Abstract
Purpose
Lower limb contractures and muscle weakness are common in children with arthrogryposis multiplex congenita (AMC). To enhance or facilitate ambulation, orthoses may be used. The aim of this study was to describe gait pattern among individuals wearing their habitual orthotic devices.
Methods
Fifteen children with AMC, mean age 12.4 (4.3) years, with some lower limb involvement underwent 3-D gait analysis. Three groups were defined based on orthosis use; Group 1 used knee–ankle–foot orthoses with locked knee joints, Group 2 used ankle–foot orthoses or knee–ankle–foot orthoses with open knee joints and Group 3 used no orthoses.
Results
The greatest trunk and pelvis movements in all planes and the greatest hip abduction were observed in Group 1, compared to Groups 2 and 3, as well as to the gait laboratory control group. Maximum hip extension was similar in Groups 1 and 2, but in Group 3, there was less hip extension and large deviations from the control data. Lower cadence and walking speed were observed in Group 1 than in Groups 2 and 3. The step length was similar in all groups and also with respect to the gait laboratory reference values.
Conclusions
Children with AMC were subdivided according to orthoses use. Kinematic data as recorded with 3-D gait analysis showed differences among the groups in trunk, pelvis and knee kinematics, and in cadence and walking speed. The step length was similar in all groups and to the gait laboratory reference values, which may be attributable to good hip extension strength in all participants.
Introduction
Arthrogryposis multiplex congenita (AMC) can be described as a complex condition characterised by deformed joints with an intact sensory system [1, 2]. To be defined as AMC, there should be contractures in multiple joints in multiple body areas present at birth [1]. Arthrogryposis refers to a large heterogeneous group of conditions [3], with an incidence reported from 1/3,000 to 1/5,100 live births [1, 4]. The underlying cause of AMC is decreased foetal movements due to various factors, such as neuropathic and myopathic processes, abnormality of connective tissue, maternal diseases or impaired foetal or intrauterine vascularity. A correlation has been found between the early onset of insufficient movement and the severity of contractures [1].
AMC has been classified into three groups; disorders with mainly limb involvement (four limb, lower limb or upper limb), disorders with limb involvement associated with other organs, and disorders with limb involvement and central nervous dysfunction [1]. Various subtypes have also been distinguished, such as amyoplasia, distal arthrogryposis and Larsen syndrome. The most common subtype is amyoplasia, which represents one-third of all cases and is also usually referred to as classic arthrogryposis. In approximately half of the patients, a specific subtype diagnosis cannot be given [1].
In children with amyoplasia, hip deformities are common and range from soft tissue contractures to hip dislocation [3, 5]. Knee joint involvement has been reported in 70%, with flexion contractures being the most common, followed by extension contractures [5]. Foot deformities are frequently observed, with equinovarus adductus foot being the most common [5, 6]. Children with AMC frequently undergo orthopaedic surgery, often with multiple procedures [3, 7]. The goals for combining operative procedures are to avoid multiple immobilisation periods [8] and reduction of muscle strength [9].
The ability to walk depends on the extent of joint range of motion, in particular in the hips and knees, as well as in the foot, with possibility of plantigrade foot position. Muscle weakness in the lower extremities, primarily in the hip and knee extensor muscles, has also been reported to influence walking ability [10], and in children with amyoplasia, muscle weakness was considered to be more influential on walking ability than the severity of contractures [9]. The involvement of upper extremities with inability of hand and arm support has also been found to influence walking function [10, 11]. Most children with AMC can achieve functional ambulation and 85% have been reported as ambulators by the age of 5 years, but for efficient community ambulation, a wheelchair may be required [3, 12, 13].
To enhance or facilitate ambulation in children with AMC, orthoses may be used to compensate for muscle weakness and to support the lower extremities in an aligned position [8, 10, 13]. Orthoses are made of different materials and with or without knee-locking mechanisms [14]. The use of ankle carbon fibre spring orthoses has reportedly led to increased plantarflexor moment and stride length in children with AMC [15]. Shoe wedges have also been recommended to compensate for plantarflexion contractures [13]. In our clinical practice, effort is made to analyse the needs of orthoses in children with AMC and to evaluate gait outcome. With the exception of one study comparing two different types of orthoses in a group which included four children with AMC [15], no published description of gait in children with AMC has been found. The aim of this study was to describe gait pattern among individuals with AMC wearing their habitual orthotic devices.
Participants and methods
The study was approved by the local ethics committee and informed consent was obtained from the participants and their parents.
