Locomotor-specific measure of spasticity of plantarflexor muscles after stroke

1696

Locomotor-Specific Measure of Spasticity of Plantarflexor Muscles After Stroke Anouk Lamontagne, PhD, PT, Francine Malouin, PhD, PT, Carol L. Richards, PhD, PT ABSTRACT. Lamontagne A, Malouin F, Richards CL. Locomotor-specific measure of spasticity of plantarflexor muscles after stroke. Arch Phys Med Rehabil 2001;82;1696-704. Objectives: To study the stretch reflex excitability (spasticity) of the plantarflexor muscles during gait in patients with hemiparesis and to study the relationships of spasticity during gait with spasticity at rest and gait speed. Design: Cross-sectional, descriptive. Setting: Rehabilitation center. Participants: Convenience sample of 30 patients (58 ⫾ 11yr) with hemiparesis (⬍6mo poststroke) and 15 healthy controls (59 ⫾ 8yr). Interventions: Patients walked at natural speed, healthy subjects at very slow speed for 10 gait cycles. Electromyographic activation of the medial gastrocremius was recorded by using surface electrodes. A 2-dimensional video camera system with reflective markers was used to acquire kinematics of the lower limbs. Main Outcome Measures: Electromyography-lengthening velocity slopes, calculated from measures obtained during the lengthening periods of the medial gastrocnemius muscle during the stance and the swing phases. Measured spatisticity (Modified Ashworth Scale [MAS]), static strength (ankle clonus), and motor control (Fugl-Meyer test). Results: Velocity-sensitive electromyographic responses, indicative of hyperactive stretch reflexes, were found on the paretic side during the stance phase of gait (in 66% of the patients), but not on the nonparetic side or in controls. In many patients, velocity-sensitive responses coexisted with low plantarflexor activation levels during the stance phase. No clear patterns of response were measured during the swing phase in either group. Spasticity during gait in the patients was found to be positively related (r ⫽ .47, p ⬍ .01; r ⫽ .57, p ⬍ .001) to spasticity at rest (MAS; ankle clonus), whereas it was found to be negatively related to gait speed (r ⫽ ⫺.47 to ⫺.53, p ⬍ .01). Conclusions: The validity of the present method is supported by the fact that it is locomotor-specific and that it allowed for a good discrimination between spastic and nonspastic limbs, as well as between stance and swing phases of the gait cycle. The results also support plantarflexor spasticity

From the Jewish Rehabilitation Hospital Research Center, Gait and Posture Research Unit, Laval, Que (Lamontagne), School of Physical and Occupational Therapy, McGill University, Montreal, Que (Lamontagne); Rehabilitation and Social Integration Interdisciplinary Research Center, Rehabilitation Institute of Quebec, Quebec City, Que (Malouin, Richards); and the Rehabilitation Department, Laval University, Quebec City, Que (Malouin, Richards), Canada. Accepted in revised form January 18, 2001. Supported by a doctoral scholarship from the Fonds de la Recherche en Sante´ du Que´bec and Health Canada. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Anouk Lamontagne, PhD, PT, Gait and Posture Unit, Jewish Rehabilitation Hospital, 3205 Place Alton Goldbloom, Laval, Que H7V 1R2, Canada, e-mail: [email protected]. 0003-9993/01/8212-6076$35.00/0 doi:10.1053/apmr.2001.26810

Arch Phys Med Rehabil Vol 82, December 2001

as a factor contributing to the poor locomotor performance after stroke. Key Words: Electromyography; Gait; Hemiparesis; Muscle spasticity; Rehabilitation. © 2001 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation AIT IN SUBJECTS WITH hemiparesis is associated with abnormal muscle activation patterns. These abnormal G muscle activation patterns reflect paresis, spasticity, and excessive muscle coactivation.1-2 Increased muscle-tendon passiverestraint of nonreflex origin, especially in plantarflexor muscles, may also contribute to the locomotor impairment after central nervous system lesion.3-5 A premature activation of the calf muscles during the stance phase, often accompanied by knee hyperextension, has been related to the expression of hyperactive stretch reflexes.1 Thus, premature activation of the calf muscles impedes muscle lengthening and forward rotation of the tibia, resulting in knee hyperextension as the body is propelled forward. Such an activation pattern during gait is present in a substantial proportion of subjects with chronic hemiparesis,1 but it is unclear how early after stroke this pattern develops.2 Although objective criteria were provided for identification of a spastic activation pattern during gait in the studies of Knutsson and Richards1 and Shiavi et al,2 spasticity was not quantified. Furthermore, though the presence of coexisting impairments was recognized, subjects were classified into mutually exclusive categories according to their predominant muscle activation impairment during gait. Such a classification likely underestimates the presence of different types of disturbed motor control (paresis, excessive muscle coactivation, spasticity) during gait after stroke. It is thus of interest to quantify the expression of hyperactive stretch reflexes (spasticity) in the presence of coexisting impairments during gait. Most of the quantitative measures of spasticity relate to the static condition, evaluating the mechanical or the electromyographic response to stretch of the muscle at rest or under a static contraction.6-15 The expression of spasticity in the resting muscle, however, differs from that observed under dynamic conditions.16,17 Other studies have shown that reflex modulation differs according to the nature of the task,18-23 or even within phases of a given task such as gait.18,21,24-26 These observations emphasize the need for a locomotor-specific measure of spasticity during gait. By the end of the 1980s and during the 1990s, a number of studies used H-reflex stimulation21,27-30 or mechanical perturbations27,31,32 to examine the expression of stretch reflexes during gait. These methods have the advantage of providing quantitative estimates of the excitability level of the motoneuron pool as a function of the time course of the locomotor cycle. The use of electric or mechanical stimuli, however, involves additional peripheral inputs not encountered in normal locomotor conditions. Crenna et al4,33-34 reported the presence of abnormally synchronous electromyographic bursts coincident with peaks of

