Secondary gait compensations in individuals without neuromuscular involvement following a unilateral imposed equinus constraint

Transcription

1 Gait and Posture 20 (2004) Secondary gait compensations in individuals without neuromuscular involvement following a unilateral imposed equinus constraint Michael J. Goodman a, Jason L. Menown a, Jay M. West, Jr. a, Kory M. Barr b, Darl W. Vander Linden a, Mark L. McMulkin b, a Department of Physical Therapy, Eastern Washington University, Spokane, WA, USA b Motion Analysis Lab, Shriners Hospital for Children-Spokane, 911 W 5th Avenue, Spokane, WA USA Accepted 27 September 2003 Abstract Ankle equinus is the most commonly identified impairment of individuals with spastic hemiplegia (SH). However, it is not clear how equinus at the ankle may contribute to gait deviations at other joints. The purpose of this study was to determine what compensatory gait deviations may occur as a result of an imposed, unilateral equinus constraint. Gait data were collected on 12 adult subjects with and without one ankle constrained in equinus using a unique taping method. Knee extension at initial contact, knee extension in mid stance, and hip extension at terminal stance were all found to be significantly reduced on the ipsilateral side as a result of the ankle constraint. On the unconstrained or contralateral side, subjects tended to adopt a foot-flat or toe-first initial contact pattern. This study suggests that stance phase limitations in both hip and knee extension in the gait of persons with hemiplegia are not necessarily caused by limited length of the involved side hamstrings and/or hip flexors, but rather that they can occur as the result of an ankle plantarflexor contracture alone. Deviations in the contralateral foot contact pattern can also occur secondary to unilateral equinus and should not be assumed to represent bilateral involvement Elsevier B.V. All rights reserved. Keywords: Plantarflexion; Toe-walking; Compensations; Hemiplegia; Gait 1. Introduction The term spastic hemiplegia (SH) describes spasticity occurring primarily on one side of the body with the other side being relatively uninvolved. It is associated with a variety of central nervous system disorders including cerebral palsy, traumatic brain injury, or cerebrovascular accident [1]. Individuals with SH present with a variety of impairments including any combination of spasticity, contractures, weakness, impaired motor control, hyperactive deep tendon reflexes, and/or skeletal deformity [2 4]. Functional limitations may include increased difficulty with activities of daily living, decreased general mobility, and abnormal gait. Individuals with SH present with a variety of pathologic gait patterns. Based on computerized kinematic analysis of the involved side sagittal plane motions of the pelvis, hip, knee, and ankle, Winters et al. [1] detailed four distinct pat- Corresponding author. Tel.: ; fax: address: mmcmulkin@shrinenet.org (M.L. McMulkin). terns. The main deviation of Group I consisted of a foot drop in swing phase along with increased knee flexion at initial contact and hyperflexion of the hip in swing. Group II was characterized by ankle plantarflexion in stance and swing, full knee extension to hyperextension in stance, and hyperflexion of the hip. Subjects in Group III had persistent ankle plantarflexion throughout, increased knee flexion at initial contact, decreased knee flexion in swing, and hyperflexion of the hip in swing. Group IV was described with persistent ankle plantarflexion throughout, limited knee extension at initial contact and midstance, limited knee flexion in swing, and limited hip extension in terminal stance. In planning orthopaedic interventions for individuals with SH, it is critical to determine which of their presenting gait deviations are primary, defined as those occurring as a direct result of a treatable impairment, and which are secondary, defined as those occurring as a compensation for another primary gait deviation. Only primary deviations should be treated, as secondary deviations will resolve upon correction of the primary cause. Excessive ankle equinus is the most commonly identified impairment found in individuals with SH [5]. Excessive gait /$ see front matter 2003 Elsevier B.V. All rights reserved. doi: /j.gaitpost

