; 2: 21– 24. Google Scholar CrossRef Search ADS   9 Duchenne G. Selections from the clinical works of Dr. Duchenne (de Boulogne). Poore GV, ed. London, England: The New Sydenham Society; 1883. 10 Stief F, Bohm H, Ebert C, Doderlein L, Meurer A. Effect of compensatory trunk movements on knee and hip joint loading during gait in children with different orthopedic pathologies. Gait Posture . 2014; 39: 859– 864. Google Scholar CrossRef Search ADS PubMed  11 Krautwurst BK, Wolf SI, Heitzmann DW, Gantz S, Braatz F, Dreher T. The influence of hip abductor weakness on frontal plane motion of the trunk and pelvis in patients with cerebral palsy. Res Dev Disabil . 2013; 34: 1198– 1203. Google Scholar CrossRef Search ADS PubMed  12 Kiernan D, Malone A, O’Brien T, Simms CK. Children with cerebral palsy experience greater levels of loading at the low back during gait compared to healthy controls. Gait Posture . 2016; 48: 249– 255. Google Scholar CrossRef Search ADS PubMed  13 Schröter J, Güth V, Overbeck M, Rosenbaum D, Winkelmann W. The ‘Entlastungsgang’: a hip unloading gait as a new conservative therapy for hip pain in the adult. Gait Posture . 1999; 9: 151– 157. Google Scholar CrossRef Search ADS PubMed  14 Svehlik M, Kraus T, Steinwender G, Zwick EB, Linhart WE. Pathological gait in children with Legg-Calve-Perthes disease and proposal for gait modification to decrease the hip joint loading. Int Orthop . 2012; 36: 1235– 1241. Google Scholar CrossRef Search ADS PubMed  15 Jahnsen R, Villien L, Aamodt G, Stanghelle JK, Holm I. Musculoskeletal pain in adults with cerebral palsy compared with the general population. J Rehabil Med . 2004; 36: 78– 84. Google Scholar CrossRef Search ADS PubMed  16 Bottos M, Feliciangeli A, Sciuto L, Gericke C, Vianello A. Functional status of adults with cerebral palsy and implications for treatment of children. Dev Med Child Neurol . 2001; 43: 516– 528. Google Scholar CrossRef Search ADS PubMed  17 Schwartz L, Engel JM, Jensen MP. Pain in persons with cerebral palsy. Arch Phys Med Rehabil . 1999; 80: 1243– 1246. Google Scholar CrossRef Search ADS PubMed  18 Kiernan D, Hosking J, O’Brien T. Is adult gait less susceptible than paediatric gait to hip joint centre regression equation error? Gait Posture . 2016; 45: 133– 136. Google Scholar CrossRef Search ADS PubMed  19 Kiernan D, Malone A, O’Brien T, Simms CK. A 3-dimensional rigid cluster thorax model for kinematic measurements during gait. J Biomech . 2014; 47: 1499– 1505. Google Scholar CrossRef Search ADS PubMed  20 Seay J, Selbie WS, Hamill J. In vivo lumbo-sacral forces and moments during constant speed running at different stride lengths. J Sports Sci . 2008; 26: 1519– 1529. Google Scholar CrossRef Search ADS PubMed  21 Brinckmann P, Frobin W, Hierholzer E. Stress on the articular surface of the hip joint in healthy adults and persons with idiopathic osteoarthrosis of the hip joint. J Biomech . 1981; 14: 149– 156. Google Scholar CrossRef Search ADS PubMed  22 Harada T, Ebara S, Anwar MM et al.   The lumbar spine in spastic diplegia: a radiographic study. J Bone Joint Surg Br . 1993; 75: 534– 537. Google Scholar PubMed  23 Wolff J. Das Gesetz der Transformation der Knoche . Berlin, Germany: Hirschwald; 1892. 24 Bogduk N, Twomey LT. Clinical Anatomy of the Lumbar Spine . London, United Kingdom: Churchill Livingstone; 
 1991. 25 Bechard DJ, Birmingham TB, Zecevic AA et al.   The effect of walking poles on the knee adduction moment in patients with varus gonarthrosis. Osteoarthritis Cartilage . 2012; 20: 1500– 1506. Google Scholar CrossRef Search ADS PubMed  26 Schwartz MH, Rozumalski A, Trost JP. The effect of walking speed on the gait of typically developing children. J Biomech . 2008; 41: 1639– 1650. Google Scholar CrossRef Search ADS PubMed  27 Granata KP, Marras WS. The influence of trunk muscle coactivity on dynamic spinal loads. Spine (Phila Pa 1976) . 1995; 20: 913– 919. Google Scholar CrossRef Search ADS PubMed  28 Paul JP. Paper 8: forces transmitted by joints in the human body. Proceedings of the Institution of Mechanical Engineers, Conference Proceedings . 1966; 181: 8– 15. Google Scholar CrossRef Search ADS   29 Gage JR, Stout JL. Gait analysis: kinematics, kinetics, electromyography, oxygen consumption and pedobarography. In: Gage JR, Schwartz MH, Koop SE, Novacheck TF, eds. The Identification and Treatment of Gait Problems in Cerebral Palsy . London, United Kingdom: Mac Keith Press; 2009: 260– 284. © 2017 American Physical Therapy Association

Journal

Physical TherapyOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

Monthly Plan

  • Read unlimited articles
  • Personalized recommendations
  • No expiration
  • Print 20 pages per month
  • 20% off on PDF purchases
  • Organize your research
  • Get updates on your journals and topic searches

$49/month

Start Free Trial

14-day Free Trial

Best Deal — 39% off

Annual Plan

  • All the features of the Professional Plan, but for 39% off!
  • Billed annually
  • No expiration
  • For the normal price of 10 articles elsewhere, you get one full year of unlimited access to articles.

$588

$360/year

billed annually
Start Free Trial

14-day Free Trial