Participants
Of 27 children with AMC born between the years 1989 and 2003 and who were treated at the orthopaedic department at Karolinska University Hospital, 17 children fulfilled the criteria of some lower limb motor involvement and independent ambulation, and were invited to participate in a gait analysis study. Inclusion criteria were independent ambulation with or without orthoses and age between 4 and 18 years. Two children were excluded, one child due to an unclear diagnosis and one child with unilateral leg involvement who used a prosthesis, thus, 15 children (8 males, 7 females; mean age 12.4 [4.3] years [range 4.7–17.7]) participated in the study (Table 1). Six children had four-limb and nine had lower-limb involvement (Table 1).
Distribution of patient characteristics, limb involvement, functional ambulation orthoses type and group according to orthosis use
R right; L left; FL four limb; LL lower limb; KAFO knee–ankle–foot orthoses; L locked knee joint; C carbon fibre spring ankle joint; AFO ankle–foot orthoses; O open knee joint with extension stop; H hinged; S solid
Gait analysis performed both with orthosis and barefoot
Orthosis subgroups
Twelve of the 15 participants used orthoses (Table 1). The children were designated into subgroups with respect to orthosis use. Prescriptions of orthoses were based on the presence of muscle weakness, joint contractures or need of joint stabilisation according to the current orthotic programme. Group 1, represented by four participants, used knee–ankle–foot orthoses with locked knee joints (KAFO-L) (Fig. 1). Three participants had grade ≤3 in knee extensor muscle strength and one participant had grade 4 and bilateral knee flexion contractures of 25–30°. Two children had the knee locking mechanisms anteriorly and two posteriorly, depending on the child's hand function. Two participants had a carbon fibre ankle joint in combination with their KAFOs with locked knee joints (KAFO-L-C) (Fig. 2).
Knee–ankle–foot orthoses with locked knee joint (KAFO-L)
Knee–ankle–foot orthoses with locked knee joint and a carbon fibre ankle joint (KAFO-L-C)
Group 2, with eight participants, used knee–ankle–foot orthoses with open knee joints (KAFO-O) or ankle–foot orthoses (AFO) of different types. Two participants had open knee joints with an extension stop and carbon fibre spring ankle joint (KAFO-O-C) (Fig. 3). In one child, this was required bilaterally to control knee hyperextension and knee valgus due to the lack of both anterior and posterior cruciate ligaments. In the other child, the control of knee valgus and knee hyperextension was indicated unilaterally, and an ankle–foot orthoses with a carbon fibre spring ankle joint (AFO-C) (Fig. 4) was used on the contralateral limb. Five participants used AFOs bilaterally and one unilaterally to stabilise the ankle joint. Four participants used an AFO-C with the aim to utilise the properties of the material to restore energy [15], one used a hinged (AFO-H) and one a solid (AFO-S).
Knee–ankle–foot orthoses with open knee joint and a carbon fibre ankle joint (KAFO-O-C)
Ankle–foot orthoses with a carbon fibre ankle joint (AFO-C)
All orthoses were made by the same orthotic company.
Group 3, with three participants, used no orthoses. Among the subjects in Group 3, one child preferred to walk barefoot and two children used shoes, of which one child had a unilateral lift due to leg length discrepancy and the other had heel wedges to compensate for plantarflexion contractures.
Functional ambulation
Functional ambulation was assessed according to a five-level scale, which has been used previously in children with myelomeningocele [16, 17]. In Group 1, all participants were designated level III, i.e. household ambulators and wheelchair users for long indoor distances. In Group 2, one participant was level III and eight participants were level II, i.e. community ambulators who require a wheelchair for long distances outdoors only. In Group 3, one participant was level II and two participants were level I, i.e. community ambulators with no need for a wheelchair (Table 1).
Muscle strength
The strength of the lower limb muscles was tested manually according to a six-graded scale [18], with grade 0 indicating no muscle strength, grade 1 activity traces, grade 2 gravity-eliminated movement, grade 3 movement against gravity and grade 4 indicating movement against gravity with some manual resistance. Grade 5, indicating normal strength, was not given in this study. The distribution of muscle strength in the lower limbs according to Groups 1–3 is shown in Table 2.