2 equinus is described as toe-strike at initial contact, with disruption of first and second rockers, and is also accompanied by excessive plantarflexion in swing [6]. Equinus may result from myostatic contracture of the triceps surae, excessive spasticity of the ankle plantarflexors, or a combination of the above two conditions [5]. This impairment disrupts normal sagittal plane ankle motion. However, it is not clear how equinus at the ankle may contribute to compensatory deviations, particularly at the ipsilateral knee and hip. The purpose of this study was to determine what compensatory gait deviations occur as a result of a unilateral ankle equinus constraint. M.J. Goodman et al. / Gait and Posture 20 (2004) Methods Twelve normal adult subjects (eight females, four males; mean age = 23.7 years) without a history of neurologic or orthopaedic deficit, acute lower extremity injury, or recent lower extremity surgery were included in this study. Subjects were screened for normal muscle length of their hip flexors, hamstrings, and ankle plantarflexors. All participants gave informed consent as approved by the local Institutional Review Board. Three-dimensional gait analysis data were captured using a six-camera Vicon 370 motion system 1 and two AMTI 2 force plates. A 15-marker set was used for the lower extremities [7]. Kinematic and kinetic modeling was done using Vicon Clinical Manager software. The ankle to be taped was randomly selected with half of the subjects having the left taped and half the right. Prior to the taping process, all hair distal to the mid-shank was shaved. This same area was lightly sprayed with a pre-taping base to improve tape adherence. Next, anchor strips of standard athletic tape were applied circumferentially around both the mid-calf and the forefoot. With the ankle held in substantial plantarflexion, several strips of heavier fabric tape were then stretched tight from the distal anchor strips up the plantar aspect of the foot, over the posterior aspect of the calcaneus, and along the posterior aspect of the calf to the anchor strips around the mid-calf. This created a pseudo-triceps surae contracture similar to that often found in individuals with SH. Finally, additional athletic tape was applied circumferentially around the entire distal shank and forefoot in order to secure the pseudo-achilles straps (Fig. 1a). Mean passive ankle dorsiflexion for all subjects after taping was 24, a measure which, because the taping technique was done entirely below the knee (Fig. 1a), was unaffected by knee position. Since flexing or extending the knee did not loosen or tighten the straps, knee position had no affect on passive ankle dorsiflexion. 1 Oxford Metrics Limited, 14 Minns Estate, West Way, Oxford OX2 OJB, UK. 2 Advanced Mechanical Technology Inc., 151 California Street, Newton, MA 02158, USA. Fig. 1. Taping techniques to impose an equinus constraint. (a) Taped into a plantar flexed position unilaterally for experimental condition and (b) fabric tape cut just below the level of the spiral tape anchor for control condition. Retroreflective markers were then placed on the subject s pelvis and lower extremities in preparation for gait analysis. Subjects were each requested to walk two times down the 12-meter runway prior to data collection in order to acclimate to the taping. Subjects were instructed to walk in whatever way was comfortable and were given no suggestions as to a strategy to employ. Kinematic and kinetic data from the first three usable trials were analyzed for both the taped and untaped sides in this, the experimental condition (mean number of trials = 8.6). A usable trial was defined as one with clean force plate strikes. With a shortened step length in the taped side/experimental (TE) condition, subjects often had clean force plate strikes for both sides in a single trial. Although kinematic data is available from all trials, the first three with clean force plate strikes for each side, and therefore available kinetic data, were used in the data analysis. The pseudo-achilles straps were then cut, allowing patients to resume normal ankle motion (Fig. 1b). However, the remainder of the tape was left on the foot and calf to prevent

3 240 M.J. Goodman et al. / Gait and Posture 20 (2004) the need for reapplication of markers, and also to control for any effect of the tape itself. Subjects were again given two warm-up walks in order to re-acclimate to their normal ankle motion. Kinematic and kinetic data for the first three usable trials for both sides were again collected for this, the control condition (mean number of trials = 14.4). Due to the longer step length in the taped side/control (TC) condition, subjects generally achieved only a single clean force plate strike by one foot per trial requiring more trials than the experimental condition to generate kinetic data for each side. Data sets were produced and compared for each subject s taped and untaped sides in both the experimental and control