Contribution of passive stiffness to ankle plantarflexor moment during gait after stroke

351

Contribution of Passive Stiffness to Ankle Plantarflexor During Gait After Stroke Anouk Lamontagne,

PhD, PT, Francine Malouin,

Moment

PhD, PT Carol L. Richards, PhD, PT

ABSTRACT. Lamontagne A. Malouin F. Richards CL. Contribution of passive stiffness to ankle plantarflexor moment during gait after stroke. Arch Phys Med Rehabil 2000;8 I :3Sl-8. Objective: To measure the contribution of passive stiffness to the ankle plantarflexor moment during gait in subjects with hemiparesis early after stroke. The relationship of passive stiffness with gait speed was also examined. Design: Cross-sectional. descriptive. Patients and Other Participants: A sample of convenience of 14 patients (54.7 + 10.9yrs) with a hemiparesis for less than 5 months and I I healthy controls (50.6 2 I I .6yrs). Main Outcome Measures: The contribution of passive stiffness to the plantartlexor moment during gait was obtained using moment-angle slope (stiffness) values. Total plantarflexor stiffness was measured during gait. and passive stiffness was measured during passive dorsiflexion imposed by an isokinetic dynamometer at velocities and ranges of movement matched with values recorded during the plantarflexor lengthening period of the stance phase. The contribution of passive stiffness was obtained by dividing the passive stiffness (dynamometer) by the total plantarflexor stiffness (gait). Results: On the paretic side. passive stiffness contributed more (16.8%: range 2.9% to 49.6%) to total plantarflexor stiffness during gait compared (p < .Ol) with both the nonparetie side (7.3%) and control values (5.9%). This increased contribution on the paretic side resulted from a large muscletendon passive stiffness, a decreased active muscle contribution, or both. Although in some patients the increased passive component led to the development of a total plantarflexor stiffness that was within normal values, it did not in others either because the active component was very small or because limited dorsiflexion during the stance phase prevented the passive component tension to develop. The contribution of passive stiffness was not significantly (p > .05) related to gait speed in both the patients and the controls. Conclusions: The increased contribution of passive stiffness to total plantarflexor moment during gait likely acts as an adaptation for a defective muscle active component, helping ankle push-off at the end of the stance phase. Although this

From the Rehabilitation Research Group. Rehabililalion Institute ofQuebec. and the Rehabilitation Depanmenl, Faculty of Medicine. Lava1 University. Quebec. Canada. Submitted for publication April 19. 1999. Accepted in revised form August 31. 1999. This work was par! of a main projecl funded by National Health Research and Development Program of Canada. and Anouk Lamontagne was the recipient of a scholarship from the Fonds de la recherche en samC du Quebec. A pilo study related IO these results was reported in a poster presentation al the Society for Neuroscience Congress, November 1997. New Orleans. LA. No commercial pany having a direct financial interest in Ihe results of the research supporting this ariicle has or will confer B benefit upon the authors or upon any organization with which the authors are associated. Reprim reques& IO Anouk Lamonlagne. PhD. PT. lnstilut de r6adaptation en d.4licience physique de Quebec. 525 boul Hamel. Bureau B-77. Quebec (Que.). Canada. G I M 2%. 0 2000 by rhe American Congress of Rehabilitation Medicine and !he American Academy of Physical Medicine and Rehabilitation 0003-9993/00/8103-5579%3.00/O

mechanism is effective in most of the patients, it cannot come into action if the dorsiflexion movement during the stancephase is prevented. for instance, by enhanced stretch reflexes. Key Words: Ankle; Gait; Stiffness; Stroke. 0 2000 6~ the American Congress cine and the American Academy Rehabilitation

of Rehabilitation of Physical Medicine

Mediand

HE ANKLE PLANTARFLEXOR muscles have several T important functions during the stance phase of gait. The plantarflexor moment during the stance phase of gait is initially related to a lengthening contraction of the plantarflexor muscles as they control ankle dorsiflexion, and then to a shortening contraction that produces the push-off propulsive power burst. The biomechanical results of the forces are usually expressed as moments (ie, the product of the muscle force and the distance to the joint center, in newton-meters [N ml). The changes in the joint moments as the angular position varies during the gait movements can be expressedas stiffness, which is the change in the moment per degree of excursion (N . m/degree). The total stiffness across a joint is composed of an active component caused by muscle contraction itself and a passive component. Passivestiffness caused by viscoelastic properties of the muscle also contributes to the production of this planttiexor moment.‘.? After a central nervous system (CNS) lesion, impairments in motor control such as paresis, excessive coactivation and spasticity, as well as structural changes in muscle passive properties occur. The differential effects of these impairments on ankle plantartlexor moment during gait, however, are still unclear. Increased passive stiffness during passive dorsiflexions has been well documented after CNS lesion.3-7 Recent findings further indicate that passive stiffness may be increased 43% on the paretic side compared with the nonparetic side as early as 4 months after stroke.8 Few studies, however, have looked at the impact of passive stiffness on the plantarflexor moment during gait after CNS lesions. Dietz and Bergerg and Berger and colleagues,1°in children and adults with hemiparesis or paraparesis, reported increased tension of the triceps surae during the stance phase of gait without a concomitant increase in electromyographic (EMG) activity, suggesting an abnormally high contribution of passive stiffness. In one subject with hemiparesis’ and in a group of children with cerebral pa1sy,2the passive stiffness was reported to contribute three to four times more to the total plantarflexor moment than in healthy controls. In these studies,‘.2 however, neither the ankle range of motion nor the lengthening velocity of the planttiexor muscles during the dynamometric measureswere matched to those recorded during the stance phase of gait. Because passive stiffness of the plantarflexors is velocity-sensitive as a result of its viscoelastic behavior,” matching muscle lengthening velocity to that measured during the stance phase of gait could be important. Moreover, neither of these two studies recorded EMG of plantarflexors and dorsiflexors during gait. As excessive coactivation of antagonist muscle, paresis, or hyperactive stretch reflexes may influence the active contractile component during gait, EMG analyses may be useful to further interpret the Arch