1697

SPASTICITY DURING GAIT AFTER STROKE, Lamontagne

lengthening velocity in spastic muscles of children with cerebral palsy (CP). They proposed a new approach to measure spasticity during gait that consists in verifying, graphically, if a muscle recruitment is velocity-dependent by plotting its electromyographic activity as a function of muscle-lengthening velocity. The use of this method for the measurement of spasticity in soleus and hamstring muscles during gait of children with CP revealed abnormalities either in the gain or the velocity threshold of the muscle electromyographic recruitment.34 The method proposed by Crenna,4,33,34 in contrast to the use of the H-reflex and mechanical perturbations, is much simpler and excludes unnatural stimuli. Our study measures spasticity of the plantarflexor muscles during gait in subjects with hemiparesis by using a method derived from that originally proposed by Crenna.4,33,34 By using a linear regression approach, the slope of the relationship between electromyographic activity and muscle lengtheningvelocity is calculated to provide a quantitative estimate of the expression of hyperactive stretch reflexes (spasticity) during gait. This aims of this study were: (1) to determine the capacity of a new quantitative locomotor-specific measure of spasticity derived from the work of Crenna4,33,34 to discriminate between

subjects having plantarflexor muscle spasticity during gait from those who do not; (2) to quantify the magnitude of spasticity in the plantarflexor muscles during the stance and swing phases of the gait cycle in a group of subjects early after stroke; and (3) to examine the relationship of spasticity measured during gait with clinical measures (spasticity at rest, static strength, motor control) and with gait speed. METHODS Subjects Thirty subjects with hemiparesis (mean age ⫾ 1 standard deviation [SD] 58 ⫾ 11yr) who had a stroke less than 6 months earlier (range, 44 –153d) and 15 healthy subjects (mean, 59 ⫾ 8yr) of comparable height and weight (controls) participated in this study (table 1). The hemiparesis was right-sided in 13 subjects. The subjects with hemiparesis (patients) had a first cerebral thromboembolic lesion confirmed by computed tomography and were able to walk at least 10 meters at a speed of at least 15cm/s with or without manual assistance. Nine of them walked with knee hyperextension in the stance phase. None had a cerebral lesion involving the brainstem or cerebellum. Additional exclusion criteria for both the patients and the

Table 1: Subject Characteristics

Subject

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Mean SD

Gender (M/F)

Age (yr)

Time Since Stroke (d)

Height (cm)

Weight (kg)

Affected Side (R/L)

F F M M M M F M M M M M M F F M M M M M M F M M F M M F M M 8F/22M

67.3 69.8 64.1 50.8 55.5 38.8 40.5 71.5 58.8 50.7 65.2 70.5 48.7 69.4 65.8 45.0 43.0 48.1 56.1 60.3 63.4 72.0 70.7 59.2 69.5 43.0 37.8 69.4 54.1 62.4 58.05 11.05

96 118 80 105 131 44 127 119 100 117 130 107 71 124 67 130 94 111 91 153 89 100 129 140 128 89 123 89 88 106 107 24

155 167 166 158.5 167 180 162.5 170 180 162.5 172.5 172.7 173.5 160 150 174 164 172 173 168 164 164 160 169 148 165 172.5 128.5 165 170 165.1 10.2

72 53 83 62 64 80 71 60 76 70 89 71 71 48 50 99 56 83 76 75 64 66 53 64 70 62 68 57 62 76 68 12

R L L L L R L R L L R R L R R L L L R L R L L L R R L R R L 13R/17L

Modified Ashworth Scale (range, 0–4)

4 3 1 3 1 2 2 1 1 3 1⫹ 1 1⫹ 0 1⫹ 3 1⫹ 3 1⫹ 1⫹ 0 2 3 3 2 2 3 3 0 1 — —

Ankle Clonus (range, 0–4)

4 4 3 4 1 3 4 1 1 4 3 3 2 2 3 1 2 3 2 4 1 3 3 4 3 2 3 4 1 3 2.7 —

Fugl-Meyer Leg (max score, 34)

Gait Speed (cm/s)

20 26 30 26 24 21 15 17 20 20 25 18 19 27 19 21 25 14 14 16 31 27 20 21 23 32 17 10 30 30 21.93 5.69

24.0 33.5 56.2 35.1 54.5 39.0 19.0 44.0 60.0 50.1 20.0 32.0 15.0 90.0 55.0 30.5 66.5 26.0 47.5 13.5 97.5 34.5 25.9 52.5 19.5 94.5 28.5 28.5 81.0 88.5 45.4 24.8

Knee Hyper Extension*

Max EMG† (␮V)

No No No No No No Yes No Yes Yes No No No No No No No No Yes Yes Yes No Yes No No No No Yes No Yes 9Y/21N

29.5 110.8 50.0 54.4 149.4 46.2 53.3 95.5 107.7 77.5 106.5 40.0 25.2 265.2 128.4 29.6 230.9 105.1 150.1 91.8 239.4 76.4 56.2 129.8 59.4 453.5 74.7 84.2 76.7 174.8 112.4 89.4

Abbreviations: M, male; F, female; R, right; L, left. * Defined as a knee extension that goes beyond 0° during the stance phase on the paretic side. † Maximal electromyographic (EMG) activation of the medial gastrocnemius muscle during the stance phase on the paretic side.