conditions, defined as: TE taped side/experimental condition, the side that had an imposed equinus constraint, with the tape intact (Fig. 1a). TC taped side/control condition, the side subjected to the taping process, with the tape cut to allow unconstrained ankle motion (Fig. 1b). UE untaped side/experimental condition, the side with unconstrained ankle motion throughout the study, while the contralateral side in imposed equinus constraint. UC untaped side/control condition, the side with unconstrained ankle motion throughout the study, while the contralateral (taped) side also had unconstrained ankle motion. Kinematic, kinetic, and temporal-distance variables were calculated for the first three usable trials for each subject in each condition and plotted to ensure consistent and quality data. No trials were removed for errors or consistency concerns. Means were then calculated for each subject s three individual trials for use in statistical analysis. Means were not generated from average kinematic and kinetic curves for each subject, rather, the variables were extracted from each subject s three trials before being averaged by subject. For example, peak dorsiflexion in stance was generated for each of a subject s trials and the mean for each subject was then calculated (rather than being the peak dorsiflexion from the average kinematic plot). For statistical analysis, a mixed two-factor ANOVA was used with repeated measures on the second factor. The first factor was body side with two levels, taped and untaped. The repeated factor was condition with two levels, experimental and control. Kinematic and kinetic data, as well as step length, were analyzed using separate ANOVAs with repeated measures. Single factor ANOVAs with the repeated measure of condition were also run on the temporal-distance parameters of velocity, cadence, and stride length. In each instance, statistical significance was established at P = Results Velocity, stride length, cadence, bilateral step lengths, and single support time on the taped side were all significantly diminished in the experimental condition (Table 1). Table 1 Time distance parameter results Parameters Mean (S.D.) Exp a Control a Velocity (cm/s) 111 (13.8) 133 (13.0) Stride length (cm) 123 (10.6) 141 (11.9) Cadence (steps/min) 108 (10.2) 113 (9.0) Step length (cm) Taped side a 65 (5.4) 73 (5.5) Untaped side a 61 (8.1) 71 (6.9) Single support (% gait cycle) Taped side a 35 (2.4) 38 (1.1) Untaped side a 41 (2.3) 39 (2.5) a Exp, experimental refers to the condition when the taped side had an imposed equinus constraint and the untaped side had no constraint. Control refers to the condition when there was unrestricted ankle motion on both the taped and untaped sides. Significant difference between experimental and control conditions P<0.01. Kinematically, there was a significant increase in plantarflexion for all phases of gait for TE compared with TC (Table 2, Fig. 2). Secondary deviations included a significant increase in knee flexion at initial contact and a significant decrease in peak knee extension during mid and terminal stance for the TE condition (Fig. 2). There were no significant differences in peak knee flexion during initial swing nor timing to peak knee flexion in swing between TE and TC conditions (Table 2, Fig. 2). At the hip, there was a significant increase in hip flexion at initial contact and a significant decrease in peak hip extension during terminal stance in the TE compared to the TC condition. A significant increase in maximum anterior pelvic tilt occurred in the TE compared to the TC condition. Also found was a significant decrease in peak knee flexion in initial swing during the UE condition compared to the UC condition, though no other significant differences were found between any other kinematic dependent variables when comparing UE and UC conditions (Table 2). Kinetic data in this study are reported in terms of an internal reference frame. At the ankle, the TE condition resulted in significantly increased peak plantarflexion moment in loading response and mid stance, decreased peak plantarflexion moment in terminal stance, increased ankle power absorption during loading response, and decreased ankle power generation in terminal stance (A2) compared to the TC condition (Table 3, Fig. 3). At the knee, TE showed decreased knee extension moment in loading response, increased knee extension moment in mid to terminal stance, and decreased knee power absorption in pre swing (K3). At the hip, TE demonstrated increased hip extensor moment in mid stance, delayed cross-over from extensor to flexor moment in mid stance, and decreased hip power in pre swing (H3). On the untaped side, there were significant secondary kinetic compensations due to the equinus constraint on the