Phys

Med

Rehabil

Vol81,

March

2000

352

PASSIVE

STIFFNESS

Table

Patient

Gender

1 2 3 4

M F M M F

5 6 7 6 9

1 OMI4F

l

Knee

50.6 67.3 63.4

158.5

(Cd

155 164 167 167 172

58.8 50.7 38.8 40.8 69.5

F F M M M

13 14

Height

55.5 69.8 48.1

M M M M

10 11 12

Age (vrs)

hyperextension

180 162.5 180 160

1: Characteristics

Weight (kg)

67 72 64

25.7 24 97.5

L

63 53 83

40 33.5 26

L L L

76 75 80 71 47

29.3 50.1 39.0 19.0 32.5 66.5

L L R L R L

43 45 64.3 54.7

56 99 83 70.6

10.9

7.5

13.7 as an extension

30.5 56.2 40.7 21.1

GAIT

Paretic Side

R R

L L lOU4R

of the knee

during

changes in the relative contribution of the active and passive components to the total plantarflexor moment. Also, both Siegler and associates’and Tardieu and coworkers* reported the contribution of passive stiffness to the total plantarflexor moment at a specific ankle angular position during the stance phase of gait. This method relies on a single point measurement and thus supposes accuracy of both gait and dynamometric angular measurements. In this study, the contribution of passive stiffness to the total plantarllexor moment during gait is measured by means of moment-angle slopes calculated during passive dorsiflexions and dorsiflexion during the stance phase of gait that were matched for velocity and excursion. The slopes describe the relationship between the plantarllexor moment and ankle angle over a range of angular positions. Moreover, the EMG activity of the dorsiflexor and plantarllexor muscles was recorded during both the passive movements and gait. Because of the importance of the interplay of active and passive stiffness to the gait of persons with hemiparesis, the primary aim of this study is to measure the relative contribution of the passive component to the total plantaiflexor moment during gait on the paretic and nonparetic sides of subjects with hemiparesis caused by a recent stroke. Further, to better understand the functional impact of the relative contribution of the passive component, the relationships between the values describing the contribution of the passive component and clinical outcomes such as gait speed, motor control of the lower leg (Fugl-Meyer Test), plantarllexor spasticity (Modified Ashworth Scale and clonus), and maximal passive ankle dorsiflexion are also studied. It was hypothesized that the relative contribution of the passive component to the total plantarllexor moment during gait would be increased on the paretic side. METHODS Subjects Fourteen persons (age range, 38.8 to 69.8yrs) who had a first stroke (patients) 62 to 130 days previously, resulting in a hemiparesis, and 11 healthy subjects (controls) matched for age and anthropometric characteristics participated in this study (tables 1 and 2). The cerebral lesion, confirmed by computerized tomography, was of thromboembolic etiology, either on the Arch

Phys

Med

Rehebil

Vol81,

March

2000

AFTER

of Subjects

Gait Speed (cmlsec)

164 174 174 166 166.6

was defined

DURING

STROKE,

With

Time Since Stroke Idays)

Hemiparesis Fugl-Meyer Leg Score (max = 34)

62 96 89

25 20 31 18 26 14

63 118 111

44

127 67 94

3 1 2 2

-8

3 0 1+ 3 1+ -

-5

beyond

No No Yes No No

-15 -12 9 3

Yes Yes Yes Yes

3 0

Yes No No No

-5

-4 1

No 8NolGYes

-0.8 7.6

-

that goes

Knee Hyperextension During Gait*

6 8 8

1

20 21.9 5.0 phase

Maximal Dorsiflexion (Knee at 0”)

3

15 20 25 21

130 94 93.7 26.4

Modified Ashworth Scale (O-41

3 4 0

20 20 21

100 117

the stance

Lamontagne

0”.

right (11= 10) or left (n = 4) hemisphere. Patients had to be able to walk lOm, without assisting devices but with manual support if needed, at a speed lower than lOOcm/sec.Exclusion criteria for all subjects were previous history of neurologic disease, diabetes, thrombophlebitis, and any orthopedic or rheumatoid conditions interfering with locomotion. 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 two rehabilitation centers involved. Measurement of Total Plantarfiexor Moment During Gait Gait was evaluated first, followed by the evaluation of passive muscle resistance to ankle movements on the same day for the controls, but on separate days (within 3 days) for the patients. The total plantarllexor moment was evaluated during gait on both sides of the patients, whereas one side was randomly chosen for each control. Patients were instructed to walk at their natural speed, and gait was assessedon each side separately by having the patients walk first in one direction and then in the other direction. Controls walked at three speeds: natural, slow, and very slow. Movements were recorded using a 2-dimensional videocamera system (250X255 pixels, Panasonic model WV-BD400a) that was connected to 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, ankle lateral malleolus. lateral epicondyle of the femur, greater trochanter at the hip joint, and lateral edge of the acromion. Three footswitches placed under each foot determined the supported or unsupported phases of each limb and allowed further calculation of spatio-temporal parameters. Ground reaction forces were recorded using a forceplate (AMTI model OR6-5- 1OOO,b)embedded in the walkway, and joint moments were calculated using an inverse dynamic model combined with a link segment Table