Arch Phys Med Rehabil Vol 82, December 2001

1698

SPASTICITY DURING GAIT AFTER STROKE, Lamontagne

healthy subjects (controls) were previous history of neurologic diseases, diabetes, rheumatic, or orthopedic conditions that could interfere with gait. As indicated in table 1, the patients had a wide range of motor impairments and plantarflexor spasticity on the paretic side, as measured by the Fugl-Meyer leg subscore, the Modified Ashworth Scale (MAS), and ankle clonus scores, respectively. All the subjects were undergoing daily physical therapy treatment at the time of the study. This project was approved by the ethics committee of the 2 rehabilitation centers involved. Evaluation of Gait A full gait analysis was made first on 1 side, and then on the other side in the patients. Only 1 side, randomly selected, was evaluated in the controls. The patients were required to walk at their natural speed without an assistive device. Manual assistance was provided if needed. Controls were asked to walk at a very slow speed, which, after analysis, was found to be comparable (p ⬎ .05) to the natural speed of the patients. Movements were recorded by using a 2-dimensional video camera systema (250⫻255 pixels), which was connected on rails parallel to the walkway. A motorized device allowed the camera to follow the subject along the walkway. Reflective markers were placed on the head of the fifth metatarsal, heel, lateral malleolus of the ankle, lateral epicondyle of the femur, greater trochanter of the hip, and lateral edge of the acromion. Three footswitches placed under the heel, midsole, and toe of each foot were used to divide the stance and swing phases of the gait cycle. Electromyographic activity was recorded by using 11-mm silver-silver chloride electrodes.b The skin was a priori rubbed with alcohol to reduce its impedance and pairs of electrodes were longitudinally placed 1cm apart over the muscle belly of the medial gastrocnemius. Sampling frequencies were 60Hz for movements and 1000Hz for the electromyography. All signals were sent to a computer and stored for offline analysis. Electromyographic signals were first preamplified by using miniature preamplifiers near the electrodes (input impedance of 10 9⍀, common mode rejection ratio of 93dB) and sent via optic fibers to a multichannel Neogenixc main receiver (bandwidth, 20 – 800Hz) and stored. The first step in the analysis was to minimize the effect of movement artifacts, if present, by filtering the electromyographic signals with a digital high-pass Butterworth filter at a frequency of 10Hz (second-order dual-pass for 0 lag). The electromyographic signals were then rectified and smoothed by using a 20-Hz low-pass filter. Such a cutoff frequency was chosen to avoid oversmoothing that would have masked, partly or completely, the electromyographic peaks associated with the reflex responses. Also, when dealing with the stretch reflex during ongoing movements such as those during locomotion, the electromechanical delay and the stretch reflex latency merge and act in opposite directions on the timing between electromyography and movements. It was thus chosen to keep the processed electromyographic signal timing close to that of the original signal and to favor the validity of the signal. Clinical Evaluations of Spasticity, Static Strength, and Motor Control Spasticity of the plantarflexor muscles at rest was assessed in the supine position by using the MAS,35 which rates the muscle resistance to manual rapid stretch (0 ⫽ no catch; 4 ⫽ limb rigid in flexion or extension). Ankle clonus was also measured with a rating scale (0 ⫽ no clonus; 4 ⫽ ⬎15 beats). To evaluate the strength of the ankle muscles, subjects were seated on an adjustable chair, arms crossed over the chest, hip flexed at 90°, Arch Phys Med Rehabil Vol 82, December 2001

lower leg supported at 45° of knee flexion, and ankle positioned at 10° of plantarflexion. Stabilization was provided by adjustable straps around the waist, just above the knee, and ankle malleoli. A hand-held dynamometerd was held perpendicular to the foot, at the metatarsal level for the assessment of plantarflexor strength, and over the tarsal bones for the assessment of dorsiflexor strength. In response to a “make contraction” instruction,36 the subjects performed first a submaximal familiarization trial and then a maximal static plantarflexion contraction (n ⫽ 1). After a 1-minute rest, this procedure was repeated to test maximal static dorsiflexion strength (n ⫽ 1). Peak force values were retained for analysis. Motor control was assessed by means of the lower limb subscore (max score ⫽ 34) of the Fugl-Meyer test, which assesses the capacity of the subject to move the limb out of stereotyped synergies.37 All clinical tests were performed following a standardized protocol by the same experienced physical therapist throughout the study. Data Analysis The measurement of spasticity during gait, which is based on the analysis of concomitant electromyographic activity and muscle-lengthening velocity, was inspired by the work of Crenna4,33,34 in children with CP. First, the muscle length and lengthening velocity of the medial gastrocnemius were calculated by using a model developed by Winter and Scott.38 This model takes into account both ankle and knee displacements. Second, analysis of muscle length revealed 2 lengthening periods for the medial gastrocnemius, one in the stance phase (about 5%–50%) of gait cycle and the other in the swing phase (about 70%–95%) of the gait cycle (fig 1). The maximal dorsiflexion during stance on the paretic side varied from ⫺4.08° to 22.8° (mean maximal values for each subject), and during swing it varied from ⫺7.26° to 19.9°. Next, the electromyographic activity of the medial gastrocnemius for each lengthening phase was normalized to 100% of maximum in each phase and the percentage activation was then plotted as a function of muscle-lengthening velocity specific to each phase of the gait cycle. The slope of these plots for individual gait cycles was then calculated and the mean slope was obtained by averaging the slopes obtained for 10 gait cycles. A mean positive slope is indicative of a velocity-dependent increase in muscle activation, suggesting the presence of spasticity,33 while its steepness, if positive, gives an estimation of the reflex gain. Because not all gait cycles are accompanied by a positive slope, the frequency of positive slopes among the 10 gait cycles studied gives an estimate of the preponderance of enhanced stretch reflexes. Statistical Analysis Between-group comparisons of electromyography–musclelengthening velocity slopes for the stance and swing phases were made by using analyses of variance followed by Scheffe´’s post hoc comparisons. Pearson’s correlation coefficients (continuous measures) and Spearman’s correlation coefficients (ordinal measures) were used to study the relationship of spasticity measured during gait (slope) with gait speed as well as with clinical measures of spasticity at rest, strength, and motor control. The alpha probability level was set at .05 and adjusted for the number of planned comparisons. RESULTS Measurement of Spasticity During Gait Figure 1 gives typical results in a control subject and a patient (both limbs) during 1 gait cycle. In the healthy subject

SPASTICITY DURING GAIT AFTER STROKE, Lamontagne

1699

Fig 1. Lengthening velocity, lengthening, and electromyographic activity of medial gastrocnemius muscle (MG) during gait, as well as the electromyography–lengthening velocity plots of the medial gastrocnemius obtained during the stance phase, for (A) a healthy subject, as well as for (B) the paretic and (C) and nonparetic sides of a patient. Two critical lengthening periods were first identified: 1 during stance and 1 during swing. The electromyogram of the medial gastrocnemius during those critical periods was normalized to its maximal value, then plotted as a function of lengthening velocity (only stance phase is shown at the bottom of the figure). The data points shown on the scatter plots were extracted every 2% of the gait cycle. Note that the number of data points used for the regression analysis was not always the same across subjects and between the paretic and nonparetic limb of the same subject because the muscle-lengthening period varied. Both the direction and the magnitude of the electromyography–lengthening velocity slopes were analyzed.