4 M.J. Goodman et al. / Gait and Posture 20 (2004) Table 2 Sagittal plane kinematics, experimental vs. control conditions on both sides Parameters Taped side mean (S.D.) Untaped side mean (S.D.) Exp a Control a Exp a Control a Ankle Plantarflexion at initial contact 28 (6.0) 1 (4.9) 4 (8.2) 1 (2.9) Peak dorsiflexion in stance 7 (7.8) 15 (6.9) 12 (3.9) 12 (2.5) Peak dorsiflexion in swing 20 (6.5) 3 (3.8) 3 (4.1) 4 (4.3) Knee Flexion at initial contact 26 (7.8) 8 (3.7) 6 (3.5) 6 (3.4) Peak extension in stance 22 (9.5) 4 (5.1) 1 (5.6) 1 (5.7) Peak flexion in swing 62 (3.9) 61 (3.9) 48 (4.6) 55 (2.7) Timing to peak flexion in swing 30 (4.3) 28 (3.2) 28 (3.6) 28 (2.5) Hip Peak extension in stance 6 (7.2) 5 (7.8) 7 (7.5) 8 (8.0) Peak flexion in late swing 49 (7.0) 40 (5.5) 35 (6.9) 37 (5.3) Pelvis Max Anterior Pelvic Tilt 17.2 (5.1) 14.5 (4.5) 17.2 (4.9) 14.7 (4.7) Kinematic values in degrees, except for timing to peak knee flexion expressed in percent of swing phase. Ankle: positive-dorsiflexion, negative-plantarflexion; knee: positive-flexion, negative-extension; hip: positive-flexion, negative-extension. a Exp, experimental refers to the condition when the taped side had an imposed equinus constraint and the untaped side had no constraint. Control refers to the condition when there was unrestricted ankle motion on both the taped and untaped sides. Significant difference between experimental and control conditions P<0.01. contralateral side (Table 3, Fig. 4). The UE ankle plantarflexion moment in loading response and mid stance was increased along with an increase in ankle power absorption in loading response. These changes are explained by the fact that several subjects adopted a foot-flat or toe-first contact pattern on the untaped side. The UE knee showed decreased knee absorption power in pre swing, while the UE hip had decreased hip power generation in pre swing. Fig. 2. Sagittal plane kinematics for the taped side for the experimental and control conditions. Angles in degrees. Ant, anterior; Post, posterior; Flx, flexion; Ext, extension; DF, dorsiflexion; PF, plantarflexion. Curves normalized to percent gait cycle. Data are averaged each 2% of gait cycle for all subjects and trials by condition.

5 242 M.J. Goodman et al. / Gait and Posture 20 (2004) Table 3 Sagittal plane kinetics Taped side mean (S.D.) Untaped side mean (S.D.) Parameters Exp a Control a Exp a Control a Ankle Moment Double-bump pattern (%) First peak magnitude (Nm/kg) 0.8 (0.19) 0.1 (0.14) 0.4 (0.26) 0.1 (0.13) Second peak magnitude (Nm/kg) 0.8 (0.22) 1.4 (0.19) 1.5 (0.15) 1.5 (0.21) Power, LR absorption (watts/kg) 1.1 (0.42) 0.6 (0.21) 1.1 (0.95) 0.7 (0.16) A2 (watts/kg) 1.6 (0.70) 3.8 (0.65) 3.8 (1.07) 4.0 (0.95) Knee Moment First extension peak (Nm/kg) 0.4 (0.20) 0.6 (0.22) 0.5 (0.33) 0.4 (0.22) Second extension peak (Nm/kg) 0.4 (0.17) 0.2 (0.09) 0.2 (0.04) 0.2 (0.06) Power K3 peak (watts/kg) 0.8 (0.17) 0.9 (0.29) 0.6 (0.20) 0.8 (0.27) Hip Power H3 (watts/kg) 1.0 (0.23) 1.3 (0.31) 0.9 (0.27) 1.2 (0.27) Internal ankle plantarflexion moment is designated as positive, internal knee extension moment is positive. Power generation is positive, power absorption is negative. a Exp, experimental refers to the condition when the taped side had an imposed equinus constraint and the untaped side had no constraint. Control refers to the condition when there was unrestricted ankle motion on both the taped and untaped sides. Significant difference between experimental and control conditions P< Discussion The major finding of the present study is that secondary gait deviations occur as a result of the primary deviation of a unilateral, imposed equinus constraint. Subjects increased their hip and knee flexion in stance in response to the ipsilateral imposed equinus constraint. On the contralateral side, subjects tended to adopt a foot-flat or toe-first initial contact pattern. Subjects walked with a slower velocity due to both shorter step lengths and decreased cadence with the unilateral equinus. The ankle, knee, and hip moments occurring in the TE condition have similarities to previously reported kinetic patterns. The ankle moment in the TE condition, although not Fig. 3. Internal sagittal plane kinetics (moments in Nm/kg and powers in watts/kg) for the taped side for the experimental and control conditions. Flx, flexion; Ext, extension; DF, dorsiflexion; PF, plantarflexion; Gen, generation; Abs, absorption. Curves normalized to percent gait cycle. Data are averaged each 2% of gait cycle for all subjects and trials by condition.