Muscle strength, joint range of motion and previous orthopaedic surgery in all participants
Flex flexion; Ext extension; Abd abduction; Dors dorsiflexion; Plant plantarflexion; L left; R right; bi bilaterally
Muscle strength grade <4 was not applied in this study
Selected side for data analysis
0 indicates neutral joint position and − indicates <0
Flexion indicates joint angle from 0 to full knee flexion range (140°), extension indicates <0 (−) and >0 (+)
0 indicates neutral joint position and − indicates <0
Soft tissue surgery
Joint contractures
Passive range of motion was measured with a goniometer. Hip and knee flexion contractures were defined when measured ≥10° and plantarflexion contractures as >0° from a neutral joint position. Two children had hip flexion contractures of 10–20°. Seven children had knee flexion contractures of 10–30° and eight children, five in Group 2 and three in Group 3, had knee hyperextension of 10–20°. Seven children had restricted knee flexion of 20–110° from the neutral position. Eight children had plantarflexion contractures of 10–20° (Table 2).
Orthopaedic surgery
Fourteen participants had undergone orthopaedic surgery, of which 12 children underwent bony surgery and two children had only soft tissue surgery (Table 2).
Gait analysis
All children underwent 3-D gait analysis using an eight-camera motion analysis system (Vicon®, Oxford, UK). The children were equipped with 34 reflective markers aligned with anatomical landmarks on the head, trunk and pelvis, and bilaterally on the arms, thighs, shanks and feet. The markers were placed by the same examiner (ÅB). The lower body was modelled according to the Newington model [19] and the upper body was modelled as the thorax, upper and lower arms, hands and head according to the Plug-in Gait model (Vicon). In the children who wore orthoses, the markers were placed as near as possible to the correct anatomical position. The subjects were asked to walk at a self-selected comfortable pace along a 10-m walkway until complete information from several gait cycles for each side was collected. In five of the six children who were able to walk without orthoses, gait analysis was performed barefoot.
Data analysis
Three kinematic gait cycles were generated for each subject. The following kinematic parameters were obtained from each gait cycle and averaged for each side to describe gait: range of lateral trunk sway, average trunk tilt, range of trunk rotation, pelvic elevation range, average pelvic tilt, pelvic rotation range, maximum hip abduction, maximum hip flexion and extension, hip rotation at initial contact, knee flexion at initial contact and in mid stance, maximum knee flexion, knee flexion/extension range and average foot progression in stance. For Group 3, the maximum dorsiflexion and plantarflexion were also analysed. Time and distance parameters were analysed, wherein velocity, step length and stride length were normalised to the leg length.
Statistical analysis
The Wilcoxon signed ranks test was used to test for differences between the left and right sides within each of the groups and between gait during barefoot walking and with orthoses in five participants in Group 2 (Table 1). The Kruskal–Wallis test was used to compare values between the three groups. All statistical analyses were carried out using commercially available software (SPSS version 16.0). A P-value of ≤0.05 was considered to be statistically significant.
Results
There were no statistical significant differences between the right and left sides in any parameter within any group. For data presentation and subsequent between-group data analysis, the limb with less muscle strength and/or greater contractures was selected. When there was no difference between the limbs, the right side was selected (Table 2).
The results are presented with respect to the orthosis groups. Reference kinematic data is illustrated by the gait laboratory control group, consisting of 23 healthy children aged 5–14 years.
Trunk kinematics
In Group 1, anterior-posterior trunk tilt movements tended towards posterior in stance and swing, which was also slightly pronounced in Group 2. In Groups 2 and 3, the participants were relatively similar in all planes except one participant in Group 3 who showed greater posterior trunk tilt of approximately 11° during the entire gait cycle. Between the groups, the trunk lateral sway range (P = 0.015) and trunk rotation range (P = 0.007) differed significantly (Fig. 5, Table 3).
Illustrations of movements in the trunk and pelvis in all planes for Groups 1–3. The shaded field represents the mean ± 1 standard deviation (SD) of the gait laboratory control group, and each line represents an individual in each group, respectively
Mean (standard deviation [SD]) of the trunk, pelvis and lower limb joint angles, cadence, step length, stride length, walking speed and step width in Groups 1–3
+ = posterior
+ = anterior
− = abduction
− = extension
Max in swing phase
− = external
Initial contact
Mid stance
− = plantarflexion
Stance
Comparison between AMC groups
Pelvis kinematics
In Group 1, internal rotation was more pronounced in stance and external rotation in swing. In all groups, there were various pelvic obliquity movements among the participants, with pelvic lift in swing frequently observed in Group 1. The mean anterior/posterior pelvic tilt values were greatest in Group 3 (Table 3) and movements varied among participants in all groups. Pelvis rotation range differed significantly between the groups (P = 0.021) (Fig. 5, Table 3).