(control), note that the activation of the medial gastrocnemius (third curve) increases with the lengthening of the muscle (second curve). In contrast, the paretic medial gastrocnemius shows an early burst that increases with the lengthening velocity (upper curve). The impact of medial gastrocnemius–lengthening velocity on its activation pattern is depicted in the bottom-most curve; this curve shows the relationship between the level of medial gastrocnemius activity (% of maximum) and concomitant lengthening velocity during the lengthening period of the stance phase. The direction of the slope (positive) indicates a response that increases with lengthening velocity (velocity-sensitive) in the paretic limb that contrasts with the negative slope in the control and nonparetic limbs. In fact, medial gastrocnemius activation in the controls and nonparetic limbs increased with muscle lengthening rather than lengthening velocity, whereas this electromyography-lengthening relationship was disrupted in the paretic limbs. No electromyo-

graphic activity burst occurred in the medial gastrocnemius during lengthening period of the swing phase in both subjects. Figure 2 compares the responses recorded in the 2 groups during the stance (fig 2A) and swing (fig 2B) phases of gait. During the stance phase, the positive direction of the mean slope on the paretic side (mean, ⫹14%/L0 䡠 s⫺1) indicates a velocity-sensitive response that contrasts with negative slope responses on both the nonparetic side (mean, ⫺35.8%/L0 䡠 s⫺1) and in controls (mean, ⫺70.7%/L0 䡠 s⫺1). Two thirds of the patients had a positive slope response on the paretic side. Although positive and negative slopes coexisted in the same muscle group or side, there was a clear pattern of responses in the frequency of slope direction and in slope magnitude. For instance, on the paretic side, patients with a positive slope response displayed on average a positive slope in 60% of the trials (n ⫽ 10); when only 30% or fewer of the trials were positive, it resulted in a negative slope response. In the latter Arch Phys Med Rehabil Vol 82, December 2001

1700

SPASTICITY DURING GAIT AFTER STROKE, Lamontagne Table 2: Relationship Between Spasticity of Gastrocnemius Muscle Measured During Stance Phase of Gait With Clinical Measurements of Impairment: Correlation Coefficients (r ) (n ⴝ 30) Spasticity During Gait (slope)

Spasticity at rest Modified Ashworth Scale Ankle clonus Motor control of lower leg Fugl-Meyer leg score Static strength† Dorsiflexors Plantarflexors

.47* .57† ⫺.28 ⫺.19 ⫺.12

NOTE. Twenty-five subjects were tested for isometric strength; 4 were not tested; and 1 was unable to perform voluntary movements at the ankle. * p ⬍.01; † p ⬍.001, 1-tail probability.

activation of less than 100␮V (⬍50% of controls) of the plantarflexors during the stance phase also had a positive slope response, indicating that spasticity frequently coexists with paresis. Relationship of Spasticity During Gait With Clinical Measures Table 2 gives the correlation coefficients between the locomotor measure of spasticity (slope for the paretic side) and clinical measures of spasticity at rest, motor impairment, and ankle muscle strength. Only the 2 measures of spasticity at rest (MAS, clonus) correlated significantly with the locomotorspecific measure of spasticity.

Fig 2. Mean electromyography–lengthening velocity slope values for the medial gastrocnemius in the paretic (n ⴝ 30) and nonparetic side (n ⴝ 30) of the patients and in the healthy group (n ⴝ 15) during the (A) stance and (B) swing phases of gait. The squares and the boxes show, respectively, the mean slope values and their 95% confidence intervals.

case, the magnitude of the negative slope response was smaller compared with responses from the nonparetic side or from controls. The magnitude of the mean negative slopes was larger in controls (p ⬍ .01) than on the nonparetic side. The frequency of negative slopes on the nonparetic side and in controls was 80% and 90%, respectively, and the responses with a positive slope were small compared with responses from the paretic side. During the swing phase, there was no clear pattern of responses in either group (fig 2B). Only 30% to 35% of the trials had a positive slope response of very small amplitude irrespective of the groups, resulting in mean slopes that had no specific direction. The peak muscle-lengthening velocity during the stance phase on the paretic side (.16 –.89L0 䡠 s⫺1) was similar ( p ⬎ .05) to the range found in controls (.26 –.68L0 䡠 s⫺1). For both the paretic side (.06 –2.50L0 䡠 s⫺1) and the controls (.97– 2.96L0 䡠 s⫺1), the peak muscle-lengthening velocity recorded during the swing phase was higher than that measured for the stance phase. It was also noted that 76% of the 17 patients with a peak Arch Phys Med Rehabil Vol 82, December 2001

Relationship Between Spasticity and Gait Speed Figure 3 and table 3 give the relationship between spasticity (at rest and during gait) and gait speed. All measures of spasticity correlated negatively ( p ⬍ .009 –.003) with gait speed, indicating that the more spastic patients walked more slowly. Closer examination of the locomotor-specific measure of spasticity with gait speed, in figure 3, shows that the patients with a positive slope response generally had a gait speed of less than 50cm/s. The individual slope responses also indicate that

Fig 3. Relationship of gait speed with the gain (electromyography– lengthening velocity slopes) of the stretch reflexes calculated on the paretic side of the patients. The gain of the stretch reflexes decreased as gait speed increased.