6 M.J. Goodman et al. / Gait and Posture 20 (2004) Fig. 4. Internal sagittal plane kinetics (moments in Nm/kg and powers in watts/kg) for the untaped side for the experimental and control conditions. Flx, flexion; Ext, extension; DF, dorsiflexion; PF, plantarflexion; Gen, generation; Abs, absorption. Curves normalized to percent gait cycle. Data are averaged each 2% of gait cycle for all subjects and trials by condition. as dramatic, is similar to a double bump ankle moment pattern with the knee moment being similar to a knee flexor moment pattern, and the hip moment much like the hip extensor moment pattern previously described [8]. In the TE condition, the increased hip extensor moment in mid stance and delayed crossover from extensor to flexor moment in mid stance are particularly interesting because these two indicators are often used for hip treatment decisions. However, these two indicators may not be the result of hip pathology at all given the result of the current study. Initial contact occurring with the ankle joint in excessive equinus has been reported to result in either an internal knee flexion moment or internal knee extension moment pattern [9]. The subjects in this study adopted an internal knee extension moment pattern. Hip and knee flexion in stance might have been adopted in order to compensate for the functionally longer leg created by the equinus constraint. Since peak knee flexion in swing remained unchanged between TE and TC conditions, increased ipsilateral hip flexion during swing, rather than increased knee flexion, appeared to be the compensation for clearance of the plantarflexed ankle. When the equinus constraint was removed, all secondary gait deviations resolved. In a previously reported study, bilateral voluntary toe walking resulted in a similar magnitude of mean ankle plantarflexion during stance (primary deviation) and a similar secondary deviation of increased anterior pelvic tilt of a few degrees [5]. Both bilateral voluntary toe walking and the current unilateral imposed equinus led to decreased velocity, stride length, and single limb support percentage, while cadence decreased only in the current study. Timing to peak knee flexion during swing was unchanged in both studies. In contrast, voluntary bilateral toe walking did not result in nearly the magnitude of increased knee flexion at initial contact, decreased peak knee extension in stance, decreased hip extension in stance, or increased hip flexion in swing as occurred in the current study [5]. Also, voluntary bilateral toe walking produced diminished peak knee flexion in swing [5,10] not seen in our study between TE and TC conditions. The contrasting results might be attributed to either the unilateral versus bilateral equinus or to the external biomechanical constraint used here versus voluntarily maintaining the plantarflexed position. The taping method was believed to more closely mimic a true equinus contracture, as subjects could load the mechanical constraint passively rather than maintain increased plantarflexion via activation of the triceps surae. The unilateral equinus constraint in the current study essentially created a hemiplegic type gait pattern and can be compared to previous gait pattern classifications of persons with SH [1,11,12]. The pattern elicited in the current study did not precisely fit into any of the categories defined by Winters et al. [1]. There was persistent plantarflexion similar to the ankle patterns of Groups II IV. The hip and knee patterns did not precisely match any of the Group II IV knee and hip patterns, though most closely matched the Group IV pattern, particularly in stance. Hullin et al. [12] identified five hemiplegic gait patterns, including Group III characterized by loss of ankle dorsiflexion with a flexed knee and hip, termed triple flexion gait. This pattern is very similar to the pattern elicited in the present study. Hullin and colleagues attributed the cause of a triple flexion gait pattern to be a tight gastrocnemius and tight hip flexors. However, the current study elicited this same triple flexion gait pattern as a result of equinus alone. It has also been proposed that decreased knee extension in stance might be caused by the biarticular nature of a contracted gastrocnemius [12]. The results of the current study seem to counter this assertion for all cases, since our monoarticular model had the same effect on the knee during gait.