Hip kinematics
Group 1 showed an average of 2–3 times greater maximum hip abduction than Groups 2 and 3 (Table 3), although there were variations among participants in Groups 1 and 2. The maximum hip extension was similar in Groups 1 and 2, reaching almost neutral position, with reasonably similar movements among participants. Group 3 had less hip extension than Groups 1 and 2 (Table 3). The maximum hip flexion was similar in all groups (Table 3), with relatively similar movements among the participants in all groups. The mean hip rotation at initial contact was similar in the groups (Table 3), but large deviations among participants were observed in Group 1 (Fig. 6).
Illustrations of movements of the hip in all planes, and in the knee and foot in the sagittal plane for Groups 1–3. The shaded field represents the mean ± 1 SD of the gait laboratory control group and each line represents an individual in each group, respectively
Knee kinematics
The maximum knee flexion in swing as well as the mean flexion/extension range were lowest in Group 1 (P = 0.048 and 0.015, respectively) (Table 3), as a consequence of the locked orthotic knee joints. Despite the locked orthoses in Group 1, corresponding to the knee flexion contractures, there was greater knee flexion at initial contact and greater knee flexion at mid stance compared to Groups 2 and 3, with variation among Group 1 (Table 3). In Group 2, the knee flexion/extension movements were similar among the participants. Reduced knee flexion during swing was observed in one participant in Group 3 due to restricted knee flexion motion (Fig. 6).
Foot kinematics
Dorsiflexion and plantarflexion for Groups 1 and 2 were not analysed due to various orthoses and footwear conditions. Group 1 showed large variation among the participants, with both large internal as well as external foot progression (Fig. 6, Table 3).
Time and distance parameters
Lower cadence and lower walking speed were observed in Group 1 than in Groups 2 and 3 (P = 0.019 and 0.033, respectively). The step length and stride length were similar in the groups (P = 0.341 and 0.460, respectively). Group 1 showed somewhat greater step width compared to Groups 2 and 3, though the difference was not statistically significant (P = 0.072) (Table 3).
Observations with respect to the gait laboratory kinematic control data and time and distance parameters reference values
Upon visual comparison of the AMC groups’ data with respect to the gait laboratory control data, the greatest deviations in trunk movements were seen in Group 1 in all planes. In the pelvis, great deviations were observed in Group 1 in all movements, in Group 2 in increased pelvic obliquity in swing and in both Groups 2 and 3 in increased anterior pelvic tilt. At the hip, obvious discrepancy was observed in Group 1 as increased abduction and hip rotation, and in Group 3 as less maximum hip extension. At the knee, discrepancy was observed in Group 1 in knee flexion extension movements during the entire gait cycle, and in Group 3 as less flexion in swing in one participant. At the foot, discrepancy was observed in Group 1 in variation among the participants and in Group 2 in increased internal foot progression in swing (Figs. 5 and 6).
Compared to the control group's average (standard deviation) time and distances parameters, cadence 130 (14) steps/min, step length/leg length 0.79 (0.06), stride length/leg length 1.59 (0.12), walking speed/leg length 1.72 (0.26) s−1 and step width 0.14 (0.02) m, it was notable that Groups 2 and 3 had similar cadence, all groups had step length relatively close to the control group, Group 2 had longer stride length and very similar walking speed, and Group 1 had almost twice the step width.
Time and distance parameters during barefoot walking and with orthoses
The results were compared between walking barefoot and with orthoses in five participants in Group 2. With orthoses, the mean cadence was slightly lower, the mean step length increased by approximately 0.09 m, the mean stride length increased by 0.17 m and mean walking speed increased by approximately 0.25 m/s, but there were no significant differences (Table 4).
Mean (SD) of cadence, step length, stride length and walking speed in five patients in Group 2 comparing barefoot and walking with orthoses
Discussion
By subdividing the participants according to the orthoses they had been prescribed, we identified different walking patterns. This subject population is heterogeneous, and orthosis prescription, while not an obvious criterion for group definition, was indicative to a large extent of subjects’ physical properties and body function. The participants in Group 1 who used KAFOs with locked knee joints displayed more extensive trunk and pelvic movements, hip abduction/adduction as well as hip rotation movements compared to the other groups, including those three participants who used no orthoses. A typically wide-based walking pattern in children with AMC wearing KAFOs has been described [20]. In Group 1, wider steps and more abducted hip movements were found compared to the other two groups, which may be interpreted as increasing their support base to improve balance, but also as compensatory mechanisms for the extended knees.