Mean SD

2: Characteristics

of Control

Subjects Gait Speed lcmlsec)

Age

Height

Weight

Gender

(yrs)

(cm)

(kg)

Very Slow

Slow

Natural

6M/5F

50.6 11.6

169.6 4.8

69.7 9.6

56.6 14.7

87.6 17.4

120.6 17

PASSIVE

STIFFNESS

DURING

GAIT’AFTER

model.‘? Electromyographic activity was recorded using I l-mm bipolar silver-silver chloride surface electrodes (Medi-Trace Pellet Electrodes. model ECE 1801’). Skin was a priori rubbed with alcohol to minimize its impedance. and pairs of electrodes were longitudinally placed Icm apart over the upper third of tibialis anterior. on the muscle belly of the medial gastrocnemius and. for the soleus. below the lateral gastrocnemius and posterior to the tendons of the peronei. The sampling frequencies were set at 6OHz for movements and forces and at I .OOOHz for EMG.

Measurement of Stiffness During Passive Movements The passive stiffness of the plantarflexor muscles was evaluated during passive movements using a Kin-Corn isokinetic dynamometeti (model 500-4). This dynamometer was modified to give a twofold increase in sensitivity by mechanically doubling the load cell’s lever arrn,‘j and a custom-made software was used for data acquisition. With this modification. the Kin-Corn dynamometer provided a resolution of .15N.m and an accuracy of ‘0.6N.m. Subjects were evaluated while seated on an adjustable chair with the hip flexed at 70” and the knee flexed at an angular position that matched the mean knee position recorded during the lengthening phase of the plantarflexor muscles in the stance phase of gait. This knee angular position varied from 0” to 30” of flexion in the patients, whereas for the controls. a standardized knee angle of IO” of flexion was chosen because of small intersubject variability.” The trunk was stabilized with straps at the chest and waist levels, whereas the thigh was immobilized distally and the foot was kept in the footplate using two adjustable straps. The center of rotation of the ankle joint was aligned with the rotational axis of the dynamometer. The EMG activity of the tibialis anterior, medial gastrocnemius. and soleus was recorded during the passive movements. For each subject, a series of six passive ankle dorsiflexions were imposed at a velocity corresponding to the mean dorsiflexion velocity (range in the patients, 12”/sec to 37”/sec) measured during the stance phase of gait. The passive movement was imposed from 30” of plantarflexion to 5” more than the maximal dorsitlexion angle recorded during the stance phase of gait (range in the patients, 5” to 23”). One-second pauses separated the ankle dorsiflexions. The sampling frequency for force, angle. and EMG signals was set at 1,OOOHz. All signals from the gait and dynamometric tests were sent to a computer and stored for off-line analysis. Electromyographic signals recorded during both tests were first preamplified using miniature preamplifiers near the electrodes (input impedance of IO” a, common mode rejection ratio of 93dB) and sent via optic fibers to a multichannel Neogenix main receivef (bandwidth. 20 to 800Hz) and stored. The first step in the analysis was to minimize the effect of movement artifacts, if present, by filtering the EMG signals with a digital high-pass Butterworth filter at a frequency of IOHz (second-order two pass for zero lag). The EMG signals were then rectified and smoothed using a 20-Hz low-pass filter.

STROKE,

353

Lamontagne

during the passive dorsiflexion. was given a score from 0 to 4 (0. no clonus; 4. more than IS beats). These tests were administered after standardized procedures by an experienced physiotherapist.

Data Analysis The relative contribution of the passive component to the total plantarflexor moment during gait was measured by means of moment-angle slopes ratios. Because the units for the slopes are in N.m/degree. the terms passive stiffness and total plantarflexor stiffness were used to designate. respectively, the passive component. as calculated from isokinetic testing, and the sum of the active and passive component contributing to the plantarflexor moment during gait. For calculation of the plantarfexor total stiffness during gait, two gait cycles for the patients and three gait cycles for the controls with concomitant data on movements. force, and EMG for each side were retained for analysis. The ankle momentangle slopes during the stance phase were calculated for each gait cycle and subsequently averaged for each subject. The calculation of these slopes consisted of first, plotting ankle moment as a function of ankle angular position during the gait cycle (fig I A). Second, the part of the moment-angle curve during the stance phase of gait where the ankle was dorsiflexing was retained for analysis, and that where the ankle is plantarflexing was not included in the analysis. Third. the moment-angle slope was calculated using all data points between 0” and the maximum dorsiflexion position recorded during the stance phase (fig I B). For the measurement of passive stiffness during passive movements, correction was made for gravitational forces acting on the foot and the Kin-Corn footplate.lh The first movement of each series of six movements was excluded, because it always gives higher moment values than the subsequent ones.” Moment-angle slopes were calculated for the next five movements and subsequently averaged for each subject. These moment-angle slopes were calculated in the same dorsiflexion range as that used for the calculation of the plantarflexor total stiffness during gait. Care was taken to ensure that passive stiffness was measured in the range of movement unaffected by the acceleration or deceleration phases of the isokinetic movements.“.lj To eliminate the effect of muscle activity on records of passive stiffness, trials with EMG activity 2SD larger than baseline values were systematically rejected.“.l”.‘6 The contribution of passive stiffness was obtained by dividing the slope calculated for the passive stiffness measured during the passive movements by the slope for the total plantarflexor stiffness calculated during gait (fig IB): [Passive stiffness (N.m/degree)/ Total stiffness (Nm/degree)]

X 100%

In two patients (No. I and l I ), who were unable to reach 0” of dorsiflexion during gait, the moment-angle slope for the paretic side was calculated over the available range from -20” to -5”.

Clinical Measurements

SLatistical Analysis

In addition to the gait and passive movement testing, the patients underwent clinical evaluations of the impairments of the paretic lower limb. Maximal ankle dorsiflexion on the paretic side was measured with a goniometer while the patient was supine and the knee was at 0”. Motor control was assessed by the lower limb section of the Fugl-Meyer test.‘? Spastic hypertonia of the plantarflexors was estimated with the modified Ashworth Scale,‘5 which grades muscle restraint to a manual passive ankle dorsiflexion, while ankle clonus, elicited

Because of heterogeneous variances, differences in the contribution of passive stiffness between the controls and both sides of the patients were compared with the Kruskal-Wallis one-way analysis of variance (ANOVA) followed by MannWhitney or Wilcoxon signed-rank test post hoc comparisons. The effect of gait speed on the contribution of passive stiffness in the controls was examined using a one-way ANOVA followed by Scheffe post hoc comparisons and Spearman correlation coefficients. Pearson correlation coefficients were Arch

Phys

Med

Rehabil

Vol 81, March

2(100

354

PASSIVE

A

STIFFNESS

Total plantarflexor stiffness during gait position

Ankle 4

DURING

B

t5

STROKE,

Lamontagne

(deg)

Total

planterflexor

stiffness

(gait)

20 10

SE 22

0

Em= zg =o

-10 I

AFTER

Contribution of passive stiffness in a subject with hemiparesis

_ 5 ‘E a, c E

GAIT

PE Ankle

1.5,

moment

(N.mlKg)

72 - Pantic llnlb - - Non-pnmtlo limb

1.92 N.m,deg 2.64 N.nJdag

,, ,

36-

I

I 30

Passive

0

20

40 60 Gait cycle (%)

80

stiffness

(passive

movements)

100

-30

-20

-10

0

10

20 .

Dorsiflexion Paretic side: Non-paretic Dorsiflexion

side:

(deg) 33.3 % 11.7%

(deg)

30

Fig 1. (A) The angular-dependent moment measured during gait. (B) An example of the contribution of passive stiffness on the paretic and nonparetie sides of a patient. The lower graph in (A) is obtained by plotting the ankle moment during gait as a function of the ankle position; the right-directed arrow in the lower graph indicates that the ankle is dorsiflexing, while the leftdirected arrow indicates that the ankle is plantarflexing. The upper graph in (B) represents the total stiffness (momentangle slope) measured during gait, and the lower graph represents the passive stiffness measured with the isokinetic dynamometer. Only the dorsiflexing movement, as indicated by the right-directed arrow, is needed for the contribution of passive stiffness calculation. The contrfbution of passive stiffness is obtained by dividing the passive stiffness by the total stiffness. Note that in this subject, the large passive stiffness of the paretic limb led to a higher contribution of passive stiffness on the paretic side than on the nonparetic side.

used to estimate the linearity of the moment-angle slope used to measure passive stiffness during passive movements and gait. Relationships between the contribution of passive stiffness in the patients with clinical outcomes such as the Fugl-Meyer leg score, the Modified Ashworth Scale, and ankle clonus were examined using Spearman correlation coefficients. The a probability level was set at .05 and was adjusted for the number of planned comparisons when applicable. RESULTS Contribution of Plantartlexor Passive Stiffness Figure 2 shows the mean (?SD) contribution of passive stiffness for the patients and the controls. The passive stiffness contribution on the paretic side was 16.8%, which is two to three times higher (p < .Ol) than that measured on the nonparetie side (7.3%) and in controls (5.9%) walking at very slow speed. The nonparetic side had a passive stiffness contribution that was similar (p > .05) to control values, although a tendency for slightly higher values was observed. Table 3 gives the correlation calculated between the ankle plantarllexor moment and the corresponding angle over the range of movement. The moderately high correlations calculated for the paretic and nonparetic sides of the patients (r = .70 to .74) support the choice of a linear model to quantify moment-angle relationship for both conditions. The correlation levels, however, tended to be lower for the patients than for the controls. Figure 3 gives the slope ratio for the passive stiffness contribution of each patient in comparison with mean values and the 95% confidence interval of the values obtained in the controls walking at very slow speed. Ten patients had a ratio higher than control values on the paretic side, and the passive stiffness contribution varied from 8% to 49.6%. The large variability in the passive stiffness contribution scores for the paretic side may in part by explained by the interplay between the passive stiffness (numerator) and the total stiffness (denomiArch

Phys

bled

Rehabil

Vol81,

March

2000

Paretic side

Non-paretic side

Controls

Passive ouffnsoo=o.43 Total sIUlners=3.6

Fig 2. The mean contribution of passive stiffness to total plantarflexor stiffness during gait, calculated for the paretic and nonparetic sides of the patients (n = 14) and for the controls walking at very slow speed. The mean passive stiffness and total stiffness values in each group are indicated at the bottom of the figure. Note that the latter values are means and that simply dividing one by the other will not necessarily reflect the contributions of passive stiffness illustrated by the bar graph. The contribution of passive stiffness on the paretic side was significantly larger compared with that of the nonparetic side and controls (p < .Ol). The mean contribution of passive stiffness on the nonparetic side was similar to that of the controls. l +p C .Ol.

PASSIVE

Table 3: Pearson Determination (w)

Correlation Calculated

STIFFNESS

DURING

Coefficients (r) and Coefficients of Between Ankle Moment and Angular Displacement Passive Stiffness (Passive Movements) Rz

r

RI

Paretic side Nonparetic side Control group

.74 .70

.73 .66

.71 .70

.67 .65

Natural speed Slow speed Very slow speed

.99 .99 .99

.98 .98 .98

.95 .97 .96

.91 .94 .92

with

AFTER

STROKE,

hemiparesis

in the calculation of the ratio. This interplay was examined by means of individual analyses of total and passive stiffness. This led to a classification of 12 subjects into four different groups; 2 other patients were unable to dorsiflex the ankle during the stance phase of gait and were excluded from the groups (patients I and IO). Figure 4 illustrates four different patterns of active and passive contributions to the total plantarflexor stiffness during gait on the paretic side of the patients. Patients in group 1 (II = 4) had a large passive stiffness component combined with a low active component to yield a low total plantarflexor stiffness compared with control values during gait, which led to a high relative contribution (>25%) of the passive component. Patients in group 2 (II = 3) had higher passive stiffness values than controls and a slightly decreased active component so that total plantarflexor stiffness values were within normal limits. In group 3. patients (II = 3) had passive stiffness values similar to control values but an active component that was smaller than control values to produce lower total plantarflexor stiffness values during gait. In both groups 2 and 3, the passive stiffness contribution ratio was moderately increased (8% to 15%) compared with controls. Subjects in group 4 (n = 2) had normal passive stiffness contribution ratios, with passive stiffness

nator)

Relationship Between the Passive Stiffness Contribution and Gait Speed There was no significant difference (p > .OS) in the passive stiffness contribution among the gait speeds (natural, slow, very

(%) 60

m o F-

Fig 3. Individual contribution of passive stiffness to total plantarflexor stiffness on the paretic and nonparetic sides of the patients (n = 14), with the 95% confidence interval for the controls walking at very slow speed in background. The subjects are classified in group 1, 2, 3, or 4. according to the magnitude of the passive stiffness and total stiffness used in the calculation of the contribution of passive stiffness ratios. The characteristics of these groups are described in the text. Ten subjects had higher passive stiffness contribution ratios on the paretic side compared with healthy values. Although the contribution of passive stiffness ratios of the nonparetic side was usually within the control values, six subjects had higher contribution of passive stiffness ratios.

355

Lamontagne

values similar or higher (patient 9) and an active component that was also similar or even higher (patient 3, who walked faster than the controls) than control values. The two patients (No. I and IO) who were excluded from the groups had passive stiffness contribution ratios within normal limits. To further investigate the potential effect of a larger dorsiflexion range in these patients, an adjusted measure of passive stiffness was calculated to estimate the stiffness if the paretic ankle was able to attain the same dorsiflexion angle as the ankle on the nonparetic side. This was performed by measuring the passive stiffness in the same dorsiflexion range, or as close as possible, to that measured for the nonparetic ankle during the stance phase of gait. After adjustment for the dorsiflexion range of motion, the passive stiffness contribution of the paretic side was 14.9% and 16% for patients I and 10, respectively, which is more than double the contribution in controls. Six patients, four of them in group I. also had a higher passive stiffness contribution ratio on the nonparetic side (fig 3). These higher ratios were associated with either a decreased total plantarflexor stiffness and a lower active component compared with control values (n = 3) similar to patients in group 3 on the paretic side, or both an increased passive stiffness and a decreased total plantarflexor stiffness (n = 3) similar to patients in group I. The muscle active component was also found to contribute less to the total plantarflexor moment during gait in the patients than in healthy controls. At least two muscle activation patterns were associated with the decreased active component. A paretic pattern characterized by a decreased soleus and medial gastrocnemius activation during the stance phase of gait was observed in the majority (n = 9), whereas an abnormal agonistantagonist coactivation of the ankle muscles during the stance phase of gait was noted in three of the patients.

Total Stiffness (Gait)

r

Subjects

GAIT

Paretic side Non-paretic side CTLs

(95% confidence

interval)

40

20

0 2

4

5

6

8 13 14 7 11 12 Subjects (no.)

3

9

1 10

HHHH Gr. 1

Gr. 2

Gr. 3 Arch

Phys

Gr. 4 Bled

Rehabil

Vol81,

March

2ooo

356

PASSIVE

CTLS

2

4

5

6

STIFFNESS

8

DURING

13

Subjects

GAIT

14

slow) in the controls (5.5%, 5.3%, and 5.9%, respectively), indicating that the relative contribution of passive stiffness to the total plantarllexor moment is not dependent on gait velocity. Further support for this finding is provided by the nonsignificant correlation coefficients calculated between the contribution of passive stiffness and gait and gait speed (at each of the three gait speeds). Similarly, in the patients, the contribution of passive stiffness to the total plantarflexor stiffness on both the paretic and nonparetic sides was not significantly related to gait speed. Relationship Between the Passive Stiffness Contribution and Clinical Outcomes The relationship of the passive stiffness contribution to the total stiffness on the paretic side with maximum dorsiflexion attained passively at the ankle on the paretic side (knee at 0”) is illustrated in figure 5. The passive stiffness contribution significantly increased with the decreasing dorsiflexion range of motion. Note that the passive stiffness itself, not illustrated in figure 5, also significantly increased (r = -.46, p < .05) with the decreasing dorsiflexion range of motion. Passive stiffness contribution was not significantly related (p 5: .05), however, to measures associated with hyperactive stretch reflexes, such as ankle clonus and Ashworth Scale scores. In addition, no significant relationship was found between the contribution of passive stiffness on the paretic side and motor impairment, as measured by the Fugl-Meyer leg test. DISCUSSION The main finding in this study is that as early as2 to 5 months after a stroke, the relative contribution of the passive component to the total plantarflexor moment during gait on the paretic side is two times larger than that of the nonparetic side and three times larger than values obtained in normal controls walking very slowly (5.9% 2 1.8%). Such a finding is consistent with the increased resistive torque of plantarllexor muscles measured by myometry as early as 2 to 4 months after stroke.* The contribution of passive stiffness values measured in this study on the paretic side is in the same range of values reported previously in one subject with hemiparesis’ (21%) and in children with cerebral palsy* (1.4% to 43%). It is reasonable to assume that the functional role of the increased contribution of the passive component to the total plantartlexor moment during Phys

Med

Rehabil

Vol81,

March

2000

7

STROKE,

11

Lamontagne

12

3

9

(no.)

Gr. 2

Arch

AFTER

II

Gr. 3

‘Gr.4

Fig 4. The active and passive components contributing to total plantarflexor stiffness during gait on the paretic side of each of the patients (n = 12, subjects 1 and 10 were not included in the groups) and in the controls (mean f 2SE). Patients in group 1 had high contribution of passive stiffness ratio because of an increased passive component and a decreased active component. Patients in group 2 had a moderately increased contribution of passive stiffness because of an increased passive component but an active component within normal limits. Patients in group 3 also had a moderately increased contribution of passive stiffness because of a reduced active component. In group 4. the patients had normal contribution of passive stiffness ratios.

the stance phase of gait helps minimize the effect of the lack of generation of an adequate active plantarflexor moment. An increased passive component could thus contribute to the forward propulsion of the body by an energy storage and release mechanism in late stance at push off.18 Adaptive Use of Excessive Passive Stiffness Present results indicate that although the increased contribution of passive stiffness to the plantarflexor stiffness may serve as an adaptation to improve locomotor function, its impact is variable. In fact. when both the passive and total plantarflexor stiffness values are abnormal (increased and decreased, respectively, relative to control values), it resulted in the highest level (25% to 50%) of passive stiffness contribution (group 1). When only one factor was abnormal (either the passive stiffness, group 2; or the total plantarflexor stiffness, group 3), a moderate level of passive stiffness contribution (8% to 15%) resulted. The

.

-20

-15

-10

Dorsiflexion

-5

range

0

of motion

5

10

(deg)

Fig 5. The contribution of passive stiffness to total plantarflexor stiffness of the paretic side as a function of the maximal passive dorsiflexion in the supine position with the knee at 0’ (n = 14). The 95% confidence interval of the contribution of passive stiffness measured in the controls is given for comparison. Positive angular values of range of motion indicate the ankle was dors-fflexed, whereas negative values indicate that the ankle was plantarflexed. The contribution of passive stiffness on the paretic side was negatively related to maximal passive dorsiflexion. r= -.60; p < .Ol.

PASSIVE

STIFFNESS

DURING

GAIT

common factor in patients demonstrating an increased (high. group I : or moderate. groups 2 and 3) contribution of passive stiffness to total plantarHexor stiffness was either an increased passive stiffness or a decreased total plantarHexor stiffness. Present data indicate that only one factor needs to be abnormal to induce a significant increase of the passive stiffness contribution. The increased passive stiffness seen in groups I and 2 (tig 4) likely compensates for the lack of adequate force produced by the active component. Further examination of figure 4. however. shows that patients in group 2. but not group I. had total stiffness values within normal limits, suggesting this adaptation to be insufficient in patients with a weaker active component. Thus. when both factors are abnormal. the increased contribution of passive stiffness remains of limited value as a functional adaptation. In the latter case (patients in group I ). the contribution of the passive component may be so high that it acts like a crutch. restraining forward rotation of the tibia over the foot. Likewise. observations in the two patients not included in the groups indicated that excessive passive stiffness and enhanced stretch reHexes preventing dorsiHexion during gait may interfere with this adaptative behavior. Present results also support the idea that increased passive stiffness does not always develop concomitantly with a decreased total plantarHexor stiffness. as for patients in group 3. One may question why passive stiffness increases markedly in some, but not in the others, or why some patients adapt better than others for a weak active component by developing an increased contribution of the passive stiffness. The development of passive stiffness after stroke has been attributed mainly to degenerative changes that affect the passive properties of the muscle-tendon unit.“.J,‘q Although the stimuli that lead to increased passive stiffness are still strongly debated, disuse and spasticity have both been incriminated. The fact that increased passive stiffness develops not only on the paretic but also on the nonparetic plantarHexorsh.x suggests that stimuli (or lack of) leading to the development of passive stiffness affect ultimately both sides. The finding that a moderate increase (8% to 15%) in the contribution of passive stiffness on the nonparetic plantarHexors coexisted only in the patients with the highest level of contribution of passive stiffness on the paretic side (group 1). further supports this view. For instance, changes on the nonparetic side may in part be a consequence of the very low gait speed and reduced movement amplitudes related to reduced total plantarHexor moment on the paretic side. and in part be related to increased passive stiffness of the plantarHexor muscle-tendon unit of nonreHex origin. Therefore, a reduced level of locomotor function and general disuse could have an impact that eventually contributes to the bilateral locomotor impairment in subjects with hemiparesis.‘“.”

Contribution of Passive Component and Gait Speed The relative contribution of the passive stiffness to the total plantarflexor stiffness remained similar at all three gait speeds tested in controls (natural, slow, very slow), indicating that the very slow walking speed of the patients cannot be regarded as a confounding factor. The lack of a significant correlation between gait speed of the patients and their respective passive stiffness contribution is not surprising, given the proposed role of the passive stiffness. It likely acts as a restraining or controlling force to limit ankle dorsiHexion and then later, the lengthened musculotendinous system may unleash a force to assist the push-off at the end of stance. Moreover. although walking speed is variable among the patients (I9 to 97Scm/ set), the mean lengthening velocity of the plantarflexors during the stance phase is quite low (I 2”/sec to 37”/sec). This low variability in the lengthening velocity of plantarflexor muscles

AFTER STROKE,

357

Lamontagne

suggests little velocity-dependent effect on both the active and passive components contributing to total plantarflexor stiffness during gait. Such a hypothesis, at least in part, is in accordance with previous studies that show plantarHexor passive resistive torque to significantly increase at lengthening velocities of 40”/sec and higher. compared with that measured at 5”/sec.tt~t3

Relationship With Clinical Outcomes The only clinical measure that was statistically correlated (I’ = .60) with the contribution of passive stiffness was the passive dorsiHexion range of motion measured at rest with the knee in extension. Why this measure stood out is difficult to explain. A possible reason may be that patients with a more normal passive dorsiHexion range have the potential for more appropriate plantarHexor dynamics during gait, and rely less on the passive component. From figure 5. one can see that short passive dorsiHexion range is not necessarily associated with high contribution of passive stiffness during gait and, inversely, that normal dorsiHexion range may be associated with high contribution of passive stiffness. Present data also indicate that patients with a very limited range of passive movement at the ankle (- 15” of dorsiflexion) will strongly rely on this adaptive behavior. As for the predictive value of this clinical measure, it remains modest given that it can explain only 36% of the variance of the contribution of passive stiffness to total plantarflexor stiffness. The lack of significant relationships between the contribution of passive stiffness and clinical measures of spasticity and motor impairment is in agreement with previous studies that failed to relate any one of these measures to the development of passive stiffness.3.8 As suggested previous1y.s a multifactorial analysis in a large sample of patients may lead to a better understanding of the factors involved in the development of passive stiffness.

Study Limitations The use of a moment-angle slope in this study, instead of a single point, minimized the error associated to the matching of the ankle angular measures between the gait and the dynamometric measurements. On the other hand, with the present method, one has to assume the existence of a linear relationship between ankle moment and ankle angle for both gait and passive movements. The high correlation coefficients calculated in the controls support the use of the linear model to quantify moment-angle relationship, as previously reported for gait in healthy individuals’z and in children with cerebral palsy.23 The moderate correlation coefficients calculated in the patients, however, suggest a partial disruption of the linear momentangle relationship during gait. The relative small sample size of subjects may also have contributed to a type 2 error.

Clinical Implications Present data further confirm that excessive passive stiffness of the plantarflexors develops early after stroke. Such an increase in passive stiffness resulted in a larger contribution of the passive component to the total planttiexor stiffness during gait, indicating that this abnormal stiffness compensates for a weak plantarflexor active component. Whereas in some patients this adaptation led to the development of a plantarflexor stiffness that was within normal limits, it did not in others who relied even more on the passive component because of a very weak active component. It was also observed that when no dorsiflexion occurred during gait (either because of excessive passive stiffness or enhanced stretch reflexes), as expected because there was no muscle lengthening, this adaptation did not occur. Arch

Phys

Med

Rehabil

Vol 81, March

2ooO

358

PASSIVE

STIFFNESS

DURING

GAIT

The passive range of dorsiflexion measured at rest was the only clinical measure significantly related to the relative contribution of the passive stiffness to the plantarflexor stiffness during gait. Such a relationship means that the patients with the least dorsiflexion at rest had the largest relative contribution of passive stiffness. Examination of individual data points (fig 5) showed that even patients with severe ankle contractures (- 15” of dorsiflexion) were able to use their passive stiffness to augment the total plantarllexor stiffness. On the other hand, patients with a good dorsiflexion at rest may not be able to dorsiflex the ankle during gait, as was the case with two patients who failed to use their passive stiffness to improve their plantarflexor total stiffness. Thus, present data indicate that passive range of motion appears less critical than the occurrence of muscle lengthening during gait (even over a shortened range) for the use of passive stiffness as an adaptation. On the other hand, although maintaining a good range of passive dorsiflexion will not prevent the development of passive stiffness,3 maintaining good ankle mobility may promote the dorsiflexion during gait that is required to make use of the passive component to augment ankle plantarflexor moment.

CONCLUSION The results of this study showed that within 5 months after stroke, subjects with hemiparesis develop a larger proportion of passive forces during stance phase ankle dorsiflexion than do controls, which augments the plantartlexor moment during gait. It is likely that the increased contribution of the passive component acts as an adaptation for a defective muscle active component, helping ankle push-off at the end of the stance phase of gait. Although this mechanism is effective even in patients with a plantarflexor contracture, it cannot come into action if the dorsiflexion movement is prevented, for instance, by enhanced stretch reflexes in early stance. Acknowledgments:

The

authors

thank

Ms. Francine

Dumas,

M.

Daniel Tardif. and M. Francois Comeau for their assistance: we also thank

all the subjects

who volunteered

to participate

in this stidy.

References

1. Siegler S, Moskowitz GD, Freedman W. Passive and active components of the internal moment developed about the ankle joint during human ambulation. J Biomech 1984;17:647-52. 2. Tardieu C, Lespargot A, Tabary C, Bret M-D. Toe-walking in children with cerebral palsy: contributions of contracture and excessive contraction of triceps surae muscle. Phys Ther 1989;69:656-62. 3. Thilmann AF, Fellows SJ, Ross HF. Biomechanical changes at the ankle joint after stroke. J Neurol Neurosurg Psychiatry 1991;54: 134-9. 4. Hufschmidt A, Mauritz K-H. Chronic transformation of muscle in spasticity: a peripheral contribution to increased tone. J Neural Neurosurg Psychiatry 1985;48:676-85. 5. Dietz V, &intern J, Berger W. Electrophysiological studies of gait in spastic&v and rigidity. Brain 198 1: 104:43 l-49. 6. Sinkjaer T: Magnissen I. Passive, intrinsic and reflex-mediated stiffness in the ankle extensors of hemiparetic patients. Brain 1994;117:355-63.

Arch

Phys

Msd

Rehabil

Vol81,

March

2ooO

AFTER

STROKE,

Lamontagne

7. Given JD. Dewald JPA. Rymer WZ. Joint dependent passive stiffness in paretic and contralateral limbs of spastic patients with hemiparetic-stroke. J Neural Neurosurg Psychiatry 19$5;59:27 I-9. 8. Malouin F, Bonneau C. Pichard L. Corriveau D. Non-retlcx mediated changes in plantarflexor muscles early after stroke. Stand J Rehabil Med 1997:29: 147-53. 9. Dietz V. Berger W. Normal and impaired regulation of muscle stiffness in gait: a new hypothesis about muscle hypertonia. Exp Neural 1983;79:680-7. IO. 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 rcflexes. J Neurol Neurosurg Psychiatry 1984;37: 1029-33. II. Lamontagne A, Malouin F. Richards CL. Viscoelastic behavior of plantar flexor muscle-tendon unit at rest. J Orthop Sports Phys Ther 1997;26:244-52. 12. Winter DA. The biomechanics and motor control of human gait: normal, elderly and pathological. Waterloo. Ontario: University of Waterloo Press: 199 I. 13. Lamontagne A, Malouin F. Richards CL, Dumas E Impaired viscoelastic behavior of spastic plantarflexors during passive stretch at different velocities. Clin Biomech 1997; I2:508- 15. 14. Fugl-Meyer AR, Jaasko L. Leyman I. Olsson S. Steglind S. The post-stroke hemiplegic patient: method for evaluation of physical performance. Stand J Rehabil Med I975:7: 13-3 I. 15. Bohannon RW. Smith MB. lnterrater reliability of a modified Ashworth scale of muscle spasticity. Phys Ther 1987;67:206-7. 16. Boiteau M, Malouin F, Richards CL. Use of a hand-held dynamometer and a Kin-Corn dynamometer for evaluating spastic hypertonia in children: a reliability study. Phys Ther 1995;75:796-802. 17. Lamontagne A, Malouin F. Richards CL. Dumas F. Evaluation of reflex- and nonreflex-induced muscle resistance to stretch in adults with spinal cord injury using hand-held and isokinetic dynamometry. Phys Ther 1998:78:964-75. IS. Hof AL. Geelen BA. Van Den Berg JW. Calf muscle moment. work and efficiency in level walking: role of series elasticity. J Biomech 1983;16:523-37. 19. Lowenthal M, Tobis JS. Contractures in chronic neurological disease. Arch Phys Med Rehabil 1957;38:640-7. 20. Shiavi R, Bugle HJ. Limbird T. Electromyographic gait assessment. Part 2: Preliminary assessment of hemiparetic synergy patterns. J Rehabil Res Dev 1987;24:24-30. 21. blney SJ, Richards C. Hemiparetic gait following stroke. Part I: 22.

23.

Characteristics. Gait Posture I996:4: 136-48. Fig0 C. Crenna I? Jensen LM. Moment-angle relationship

at lower limb joints during human walking at -different velbcities. J Electromvom Kinesiol 1996:6: 177-90. Davis RB, be Luca PA. Gait characterization via dynamic joint stiffness. Gait Posture 1996;4:224-3 I. Suppliers

a. Panasonic, 5770 Ambler Drive, Mississauga, Ontario, L4W 2T3 Canada. b. Advanced

Mechanical

Technology.

Inc..

176

Waltham

Watertown, MA 02172. c. Graphic Controls Canada Ltd.. 215 Herbert, Gananoque,

Street. Ontario,

K76 2Y7 Canada.

d. Chattecx Corp., Chattanooga Group. TN 37343. e. Neogenix,

3175

Quatre-Bourgeois

Quebec G 1W 2K7, Canada.

4717

Adams

Road,

Hixson,

Street,

Bureau

100.

Ste-Foy,