SPASTICITY DURING GAIT AFTER STROKE, Lamontagne Table 3: Relationship Between Spasticity and Gait Speed (n ⴝ 30) Gait Speed

Spasticity

At rest MAS Clonus* During gait Slope (gain)

Correlation Coefficient (r)

Determination Coefficient (R 2)

p (2-tailed)

⫺.53 ⫺.48

— —

.003 .007

⫺.47

.22

.009

* Nonparametric Spearman’s correlation tests.

a third of the patients had no locomotor spasticity (negative slope). DISCUSSION Specificity of the Locomotor Measure of Spasticity In this study, a new quantitative and locomotor-specific measure of spasticity was developed. This measure, which relates the activation of the plantarflexors (electromyography) and the velocity of muscle lengthening to give a measure of slope, allowed for the identification of velocity-sensitive responses in the plantarflexor muscles on the paretic side of subjects with hemiparesis. As in the spastic activation pattern (type I) described by Knuttson and Richards,1 a premature activation of the plantarflexor muscles occurring during the lengthening period of the plantarflexor muscles and consisting of 1 or several bursts of low amplitude was observed on the paretic side (fig 1). Our study further shows that the medial gastrocnemius activation bursts during the stance phase were coincident in time with peaks of lengthening velocity. These velocity-coincident bursts accounted for most of the total activation observed in the medial gastrocnemius on the paretic side. The positive slope responses measured in paretic limbs can thus be related to a dominant velocity-sensitive activation pattern, which contrasts with the length-dependent activation pattern in nonparetic limbs. Such a velocity-dependent activation (positive slope response) is compatible with an increased excitability of the stretch reflex.39,40 Moreover, because muscle-lengthening velocities were not higher on the paretic side compared with controls, the velocity-dependent responses measured on the paretic side are consistent with a lower stretch reflex threshold.12,41 These observations altogether support the idea that the paretic side displayed abnormal responses to muscle stretch that are compatible with spasticity.42,43 Our findings that only the paretic side had such abnormal responses further support the discriminant validity of this locomotor measure of spasticity. The lack of abnormal responses on the paretic side during the swing phase of gait, despite higher muscle-lengthening velocities, indicates the expression of locomotor spasticity to be phase-dependent. Such a finding is in accordance with observations made by Crenna34 in children with CP. Previous investigations in patients with multiple sclerosis29 and spinal cord injuries,27,28 by using the H-reflex as a measure of spinal excitability during locomotion, reported similar phase-dependent responses. Yang et al28 hypothesized that the phasedependent modulation of the stretch reflex in spastic patients was preserved either because the intact descending commands were able to generate the reflex modulation or that the modulation originated from the spinal cord level. It was also proposed, in children with CP, that the cerebral insult would affect

1701

the supraspinal regulation over inhibitory segmental mechanisms in a phase-dependent manner, “disrupting descending control on the gating systems acting around ground contact, and interfering much less with the more robust inhibitory effects which operate during the flexion phase of the stride.”34 Our results also suggest that the tendency for reduced dorsiflexion during the swing phase reported on the paretic side in subjects with hemiparesis44-46 is not likely to be related to pathologic stretch reflexes in the plantarflexor muscles. The occurence of a positive slope response does not eliminate the presence of coexisting impairments such as an abnormal coactivation pattern, paresis, or increased muscle-tendon passive stiffness. One might question, however, whether the presence of such coexisting impairments are responsible for the occurrence of positive slope responses. Because an abnormal coactivation pattern is not likely to be length dependent,1 it should not be at cause in the abnormal velocity-sensitive muscle activation that occurs specifically during muscle-lengthening periods of the gait cycle. The influence of paresis, by normalizing the electromyogram to the maximal value recorded during the muscle’s lengthening period, is also taken into account so that electromyographic amplitude does not explain the presence of positive slopes. Paresis, however, may still play a role in the magnitude of the gain of the positive slopes. As mentioned earlier, positive slope responses appear to be the result of a dominant velocity-sensitive behavior over the normal length-dependent recruitment pattern. It is clear that the combination of a low level of activation in late stance (when lengthening velocity is low) and of a premature activation burst in early stance (when lengthening velocity is high) favors a high gain of the electromyography-velocity relationship. Increased plantarflexor passive stiffness may also influence the gain of the slope response, but it is uncertain in which direction it would act. On the one hand, excessive passive stiffness may contribute to limit the velocity of lengthening and possibly the amplitude of lengthening, thus, reducing the magnitude of the “perturbation” and thus the magnitude of the reflex response of the spastic muscle. Higher muscle stiffness and spasticity, however, are thought to potentiate each other, so that increased stiffness of plantarflexor muscles could, indirectly, increase the likelihood of a hyperactive stretch reflex response. Functional Significance of Locomotor Spasticity The number of patients with locomotor spasticity (positive slope response) corresponded to about two thirds of the present sample, indicating that spasticity is a frequent impairment 4 to 6 months after stroke onset. This proportion of patients with spasticity during gait appears higher compared with the proportion of chronic patients that were classified as spastic (type I) in Knutsson and Richards1 (9/26) and Shiavi et al2 (2/12). In the latter studies,1,2 the patients were classified based on visual inspection of muscle activation patterns, in mutually exclusive categories. The presence of coexisting impairments such as paresis and spasticity, however, was also reported by Knuttson and Richards1 (complex activation patterns), as in our study. The slope method thus has the advantage of detecting spasticity (positive slope) regardless of the presence of other impairments, which can explain the high proportion of patients with spasticity reported in our study. This slope method, however, gives no information about the timing and the magnitude of the plantarflexor activation during gait. Thus, contrary to studies that visually inspected muscle activation patterns, it is more difficult to relate the actual slope measure to specific abnormal gait patterns. For instance, plantarflexor spasticity may be associated with knee hyperxtenArch Phys Med Rehabil Vol 82, December 2001

1702

SPASTICITY DURING GAIT AFTER STROKE, Lamontagne

sion.1 In our study, however, neither the amplitude nor the location in the stance phase of the activation burst related to hyperactive stretch reflexes can be determined, thus, precluding the possibility of relating spasticity (positive slope) to the knee movement. Although the occurrence of slopes of opposite polarity for a given muscle group in successive gait cycles was a common feature, 1 polarity always predominated, resulting in a clear mean slope response. Moreover, the magnitude of the slope of nonpredominant polarity was very small. These observations underline an intrinsic variability of the stretch reflex excitability that could be dependent, from 1 cycle to the next, on minute variations from peripheral inputs and descending influences that modulate the neuronal excitability. Ten gait cycles appear to be sufficient for a clear slope response to emerge and to provide a discriminative measure of locomotor spasticity. Last, the negative slope responses in the controls was steeper compared with that in the nonparetic side. Closer examination of figure 1 may provide some explanation for this difference in slope magnitude. Indeed, peak electromyography (bottom curve) in the healthy subject corresponds with the lowest lengthening velocity, whereas on the nonparetic side, the peak electromyography is not so clearly associated with the lowest lengthening velocity. A possible explanation is the gradual decline of activation in the nonparetic medial gastrocnemius (third curve) instead of a sudden drop after the peak as in the healthy subject; this difference in the activation pattern is associated with a biphasic lengthening velocity curve (top curve) that contrasts with the monophasic curve seen in the healthy subject. Locomotor-Spasticity and Clinical Tests The 2 measures of impairment (static muscle strength, FuglMeyer leg motor score) were weakly associated with the locomotor measure of spasticity, indicating these 2 clinical measurements to be poor predictors of spasticity during gait. The scores from the 2 tests that assess the stretch reflex excitability (MAS, ankle clonus) were the only clinical measures to correlate significantly with the locomotor measure of spasticity. Even though the response to a manual passive dorsiflexion recorded in a seated subject at rest may help predict the plantarflexor response to lengthening during the stance phase of gait, the MAS and ankle clonus scores each explain a small proportion (22%–32%) of the variance associated with the positive slope response. The latter findings are not surprising given the numerous studies suggesting the task specificity of the stretch reflex excitability,18-26 and underline the need for a locomotor-specific measure of spasticity. Spasticity and Locomotor Performance In our study, spasticity of plantarflexor muscles measured both at rest and during gait was negatively related to gait speed, which is known to be indicative of locomotor performance after stroke.47,48 The negative relationship of plantarflexor spasticity with gait speed in the patients apparently contrasts with previous reports that spasticity of quadriceps or hamstring at rest is not significantly correlated with gait speed.49-51 It is likely that the occurrence of hyperactive stretch reflexes in the lengthening period of the plantarflexors in early stance phase perturbs the kinematics of the lower limb and compromises the efficiency of ankle push-off in late stance, which is well known to be closely related to gait speed after stroke.46-52 The knee muscles, on the other hand, tend to function as energy absorbers and are less related to propulsion.46 Our findings are consistent with the fact that reducing spasticity with antispastic medication leads to improved movement excursions during Arch Phys Med Rehabil Vol 82, December 2001

gait, especially of ankle dorsiflexion during the stance phase, and improved gait speed.53-55 Other factors, however, such as ankle plantarflexor moment (r ⫽ .68 –.85)46,52 and activation of the plantarflexors (r ⫽ .64, Richards et al56; r ⫽ .73, Lamontagne et al [unpublished observation]) during the stance phase are more strongly related to gait speed, emphasizing the need to consider also the important role of paresis in the locomotor disorder after stroke. Advantages, Limitations, and Future Research The main advantage of the slope measure of spasticity is that it allows for a task-specific evaluation under natural conditions. The slope measure, in contrast to the use of mechanical or electric stimuli, describes electromyographic recruitment in response to stimuli inherent to locomotion. It has to be expected, however, that locomotor-specific stimuli vary across subjects, or even within the same subject evaluated at 2 different times, depending on movement patterns and gait speed. Consequently, within- or between-subject comparisons of slope responses have to take these factors into consideration when evaluating possible changes in reflex excitability. It was clearly shown in our study that hemiparetic subjects, who walked faster (less walking disability), had slope responses closer to normal values. Task-specificity of locomotor stimuli and the presence of a dominant length-dependent activation pattern in normal locomotion also imply that numeric comparisons of the stretch reflex velocity threshold between spastic and nonspastic subjects should not be made. Instead, it can only be inferred that the velocity threshold is lower in more severely affected patients than in healthy controls and in hemiparetic subjects who walk faster. Although the slope measure can quantify the velocity dependence of muscle recruitment during locomotion, it does not provide information about the timing and magnitude of the plantarflexor activation. Thus, to interpret how positive slope responses may be related to abnormal movement patterns, muscle activations must be carefully examined in relation to ankle and knee movements. Our report is a first step toward the investigation of spasticity during a functional task such as locomotion. It would be of interest to study how changes in muscle length or lengthening velocity when walking uphill or downhill or when crossing over obstacles or running affect the expression of locomotor spasticity and performance in subjects with a neurologic insult. CONCLUSION The expression of spasticity during gait early after stroke was quantified in our study by using measures of slope between electromyography and muscle-lengthening velocity. Analysis of the relationship between the level of electromyographic activity and muscle-lengthening velocity self-imposed by the movements during gait has the major advantage of measuring the motor output of the muscle in response to locomotorspecific stimuli. This method is noninvasive and excludes unnatural or external stimuli such as those used for H-reflex stimulation or stretch perturbations during gait.34 Moreover, it requires only the electromyographic records and kinematics (sagittal plane movements) of the lower extremity to obtain the variables required by the customized software, which makes it relatively simple to use. This measure, which is a dynamic measure of spasticity, may prove to be versatile for the evaluation of enhanced stretch reflex during the performance of various tasks. The validity of this method is supported by the fact that it is locomotor specific and that it allowed for good discrimination between spastic and nonspastic limbs, as well as between stance and swing phases

SPASTICITY DURING GAIT AFTER STROKE, Lamontagne

of the gait cycle. The patients had a wide spectrum of spasticity levels during the stance phase with two thirds of them classified as having velocity-dependent electromyographic recruitment of plantarflexor muscles during the stance phase, suggesting spasticity to be more preponderant than previously expected early after stroke. The presence of very low levels of activation with velocity-dependent electromyographic recruitment during the stance phase on the paretic side in many of the patients suggests stretch reflex hyperexcitability frequently coexits with paresis during gait after stroke. Our study also showed spasticity at rest to only weakly predict spasticity during the stance phase of gait, emphasizing the need for a locomotor-specific measure of spasticity. The negative relationship measured between plantarflexor spasticity during the stance phase of gait and gait speed further supports the implication of spasticity in the poor locomotor performance after stroke, though other impairments such as muscle paresis must also be taken into consideration. Acknowledgments: The authors are very grateful to Franc¸ois Comeau, engineer, who developed the analysis software, and Francine Dumas and Daniel Tardif, whose participation in the laboratory was essential and irreplaceable.

17. 18. 19. 20. 21.

22. 23. 24. 25.

1. 2. 3. 4. 5.

6. 7. 8. 9. 10. 11. 12. 13.

14. 15. 16.

References Knutsson E, Richards C. Different types of disturbed motor control in gait of hemiparetic patients. Brain 1979;102:405-30. Shiavi R, Bugle HJ, Sideird T. Electromyographic gait assessment, part 2: preliminary assessment of hemiparetic synergy patterns. J Rehabil Res Dev 1987;24:24-30. Dietz V, Quintern J, Berger W. Electrophysiological studies of gait in spasticity and rigidity. Brain 1981;104:431-49. Crenna P, Inverno M, Frigo C, Palmieri R, Fedrizzi E. Pathophysiological profile of gait in children with cerebral palsy. Med Sport Sci 1992;36:186-98. Lamontagne A, Richards CL, Dumas F, Tardif D, Comeau F. Larger contribution of passive stiffness to ankle plantarflexor moment during gait after stroke. Soc Neurosci 1997;23:769. Abstract. Herman R, Schaumburg H. Alterations in dynamic and static properties of the stretch reflex in patients with spastic hemiplegia. Arch Phys Med Rehabil 1968;25:199-204. Broberg C, Grimby G. Measurement of torque during passive and active ankle movements in patients with muscle hypertonia. A methodological study. Scand J Rehabil Med Suppl 1983;9:108-17. Gottlieb GL, Agarwal GC, Penn R. Sinusoidal Oscillation of the ankle as a means of evaluating the spastic patient. J Neurol Neurosurg Psychiatry 1978;41:32-9. Rack PM, Ross HF, Thilmann AF. The ankle stretch reflexes in normal and spastic subjects. Brain 1984;107:637-54. Lee WA, Boughton A, Rymer WZ. Absence of stretch reflex gain enhancement in volontary activated spastic muscle. Exp Neurol 1987;98:317-35. Powers RK, Marder-Meyer J, Rymer WZ. Quantitative relations between hypertonia and stretch reflex threshold in spastic hemiparesis. Ann Neurol 1988;23:115-24. Powers RK, Campbell DL, Rymer WZ. Stretch reflex dynamics in spastic elbow flexor muscles. Ann Neurol 1989;25:32-42. Lehmann JF, Price R, deLateur BJ, Hinderer S, Traynor C. Spasticity: quantitative measurements as a basis for assessing effectiveness of therapeutic intervention. Arch Phys Med Rehabil 1989;70:6-15. Sinkjaer T, Toft E, Larsen K, Andreassen S, Hansen HJ. Nonreflex and reflex mediated ankle joint stiffness in multiple sclerosis patients with spasticity. Muscle Nerve 1993;16:69-76. Sinkjaer T, Magnussen I. Passive, intrinsic and reflex-mediated stiffness in the ankle extensors of hemiparetic patients. Brain 1994;117:355-63. Knutsson E, Mårtensson A. Dynamic motor capacity in spastic paresis and its relation to prime mover dysfunction, spastic re-

26. 27.

28. 29. 30. 31. 32.

33.

34. 35. 36. 37. 38. 39. 40.

1703

flexes and antagonist coactivation. Scand J Rehabil Med 1980;12: 93-106. Neilson PD, Andrews CJ. Comparison of the tonic stretch reflex in athetotic patients during rest and voluntary activity. J Neurol Neurosurg Psychiatry 1973;36:547-54. Capaday C, Stein RB. Amplitude modulation of the soleus Hreflex in the human during walking and standing. J Neurosci 1986;6:1308-13. Capaday C, Stein RB. Difference in the amplitude of the human soleus H reflex during walking and running. J Physiol (Lond) 1987;392:513-22. Stein RB, Capaday C. The modulation of human reflexes during functional motor tasks. Trends Neurosci 1988;11:328-32. Fung J, Barbeau H. Effects of conditioning cutaneomuscular stimulation on the soleus H-reflex in normal and spastic paretic subjects during walking and standing. J Neurophysiol 1994;72:2090104. Capaday C, Lavoie BA, Comeau F. Differential effects of a flexor nerve input on the human soleus H-reflex during standing versus walking. Can J Physiol Pharmacol 1995;73:436-49. Lavoie BA, Devanne H, Capaday C. Differential control of reciprocal inhibition during walking versus postural and voluntary motor tasks in humans. J Neurophysiol 1997;78:429-38. Crenna P, Frigo C. Excitability of the soleus H-reflex are during walking and stepping in man. Exp Brain Res 1987;66:49-60. Yang JF, Stein RB. Phase-dependent reflex reversal in human leg muscles during walking. J Neurophysiol 1990;63:1109-17. Zher PE, Stein RB. What functions do reflexes serve during human locomotion? Prog Neurobiol 1999;58:185-205. Stein RB, Yang JF, Edamura M, Capaday C. Reflex modulation during normal and pathological locomotion. In: Shimamura M, Grillner S, Edgerton VR, editors. Neurobiological basis of human locomotion. Tokyo: Japan Scientific Pr; 1991. p 335-46. Yang JF, Fung J, Edamura M, Blunt R, Stein RB, Barbeau H. H-reflex modulation during walking in spastic paretic subjects. Can J Neurol Sci 1991;18:443-52. Sinkjaer T, Toft E, Hansen HJ. H-reflex modulation during gait in multiple sclerosis patients with spasticity. Acta Neurol Scand 1995;91:239-46. Rossignol S, Chau C, Brustein E, Belanger M, Barbeau H, Drew T. Locomotor capacities after complete and partial lesions of the spinal cord. Acta Neurobiol Exp (Warsz) 1996;56:449-63. Berger W, Horstmann GA, Dietz V. Spastic paresis: impaired spinal reflexes and intact motor programs. J Neurol Neurosurg Psychiatry 1988;51:568-71. Nielsen JF, Andersen JB, Barbeau H, Sinkjaer T. Input-output properties of the soleus stretch reflex in spastic stroke patients and healthy subjects during walking. Neurorehabilitation 1998;10: 151-66. Crenna P, Inverno M. Objective detection of pathophysiological factors contributing to gait disturbance in supraspinal lesions. In: Fredrizzi E, Avanzini G, Crenna P, editors. Motor development in children. London: John Libbey; 1994. p 103-18. Crenna P. Spasticity and ‘spastic’ gait in children with cerebral palsy. Neurosci Biobehav Rev 1998;22:571-8. Bohannon RW, Smith MB. Interrater reliability of a modified Ashworth scale of muscle spasticity. Phys Ther 1987;67:206-7. Bohannon RW. Strength of lower side related to gait velocity and cadence in stroke patients. Phys Can 1986;38:204-6. Fugl-Meyer AR, Jaasko L, Leyman I, Olsson S, Steglind S. The post-stroke hemiplegic patient: method for evaluation of physical performance. Scand J Rehabil Med 1975;7:13-31. Winter DA, Scott SH. Technique for interpretation of electromyography for concentric and eccentric contractions in gait. J Electromyogr Kinesiol 1991;1:263-9. Herman R, Freedam W, Mayer N. Neurophysiological mechanisms of hemiplegic and paraplegic spasticity: implications for therapy. Arch Phys Med Rehabil 1974;55:33-8. Thilmann AF, Fellows SJ, Garms E. The mechanism of spastic muscle hypertonus: variation in reflex gain over the time course of spasticity. Brain 1991;114:233-44. Arch Phys Med Rehabil Vol 82, December 2001

1704

SPASTICITY DURING GAIT AFTER STROKE, Lamontagne

41. Levin MF, Feldman AG. The role of stretch reflex threshold regulation in normal and impaired motor control. Brain Res 1994; 657:23-30. 42. Lance JW. Symposium synopsis. In: Feldman RG, Young RR, Koella WP, editors. Spasticity: disordered motor control. Chicago: Yearbook Medical Publishers; 1980. p 485-94. 43. Burke D. Spasticity as an adaptation to pyramidal tract injury. Adv Neurol 1988;47:401-23. 44. Berger W, Horstmann G, Dietz V. Tension development and muscle activation in the leg during gait in spastic hemiparesis: independence of muscle hypertonia and exaggerated stretch reflexes. J Neurol Neurosurg Psychiatry 1984;47:1029-33. 45. Lehmann JF, Condon SM, Price R, deLateur BJ. Gait abnormalities in hemiplegia: their correction by ankle-foot orthoses. Arch Phys Med Rehabil 1987;68:763-71. 46. Olney SJ, Griffin MP, McBride ID. Temporal, kinematic, and kinetic variables related to gait speed in subjects with hemiplegia: a regression approach. Phys Ther 1994;74:872-85. 47. Richards CL, Malouin F, Dumas F, Tardif D. Gait velocity as an outcome measure of locomotor recovery after stroke. In: Craik RL, Oatis CA, editors. Gait analysis: theory and application. St-Louis (MO): Mosby; 1995. p 355-64. 48. Goldie PA, Matyas TA, Evans OM. Deficit and change in gait velocity during rehabilitation after stroke. Arch Phys Med Rehabil 1996;77:1074-82. 49. Norton BJ, Bomze HA, Sahrmann SA, Eliasson SG. Correlation between gait speed and spasticity at the knee. Phys Ther 1975; 55:355-9. 50. Nakamura R, Hosokawa T, Tsuji I. Relationship of muscle strength for knee extension to walking capacity in patients with spastic hemiparesis. Tohoku J Exp Med 1985;145:335-40.

Arch Phys Med Rehabil Vol 82, December 2001

51. Bohannon RW, Andrews AW. Correlation of knee extensor muscle torque and spasticity with gait speed in patients with stroke. Arch Phys Med Rehabil 1990;71:330-3. 52. Richards CL, Malouin F, Dumas F, Lamontagne A. Recovery of ankle and hip power during walking after stroke. Can J Rehabil 1998;4:120-9. 53. Knutsson E, Mårtensson A. Action of dantrolene sodium in spasticity with low dependence on fusimotor drive. J Neurol Sci 1976;29:195-212. 54. Knutsson E, Mårtensson A, Gransberg L. Antiparetic and antispastic effects induced by tizanidine in patients with spastic paresis. J Neurol Sci 1982;53:187-203. 55. Knutsson E. Analysis of gait and isokinetic movements for evaluation of antispastic drugs or physical therapies. In: Desmedt JE, editor. Motor control mechanisms in health and disease. New York: Raven Pr; 1983. p 1013-34. 56. Richards CL, Malouin F, Dumas F, Wood-Dauphinee S. The relationship of gait speed to clinical measures of function and muscle activations during recovery post-stroke. In: Draganich L, Wells R, Bechtold J, editors. Proceedings of NACOB II: The Second North American Congress on Biomechanics; 1992 Aug 24-28; Chicago. p 299-302. Suppliers a. Model WV-BD400; Panasonic, 5770 Ambler Dr, Mississauga, Ont L4W 2T3, Canada. b. Medi-Trace Pellet Electrodes™ model ECE 1801; Graphic Controls Canada Ltd, 215 Herbert, Gananoque, Ont K7G 2Y7, Canada. c. Model Neo 210A; Neogenix, 3175 Quatre-Bourgeois, Bureau 100, Ste-Foy, Que G1W 2K7, Canada. d. Model D60107 MK3; Penny & Giles Controls Ltd, 15 Airfield Rd, Christchurch, Dorset BH23 3TJ, UK.