Hip flexion and extension movements in all groups were similar to the gait laboratory control group, but less hip extension in late stance was seen in all groups. In Group 3, who did not require orthoses, two children had passive hip extension range of motion to the neutral position, but did not utilise their full hip extension during stance, possibly due to previous hip deformities, and one child had hip flexion contractures.
The step length was similar in all groups and to the gait laboratory reference value, which may be attributable to good hip extension strength. The importance of hip extension strength grade 4 or better for functional ambulation in patients with AMC has been reported [10] and good correlation has been found between hip muscle strength and motor function in patients with amyoplasia [9]. Even if it is not possible to measure the muscle strength over the entire range of motion, it is important to grade the resistance throughout the available arc of motion, since children with AMC may be strong in the midrange [13].
All of the children in this study were able to perform movements against gravity with or without manual resistance in hip flexion, extension and abduction, and most of them in knee extension and flexion. KAFOs with locked knee joints have been recommended if knee extensor weakness or a knee flexion contracture is present [5, 10]. KAFOs with open knee joints have frequently been used by children with myelomeningocele to stabilise the knee joint in the frontal and transverse planes [21]. In children with AMC, both flexion contractures [10] and hyperextension have frequently been reported, wherein flexion contractures, in contrast to hyperextension, have a negative influence on walking ability [5, 22]. In our group, five knee joints reached a neutral position, ten had knee flexion contractures and 15 knees could hyperextend. In one child, knee hyperextension was excessive, which made an extension stop necessary. The thigh segment also improved the control of the knee in the frontal and transverse planes.
In AMC, complete correction of clubfoot has been described as difficult to achieve [23] and residual stiffness has been reported [6], which can cause ambulatory difficulties [6, 10]. In a recently published long-term follow-up study [24], all independent or community walkers were reported to have plantigrade feet, despite foot deformities at birth. In children with joint contractures, improved postural alignment of the body segments can be obtained by adding wedges under the heels [13]. Such wedges were used in one child in Group 3 (subject 14). Six of eight participants in Group 2 used an energy-restoring carbon fibre ankle joint in their orthoses. In four of six participants, the stride length was longer with respect to the control group data, which is in accordance with previous results [15]. The greater step length and walking speed observed in Group 2 compared to Group 3 may also be attributable to the use of ankle carbon fibre springs.
Six of eight children in Group 2 were able to walk independently without orthoses, of which one child could only do so with difficulty. The results of the five participants who performed gait analysis both barefoot and with orthoses indicate that time and distance parameters improved during walking with orthoses; however, no significant differences could be found in such a small number.
All participants used their orthoses regularly throughout the day, which indicates high acceptance of the orthoses. In Group 1, all children required orthoses to achieve the ability to walk. With orthoses, they were able to walk indoors and even short distances outdoors. The children who used KAFO-L were able to lock and unlock the knee joints independently, and were also able to change their position from sitting to standing. Three children required help putting on the orthoses due to impaired hand function. Limited ambulation due to poor protective responses of the upper extremity has been reported [13] and that children wearing KAFOs often require walkers [20]. In this study group, only one child occasionally used a walking aid, in the school yard for safety reasons.
The ambulatory activity level in youths with AMC has been reported as being lower compared to a control group [25]. In our study, all four children in Group 1 were household ambulators and of the total of seven children who used a wheelchair, six also had a powered wheelchair, four of whom had impaired hand function. Future studies of energy consumption can provide more information about the physical effort during walking in children with AMC.
Conclusion
Gait pattern in children with arthrogryposis multiplex congenita (AMC) was recorded with their orthoses, ranging from locked knee joints to ankle foot orthoses and shoes only. We have shown differences among the groups in the trunk, pelvis and knee kinematics, and in cadence and walking speed. In the children requiring locked knee joints, the greatest trunk and pelvis movements and the lowest knee flexion were observed, as well as the lowest cadence and slowest walking speed.
The step length was similar in all groups and to the gait laboratory reference values, which may be attributable to good hip extension strength in all participants. Comparison between barefoot and orthotic condition was performed in only five participants, and indicated improved stride and temporal parameters, though the small participant number precludes conclusions on orthosis benefit.
Footnotes
Acknowledgments
We would like to thank the children and their parents for participating in the study. This study was supported by the Norrbacka-Eugenia Foundation.