7 244 M.J. Goodman et al. / Gait and Posture 20 (2004) The findings of this study suggest it is possible that, in some cases where contractures of the hamstrings and hip flexors are present, they may very well be the effect, rather than the cause, of the gait pattern in question. This is because the described gait pattern does not allow for elongation of the involved side hamstrings and hip flexors to their normal length, thus, potentially leading to sarcomere loss and the development of contractures over time. If this is the case for some individuals, can hamstring and hip flexor contractures be expected to resolve over time following effective lengthening of the ankle plantarflexors alone? Attention to the primary impairment of the contracted plantarflexors would allow a more normal pattern at the ankle which would, in turn, allow for more normal elongation of the involved side hamstrings and hip flexors, functionally stretching these muscle groups and possibly reversing their contractures over time. On the other hand, it could be argued that single level treatment of the plantarflexors in a patient with SH would be ineffective in treating hip and knee contractures, regardless of the cause, because a more normal pattern could not be assumed in the face of the remaining contractures at the hip and knee. 5. Clinical relevance This study demonstrated that stance phase kinematic and kinetic gait deviations at the knee and hip in the gait of persons with hemiplegia are not necessarily caused by limited length of the involved side hamstrings and/or hip flexors, but rather that they can occur as the result of ankle equinus alone. Deviations in contralateral foot contact patterns can also occur due to unilateral equinus and thus should not be assumed to represent bilateral involvement. On the other hand, if a person with hemiplegia presents with diminished or delayed knee flexion in swing, it is unlikely to be the result of a plantarflexion contracture alone. Striving to better understand primary and secondary gait deviations will assist in treatment decision-making and result in appropriate interventions for persons with spastic hemiplegia. References [1] Winters T, Gage J, Hicks R. Gait patterns in spastic hemiplegia in children and young adults. J Bone Joint Surg [Am] 1987;69A: [2] Olney SJ, Wright MJ. Cerebral palsy. In: Campbell SK, Vander Linden DW, Palisano RJ, editors. Physical therapy for children. 2nd ed. Philadelphia: WB Saunders Company; p [3] Gage JR. Gait Analysis in cerebral palsy. London: Mac Keith Press, 1991: [4] Gage JR. Gait analysis: an essential tool in the treatment of cerebral palsy. Clin Orthop 1993;288: [5] Davids J, Foti T, Dabelstein J, Bagley A. Voluntary (normal) versus obligatory (cerebral palsy) toe walking in children: a kinematic, kinetic, and electromyographic analysis. J Pediatr Orthop 1999;19: [6] Lin JP, Brown JK. Peripheral and central mechanisms of hindfoot equinus in childhood hemiplegia. Dev Med Child Neurol 1992;34: [7] Davis RB, DeLuca PA. Gait analysis: current methods and future directions. In: Harris GF, Smith PA, editors. Human motion analysis: current applications and future directions. New York: IEEE Press, 1996: [8] Õunpuu, Sylvia. Joint kinetics: interpretation and clinical decision making for the treatment of gait abnormalities in children with neuromuscular disorders. In: Harris GF, Smith PA, editors. Human motion analysis: current applications and future directions. New York: IEEE Press, 1996: [9] Mikosz RP, Andriachhi TP, Kuo KN. The use of kinetic data for research and clinical applications in human motion analysis. In: Harris GF, Smith PA, editors. Human motion analysis: current applications and future directions. New York: IEEE Press; p [10] Kerrigan DC, Nieto TJ, Riley PO. The effect of toe-walking on knee flexion. Gait Posture 2000;11:138. [11] Rodda J, Graham HK. Classification of gait patterns in spastic hemiplegia and spastic diplegia: a basis for a management algorithm. Eur J Neurol 2001;8(Suppl 5): [12] Hullin MG, Robb JE, Loudon IR. Gait patterns in children with hemiplegic spastic cerebral palsy. J Pediatr Orthop 1996;5: