Available online at

ScienceDirect www.sciencedirect.com Annals of Physical and Rehabilitation Medicine 57 (2014) 185–192

Original article / Article original

Dynamic EMG of peroneus longus in hemiplegic children with equinovarus Activite´ EMG dynamique du muscle peroneus longus dans l’e´quin varus chez l’enfant he´miple´gique C. Boulay a,*,b,c, M. Jacquemier a,c, V. Pomero a,c, E. Castanier a,c, G. Authier a,c, B. Chabrol b, G. Bollini a,c, J.-L. Jouve a,c, E. Viehweger a,c a

Laboratoire de la marche, service de chirurgie orthope´dique pe´diatrique, CHU La Timone–Enfants, 264, rue Saint-Pierre, 13385 Marseille cedex 05, France b Service de neurologie pe´diatrique, CHU La Timone–Enfants, 13385 Marseille, France c Institut des sciences du mouvement, Aix-Marseille universite´, UMR CNRS 6233, 13385 Marseille, France Received 7 October 2013; accepted 14 February 2014

Abstract Objective. – In hemiplegic children the appearance of equinovarus is correlated with premature electromyography (EMG) activity of the gastrocnemius medialis (GM) prior to initial contact. The goal was to analyze the onset of EMG activation in the GM and, more particularly, the peroneus longus (PL) in cases of equinovarus: is PL activity likewise premature? Material and methods. – As 15 hemiplegic children (age 5 years  1.5) with equinovarus walked, their PL and GM EMG activity was being recorded. The latter was normalized in terms of gait cycle percentage (0–100%) and detected through semi-automatic selection with activation threshold set at 20 mV. A paired t-test compared activation onset of the PL versus the GM muscles. Results. – As regards the healthy limb, activity onset of the GM (+14.55%) and the PL (+19.2%) muscles occurred only during the ST. In cases of equinovarus, activation of the GM ( 5.2%) and the PL ( 6.1%) occurred during the SW and was premature. For each muscle, comparison between the healthy and the hemiplegic side was highly significant (P < 0.001). Conclusion. – Premature PL and GM EMG activity preceding initial contact corresponds not to a disorder secondary to imbalance but rather, more probably, to motor command dysfunction. While the PL consequently contributes to equinus deformity, its possible role in varus genesis is less evident. EMG study needs to be completed by comparing PL and tibialis posterior strength while taking foot bone morphology into full account. # 2014 Elsevier Masson SAS. All rights reserved. Keywords: Peroneus longus; Cerebral palsy; Hemiplegia; Equinus

Re´sume´ Objectif. – L’e´quin varus chez l’enfant he´miple´gique est corre´le´ a` une activation EMG pre´mature´e du gastrocnemius medialis (GM). Le but e´tait d’analyser le de´but d’activation EMG du GM et du peroneus longus (PL) dans l’e´quin varus : le PL est-il pre´mature´ ? Mate´riel et me´thode. – Quinze he´miple´giques (aˆge 5  1,5 ans) marchant en e´quin varus ont eu un EMG PL et GM. L’activite´ EMG (0–100 %) e´tait normalise´e en pourcentage du cycle de marche et de´termine´e par une me´thode semi-automatique avec un seuil d’activation de 20 mV. Un test de Student comparait le de´but d’activation du PL versus GM. Re´sultats. – Coˆte´ sain, le de´but de l’activation du GM (+14,55 %) et du PL (+19,2 %) se faisait uniquement en phase d’appui (NS). Coˆte´ he´miple´gique, GM 5,2 % et PL 6,1 % (NS) e´taient pre´mature´s en phase d’oscillation. Pour chaque muscle, la comparaison sain versus he´miple´gique e´tait tre`s significative ( p < 0,001).

* Corresponding author. E-mail address: [email protected] (C. Boulay). http://dx.doi.org/10.1016/j.rehab.2014.02.004 1877-0657/# 2014 Elsevier Masson SAS. All rights reserved.

186

C. Boulay et al. / Annals of Physical and Rehabilitation Medicine 57 (2014) 185–192

Conclusions. – Cette activite´ pre´mature´e EMG PL et GM pre´ce´dant le contact initial ne correspond pas a` un trouble secondaire a` un de´se´quilibre mais probablement a` un trouble de la commande. Le PL participe donc a` l’e´quin mais il est difficile d’affirmer son roˆle dans la gene`se du varus. Il faut comple´ter l’e´tude EMG avec le tibialis posterior, comparer leur force et prendre en compte l’aspect constitutionnel (osseux). # 2014 Elsevier Masson SAS. Tous droits re´serve´s. Mots cle´s : Peroneus longus ; Paralysie ce´re´brale ; He´miple´gie ; E´quin

1. English version 1.1. Introduction Gage [2] has described 5 factors enabling effective and functional gait; one of them is correct pre-positioning of the foot at the end of the swing phase (SW), which prepares initial contact (IC) by the heel. When this condition is not met, there exists an equinus [2]. In the hemiplegic child with cerebral palsy (CP), equinus deformities are likely to develop as the diseased foot continues to grow. But before the bone deformities are set into place, abnormal electromyography (EMG) activity sequences for the gastrocnemius medialis (GM), as well as other muscles, are likely to trigger equinus. This will occur at a very early stage, while the infant is learning to walk [2,13,14]; during the SW, EMG activation of the GM is premature. Early EMG in the cerebral palsy child is particularly instructive. In cases of equinovarus, premature GM and tibialis posterior (TP) activation has been observed repeatedly [13,14,22]. More precisely, premature EMG activity has been shown to occur at the end of the SW; one of Gage’s conditions for functional gait has not been met [2]; IC will involve not the heel but the fifth metatarsal head, which means that equinovarus has come into being. As concerns equinovarus in a CP child, Perry [13] took note of EMG activation during the SW of the PL, and it was deemed paradoxical due to the action of the equinovalgus deformity of the PL [1]. However, Perry [13] had neither studied nor proposed a pathophysiological explanation for his observation. While examining hemiplegic CP children in our laboratory, we made the same EMG-based clinical observation. That is why the objective of this study was to evaluate and quantify the EMG activation sequence of the PL and GM muscles in hemiplegic children presenting with equinovarus. Indeed, to our knowledge the EMG activation sequence common to PL and GM in hemiplegic CP children has yet to be described in any published study. We wished to determine whether or not the synergic action of PL and GM in cases of equinovarus should be put forward as a hypothesis.

neurology unit. Inclusion criteria were: CP diagnosis by a child neurologist (B.C.), child under 6 years of age, right or left hemiplegia, good functional level (Gross Motor Function Classification System 1 [GMFCS 1], validated only for children aged from 6 to 12 years). Clinical examination and video images allowed for definition of the positions of the ankle and the forefoot during IC. In clinical examination of the plantar side of the feet, reproducibility of the equinovarus was revealed by hyperkeratosis face to the 5th metatarsal head (M5), which objectified observation of preferred M5 placement. According to Wren et al. [23], equinus is defined by ankle plantar flexion lower than 08. The key equinovarus criterion was IC with the M5 (Fig. 1). A posterior view shows the position of the forefoot as initial M5 contact is taking place (Fig. 1). Children presenting with equinovarus foot deformity confirmed by two therapists (a physical and rehabilitation medicine specialist and a physiotherapist) were included in the study. Goniometric evaluation aimed at measuring dorsiflexion amplitudes of the ankle between tibia and hindfoot was performed by two practitioners (C.B. and G.A.) with the child placed in a supine position, his or her knees first bent and then stretched; care was taken to avoid pronation or supination of the forefoot. Criteria of exclusion consisted in an equinus deformity involving retraction of the triceps sural (calf muscle) and a difference in length of the lower limbs to the detriment of the hemiplegic side greater than 1 cm. Fifteen hemiplegic children were included: 9 with deformities on the left and 6 on the right; 7 girls and 8 boys. Average age was 5  1.5 years. The mean number of gait cycles analyzed was 15  10. The children walked barefooted in accordance with their spontaneous walking velocity. Some authors [13,14,16,17] have

1.2. Materials and method We extracted from our database a series of CP children who had been given a dynamic EMG analysis. 1.2.1. Clinical data Our retrospective study was carried out from 2008 until 2011. All of the children were monitored by our pediatric

Fig. 1. Right hemiplegia (posterior view). Equinovarus: initial contact occurs at the level of the 5th metatarsal head (M5).

C. Boulay et al. / Annals of Physical and Rehabilitation Medicine 57 (2014) 185–192

demonstrated that muscle activation depends on speed. That is why, in order to optimally compare the hemiplegic side with the healthy side, we opted in favor of the most easily reproducible conditions: shoeless and with the speed specific to each child (rather than an imposed pace). 1.2.2. Data acquisition The children were given an EMG test of the healthy side and the hemiplegic side. The EMG data were recorded through use of a Wifi system (Zero Wire Aurion, Italy). The EMG signals were picked up at an acquisition frequency of 2000 Hz, amplified 1000, filtered (10 Hz high-pass filter, 500 Hz lowpass filter) and rectified. The skin had been preliminarily degreased, stripped and cleaned. Two electrodes (Ag-AgCl, diameter 10 mm) were positioned parallel to the PL fibers at ¼ of a line running from the fibular head to the edge of the lateral malleolus. The SENIAM project (Surface EMG for Non-Invasive Assessment of Muscles) [7,12] recommends a distance of 10 mm between the electrode centers. Standard SENIAM protocol was scrupulously respected in view of minimizing interference or cross-talk. More specifically, validation of electrode positioning was aimed at minimizing cross-talk between not only GM and PL but also PL and tibialis anterior (TA), and it was carried out first by plantar flexion movements with stretched knees (GM testing) and eversion (PL), and then by dorsal inversionflexion (TA). The objective of the validation procedure was to observe asynchronicity in the bursts of EMG activity during phases of voluntary movement. EMG activity onset was identified through use of a semi-automatic method; the EMG signal for a given muscle was defined as active when the rectified raw signal exceeded 20 microvolts [11]. Indeed, 20 microvolts is the commonly recognized threshold usually applied when studying motor unit potential [11,15] and attempting to exclude any artifact. 1.2.3. Data analysis The EMG activation sequences were analyzed in terms of timing with regard to the gait cycle. Muscle activity was normalized as gait cycle percentage. Onset of the activation sequence for the PL and GM muscles was expressed according to the latter (0–100%). Perry [13] has defined premature GM activity as beginning during the SW, which means prior to IC; at this stage the GM muscle presents no EMG activity in healthy children. Using the same method, Perry [13] defined onset of premature PL activity during the SW preceding IC; the PL shows no activity in healthy children at this stage. Premature EMG activation of the GM and PL muscles was consequently normalized and expressed as gait cycle percentage. By convention, EMG activation occurring prior to IC (0%) was noted in negative terms; all premature activity was considered as negative. For example, if premature activity commenced at 10% before IC, it was identified as 10%. The onset of PL and GM activation could consequently be studied and compared. A paired t-test was used. Activation onset was compared between PL and

187

GM muscles and between the hemiplegic and the healthy sides. Activation onset for PL and GM on the healthy side was compared to the data reported by Perry [13], which were considered as references. Normal distribution of the parametric variables was verified by the Kolmogorov-Smirnov test and by equality of variance tests in order to ensure that the conditions required for application of the comparison of means (Student’s) test had been fulfilled. Statistical significance was set at P < 0.05. 1.3. Results 1.3.1. The clinic The mean for ankle dorsiflexion (bended knees) was 25  58 on the hemiplegic side and 28  98 on the healthy side (P < 0.05). The mean for ankle dorsiflexion (stretched knees) was 15  108 on the hemiplegic side versus 20  88on the healthy side (P < 0.05) (Table 1). As regards the length of each lower limb, mean intraobserver variability was 3 mm; with an extra-observer, it rose to 5 mm. ST duration was 62% on the healthy side and 55% on the hemiplegic side. 1.3.2. Electromyography On the healthy side, mean PL activation onset occurred at 19.2% of the gait cycle, while on the hemiplegic side, the analogous mean was 6.1% (P < 0.001) (Table 2; Fig. 2). The PL muscle was prematurely activated during the SW. On the healthy side, mean GM activation onset occurred at 14.55% of the gait cycle, while on the hemiplegic side, the analogous mean was 5.2% (P < 0.001) (Table 2) (Fig. 2). The GM muscle was prematurely activated during the SW. Both the PL and the GM muscles showed premature activity with regard to IC, but presented no significant difference as concerns activation onset (Table 3; Fig. 2). Table 1 Initial contact and dorsiflexion amplitudes of the tibiotarsal joints. Healthy side (n = 15) Mean  S.D.

Hemiplegic side (n = 15) Mean  S.D.

Dorsiflexion ankle Bended knee Stretched knee

20  58 18  48

15  58 14  78

Initial contact

Heel

M5

M5: 5th metatarsal head. Table 2 Activation onset of the GM and PL muscles (in gait cycle %): comparison of mean (Student’s test), healthy side versus hemiplegic side. Activation onset

Healthy side

Hemiplegic side

t-test

Activation sequence GM (mean/S.D.) PL (mean/S.D.)

Normal +14.55% / 11 +19.2% / 11.8

Prematurity 5.2% / 12.4 6.1% / 15.1

p < 0.001 p < 0.001

S*** S***

GM: gastrocnemius medialis; PL: peroneus longus; S*: P < 0.05; S**: P < 0.01; S***: P < 0.001.

188

C. Boulay et al. / Annals of Physical and Rehabilitation Medicine 57 (2014) 185–192 Table 3 Activation onset of the GM and PL muscles (in gait cycle %): comparison of mean (Student’s test) on the hemiplegic side. Activation onset

Hemiplegic side

t-test

Activation sequence GM (mean/S.D.) PL (mean/S.D.)

Prematurity 5.2% / 12.4 6.1% / 15.1

P = 0.8

NS

GM: gastrocnemius medialis; PL: peroneus longus.

Fig. 2. During gait cycle: in percentage (%), onset of EMG activation of the GM and PL muscles on the healthy side versus the hemiplegic side. M5: 5th metatarsal head; GM: gastrocnemius medialis; PL: peroneus longus.

1.4. Discussion In a population of hemiplegic children, the presence of an equinus can be attributed to premature EMG activity of the GM muscle especially. Indeed, EMG activations have been documented in equinovarus with regard to the varus muscles. Abnormal premature PL activation has already been described in hemiplegic children with equinovalgus deformity [1]. In addition, Perry [13] observed premature EMG activation of the PL muscle during the SW in cases of equinovarus diagnosed amongst a population of CP children. Our results quantify the prematurity and the observations made by Perry. As in any pediatric study, especially with very young patients, variability of gait cycles may constitute a bias, as could the small population size (n < 30) and the absence of a control group. Moreover, interpretation of the EMG dynamic in pediatrics necessitates a number of references to the literature. For some authors, muscle activation sequences are considered as mature from the age of 2 years [14]. While Sutherland et al. [17] described highly exceptional premature activation sequences in children from 2 to 7 years of age, the population studied by his team was very small. Moreover, exceptional degrees of prematurity were not confirmed by Shiavi et al. [16], who did not determine a difference between the ages of 4 and 11 years. More recently, Granata et al. [6] demonstrated that EMG activation sequences pertaining to the lower limbs came into play well before the age of 7 years. In our study, stance phase (ST) duration on the healthy side (toe off 62%) was in accordance with the data in the literature [13]. On the same token, diminution of the ST on the hemiplegic side (toe off 55%) signifies the avoidance strategy or transfer of the center of gravity to the healthy side that has

been described in the CP child [17]. Gait cycle variability in children was evaluated through selection of the most repeatable and reproducible cycles. During IC, equinovarus was defined as contact with the 5th metatarsal head (Fig. 1) without the gradual knee bending that could have provoked its occurrence, which is seldom observed in hemiplegic patients. From a clinical standpoint and in the video images, the lower limbs presented no torsional abnormality, nor did any retraction of the triceps sural muscle occur. Nevertheless, ankle dorsiflexion was significantly diminished on the hemiplegic side, which meant that early modifications of the capusloligamentous, musculoaponeurotic and tendinous structures had indeed taken place. Since the latter are a source of passive resistance to dorsiflexion, they contribute to development and fixation of the equinus. According to the data in the literature, normal GM and PL activation sequences take place only during the STon the healthy side [1,13]. The two muscles show no activity during the SW, which means that previous findings have been corroborated in our study (Table 2). As regards the hemiplegic side, activation onset is premature (Table 2), but there is no significant difference in prematurity between the GM and PL muscles. In the literature, the main cause of equinovarus is overactivity involving primarily the triceps sural, and also the TP. The premature EMG activity of the PL during SW, which means in an open kinetic chain, does not correspond to an activation disorder secondary to compensation or disequilibrium, which is observed during ST in a closed kinetic chain as a means of stabilizing [13] the foot and the ankle. As a result, premature PL and GM activity in the SW would more likely be due to a problem pertaining to primary EMG activation, i.e. to misdirected commands in CP cases. Premature synergic PL and GM activity could consist in spastic co-contractions of the two muscles. It has been shown [3,4,19,20] that such co-contraction is one of the forms of muscle overactivity interfering with walking in a spastic paresis subject. Gracies et al. [5] attributed spastic co-contraction to a misdirected descending command from dorsiflexor to plantar flexor, as is seen in CP; exaggerated contraction of the plantar flexors of the ankles (PL and GM) is likely to be triggered by the ankle dorsiflexion effort observed during SW. To sum up, in synergy with the GM, the PL muscle strongly contributes to occurrence of equinus during SW, and the deleterious effects of this action may tend to be underestimated [18]. That said, and even though Thevenon et al. [18] mention this type of causation with in equinovarus in a hemiplegic adult after an ischemic stroke, it would appear problematic to

C. Boulay et al. / Annals of Physical and Rehabilitation Medicine 57 (2014) 185–192

underline the role of the PL in the genesis of the varus. While this muscle exerts an locking action on the midfoot in the middle of the ST accompanying the 1st rocker, it does not act upon the tibiotarsal joint [8]. On the other hand, the PL provides support for the longitudinal and transversal arches of the foot. At the end of the ST during the 3rd rocker for the push-off, the PL again exerts locking action, this time on the 1st and 2nd rays i.e. the medial fore foot; the forefoot is stabilized as the heel is loosened [8]. So it is that the PL brings about forefoot pronation as a means of compensation for the action of the varus muscles (essentially the GM and the TP) [10,14,15]. In equinus, however, the PL relinquishes its function as stabilizer and support provider of the 1st ray [9]. To conclude, further pathophysiological understanding entails conjoined exploration of PL and TP muscles in terms of EMG activation sequences, particularly with reference to the co-contraction indexes [19,20] involving the two muscles, the TA recruitment index [19,20] and a strength index, as well. The other key element to be taken into account is the constitutional or osseous aspect, which should be assessed both clinically (forefoot, mid- and hindfoot) and radiographically, so that foot morphology development can be monitored during growth. Westberry et al. [21] have described the pediatric radiographic norms in the hemiplegic child through which it is possible to quantify morphology of the forefoot (abduction/adduction, planus/cavus), midfoot (abduction/adduction, pronation/supination) and hindfoot (calcaneus/equinus, valgus/varus). Disclosure of interest The authors declare that they have no conflicts of interest concerning this article. 2. Version franc¸aise 2.1. Introduction Gage [2] a de´crit 5 facteurs pour une marche fonctionnelle et efficace dont le pre´-positionnement correct du pied en fin de phase d’oscillation (PO) afin de pre´parer un contact initial (CI) par le talon. Si ce pre´-requis n’est pas respecte´, il existe un e´quin [2]. Chez l’enfant he´miple´gique par paralysie ce´re´brale (PC), les de´formations dues a` l’e´quin ont un potentiel d’e´volutivite´ e´tant donne´ la croissance du pied a` venir. Avant que les de´formations osseuses ne soient fixe´es, ce sont les troubles de la se´quence d’activation e´lectromyographique (EMG) du gastrocnemius medialis (GM), entre autres, qui vont entraıˆner l’e´quin, et ce de fac¸on tre`s pre´coce lors de l’acquisition de la marche [2,13,14] : la se´quence d’activation EMG du GM est pre´mature´e en PO. L’EMG pre´coce chez le jeune enfant atteint de PC est tre`s informatif. Dans l’e´quin varus, une activation pre´mature´e du GM et du tibialis posterior (TP), essentiellement, est constate´e [13,14,22]. En effet cette activation EMG pre´mature´e se fait en fin de PO. Un des pre´-requis de Gage n’est pas respecte´ [2] : le CI ne se fera pas par le talon mais sur la teˆte du 5e me´tatarsien, un e´quin varus est constitue´.

189

Dans l’e´quin varus chez l’enfant PC, Perry [13] avait constate´ la pre´sence paradoxale d’une activation EMG en PO du PL : action paradoxale e´tant donne´ l’action d’e´quin valgus du PL [1]. Cependant Perry [13] n’avait pas e´tudie´ ni donne´ d’explication physiopathologique a` ce constat. Lors des e´valuations des enfants he´miple´giques par PC au sein de notre laboratoire, nous avons fait cette meˆme constatation clinique et EMG. C’est pourquoi le but de ce travail e´tait d’e´valuer et de quantifier la se´quence d’activation EMG du PL et du GM chez les enfants he´miple´giques pre´sentant un e´quin varus. Cette e´tude de se´quence d’activation commune PL et GM n’a pas e´te´ de´crite chez l’enfant he´miple´gique par PC. L’hypothe`se e´tait de mettre en e´vidence ou non l’action synergique du PL avec le GM dans l’e´quin varus. 2.2. Mate´riels et me´thode Nous avons extrait de notre banque de donne´es une se´rie d’enfants PC qui avaient be´ne´ficie´ d’une analyse EMG dynamique. 2.2.1. Donne´es cliniques Il s’agissait d’une e´tude re´trospective entre 2008 et 2011. Tous les enfants e´taient suivis dans le service de neurologie pe´diatrique. Les crite`res d’inclusion e´taient : le diagnostic de PC avait e´te´ re´alise´ par un neurope´diatre (B.C.), un aˆge infe´rieur a` 6 ans, une he´miple´gie droite ou gauche, un bon niveau fonctionnel (Gross Motor Function Classification System 1 [GMFCS 1], valide´ uniquement pour les enfants aˆge´s de 6 a` 12 ans). L’examen clinique et la vide´o permettaient de de´finir la position de la cheville et de l’avant-pied lors du CI. Lors de l’examen clinique de la face plantaire des pieds, la reproductibilite´ de l’e´quin varus e´tait re´ve´le´e par l’hyperke´ratose en regard de la teˆte de M5 objectivant l’appui pre´fe´rentiel a` ce niveau sur la teˆte M5. Selon Wren et al. [23], l’e´quin est de´fini comme une flexion plantaire de la cheville en dessous de 08. Le crite`re de l’e´quin varus e´tait un CI qui se faisait avec la teˆte du 5e me´tatarsien (M5) (Fig. 1). Une vue poste´rieure lors du CI re´ve`le la position de l’avant-pied lors du contact initial par M5 (Fig. 1).

Fig. 1. He´miple´gie droite (vue poste´rieure). E´quin varus : le contact initial se fait au niveau de la teˆte du 5e me´tatarsien (M5).

190

C. Boulay et al. / Annals of Physical and Rehabilitation Medicine 57 (2014) 185–192

Les enfants pre´sentant un e´quin varus confirme´ par deux the´rapeutes (un me´decin de me´decine physique et de re´adaptation et un masseur kine´sithe´rapeute) sont inclus dans l’e´tude. Un bilan goniome´trique mesurant les amplitudes de flexion dorsale de la cheville entre le tibia et l’arrie`re-pied a e´te´ re´alise´ par deux praticiens (C.B. et G.A.) : en de´cubitus dorsal, mesure de la flexion dorsale de cheville d’abord genoux fle´chis puis genoux tendus en s’assurant de l’absence de pronation ou de supination de l’avant-pied. Les crite`res d’exclusion e´taient la pre´sence d’une de´formation en e´quin avec une re´traction du triceps sural et une ine´galite´ de longueur des membres infe´rieurs supe´rieure a` 1 cm aux de´pens du coˆte´ he´miple´gique. Quinze enfants he´miple´giques ont e´te´ inclus (9 gauches et 6 droits) dont 7 filles et 8 garc¸ons. L’aˆge moyen e´tait de 5  1,5 ans. Le nombre moyen de cycles e´tudie´s e´tait de 15  10. Les enfants marchaient pieds nus suivant leur vitesse de marche spontane´e. Certains auteurs [13,14,16,17] avaient de´montre´ que l’activation musculaire e´tait de´pendante de la vitesse. C’est pourquoi nous avons choisi les conditions les plus reproductibles a` savoir l’e´valuation pieds nus et avec une vitesse de marche propre a` l’enfant (et non impose´e) afin de pouvoir comparer au mieux le coˆte´ he´miple´gique versus le coˆte´ sain. 2.2.2. Acquisition des donne´es Les enfants ont be´ne´ficie´ d’un examen EMG du coˆte´ sain et du coˆte´ he´miple´gique. Les donne´es EMG ont e´te´ enregistre´es en utilisant un syste`me Wifi (Zero Wire Aurion, Italie). Les signaux EMG e´taient acquis avec une fre´quence d’acquisition de 2000 Hz, amplifie´ 1000, filtre´ (10 Hz high-pass filter, 500 Hz low-pass filter) et rectifie´s. La peau avait e´te´ de´graisse´e, de´cape´e et nettoye´e. Deux ´electrodes (Ag-AgCl), 10 mm de diame`tre, e´taient place´es de fac¸on paralle`le aux fibres du PL a` 25 % d’une ligne allant de la teˆte de la fibula jusqu’a` la pointe de la malle´ole late´rale. La SENIAM (Surface EMG for Non-Invasive Assessment of Muscles) [7,12] recommande une distance inter-e´lectrode de 10 mm entre le centre de chaque e´lectrode. Le protocole de standardisation de la SENIAM a e´te´ respecte´ notamment pour minimiser la diaphonie (ou cross-talk). La validation pour minimiser la diaphonie entre le GM et le PL, mais e´galement entre le PL et le TA, a e´te´ re´alise´e graˆce a` des mouvements de flexion plantaire genoux tendus pour le testing du GM et par une e´version pour le PL puis par une inversion-flexion dorsale pour le TA. Le but de cette proce´dure de validation de la position des e´lectrodes e´tait d’observer l’asynchronisme dans le temps des bouffe´es d’activite´ EMG lors des mouvements volontaires. Le de´but de l’activite´ EMG e´tait identifie´ en utilisant une me´thode semi-automatique : le signal EMG du muscle e´tait de´fini comme actif lorsque le signal brut rectifie´ e´tait au-dela` de 20 microvolts [11]. Vingt microvolts est le seuil commune´ment admis pour e´tudier le potentiel d’unite´ motrice [11,15] et exclure l’arte´fact. 2.2.3. Analyse des donne´es Les se´quences d’activation EMG e´taient analyse´es en termes de timing par rapport au cycle de marche. L’activite´ musculaire

e´tait normalise´e en pourcentage du cycle de marche. Le de´but de la se´quence d’activation musculaire du PL et du GM e´tait exprime´ en fonction du pourcentage de cycle de marche (0– 100 %). Perry [13] avait de´fini une activite´ pre´mature´e du GM lorsque le de´but de celle-ci de´butait durant la phase d’oscillation, i.e. avant le CI : le GM n’a pas d’activite´ EMG pendant la PO chez les enfants sains. Selon la meˆme me´thode Perry [13] de´finissait un de´but d’activite´ pre´mature´e du PL pendant la phase d’oscillation pre´ce´dant le CI : le PL n’a pas d’activite´ EMG pendant la PO chez les enfants sains. Ainsi la pre´maturite´ de l’activation EMG du GM et du PL a e´te´ normalise´e et exprime´e en pourcentage du cycle de marche. Par convention une activation EMG qui survenait avant le CI (0 %) e´tait note´e en ne´gatif : toute activite´ pre´mature´e e´tait donc conside´re´e comme ne´gative. Par exemple si l’activite´ pre´mature´e de´butait a` 10 % avant le contact initial, elle e´tait identifie´e a` 10 %. Les de´buts de l’activation du PL et du GM e´taient donc e´tudie´s et compare´s. Un test de Student a e´te´ utilise´. Le de´but de l’activation e´tait compare´ entre PL et GM du coˆte´ he´miple´gique et du coˆte´ sain. Le de´but de l’activation du coˆte´ sain pour le PL et le GM e´tait compare´ avec les donne´es de Perry [13], conside´re´es comme des re´fe´rences. La distribution normale des variables parame´triques e´tait ve´rifie´e par le test de Kolmogorov-Smirnov et par les tests d’e´galite´ des variances afin que les conditions soient ve´rifie´es pour appliquer le test de comparaison de moyennes (Student). La significativite´ statistique e´tait de´termine´e par un p < 0, 05. 2.3. Re´sultats 2.3.1. La clinique La moyenne de la flexion dorsale de la cheville (genoux fle´chis) e´tait de 25  58 du coˆte´ he´miple´gique et de 28  98 du coˆte´ sain ( p < 0,05). La moyenne de la flexion dorsale de la cheville (genoux e´tendus) e´tait de 15  108 du coˆte´ he´miple´gie versus 20  88 du coˆte´ sain ( p < 0,05) (Tableau 1). Concernant la longueur de chaque membre infe´rieur, la moyenne de la variabilite´ intra-observateur e´tait de 3 mm et en extra-observateur elle e´tait de 5 mm. La dure´e de la phase d’appui e´tait de 62 % du coˆte´ sain et de 55 % du coˆte´ he´miple´gique. 2.3.2. L’EMG Le PL avait, du coˆte´ sain, une moyenne de de´but d’activation a` 19,2 % du cycle de marche alors qu’elle e´tait a` 6,1 % Tableau 1 Contact initial et amplitudes de la flexion dorsale des articulations tibiotarsiennes. Coˆte´ sain (n = 15= Moyenne  DS

Coˆte´ he´miple´gique (n = 15) Moyenne  DS

Flexion dorsale cheville Genou fle´chi 20  58 18  48 Genou tendu

15  58 14  78

Contact initial

M5

Talon e

M5 : teˆte du 5 me´tatarsien.

C. Boulay et al. / Annals of Physical and Rehabilitation Medicine 57 (2014) 185–192 Tableau 2 De´but d’activation du GM et du PL (en % du cycle de marche) : comparaison de moyenne (test de Student), coˆte´ sain versus coˆte´ he´miple´gique. De´but d’activation

Coˆte´ sain

Coˆte´ he´miple´gique t-test

Pre´maturite´ Se´quence d’activation Normale +14,55 % / 11 5,2 % / 12,4 GM (moyenne/DS) PL (moyenne/DS) +19,2 % / 11,8 6,1 % / 15,1

p < 0,001 S*** p < 0,001 S***

191

Tableau 3 De´but d’activation du GM et du PL (en % du cycle de marche) : comparaison de moyenne (test de Student) du coˆte´ he´miple´gique. De´but d’activation

Coˆte´ he´miple´gique

t-test

Se´quence d’activation GM (moyenne/DS) PL (moyenne/DS)

Pre´maturite´ 5,2 % / 12,4 6,1 % / 15,1

p = 0,8

NS

GM : gastrocnemius medialis ; PL : peroneus longus.

GM : gastrocnemius medialis ; PL : peroneus longus.

( p < 0,001) du coˆte´ he´miple´gique (Tableau 2) (Fig. 2). Le PL s’activait pre´mature´ment en PO. Le GM avait, du coˆte´ sain, une moyenne de de´but d’activation a` 14,55 % du cycle de marche alors qu’elle e´tait a` 5,2 % ( p < 0,001) du coˆte´ he´miple´gique (Tableau 2) (Fig. 2). Le GM s’activait pre´mature´ment en PO. Ainsi PL et GM avaient une activite´ pre´mature´e par rapport au contact initial mais n’avaient pas de diffe´rence significative au niveau de leur de´but d’activation (Tableau 3) (Fig. 2).

Comme toute e´tude en pe´diatrie notamment chez les plus jeunes, il existe une variabilite´ sur les cycles de marche ce qui est un biais possible ainsi que le faible e´chantillon (n < 30) et l’absence de groupe te´moin. L’interpre´tation de l’EMG dynamique en pe´diatrie ne´cessite des pre´cisions. Pour certains auteurs, les se´quences d’activation musculaire sont matures a` partir de 2 ans [14]. Sutherland et al. [17] ont de´crit des se´quences d’activation pre´mature´e tout a` fait exceptionnelles entre 2 et 7 ans mais sur de tre`s faibles e´chantillons. Les degre´s de pre´maturite´ exceptionnels n’ont pas e´te´ confirme´s par Shiavi et al. [16] qui n’avaient pas de´termine´ de diffe´rence entre l’aˆge de 4 et de 11 ans. Re´cemment Granata et al. [6] ont e´tabli que les se´quences d’activation EMG des membres infe´rieurs e´taient effectives bien avant 7 ans. La dure´e de la phase d’appui du coˆte´ sain (toe off 62 %) est conforme aux donne´es de la litte´rature [13]. De meˆme la re´duction de la phase d’appui du coˆte´ he´miple´gique (toe off 55 %) signifie une esquive de l’appui du coˆte´ he´miple´gique de´ja` de´crit chez l’enfant PC [13]. La variabilite´ des cycles de marche chez les enfants e´tait e´value´e par une se´lection la plus re´pe´table et reproductible des cycles de marche. Lors du CI, l’e´quin varus e´tait de´fini par un contact avec la teˆte du 5e me´tatarsien (Fig. 1) sans flexion progressive du genou qui aurait pu provoquer un varus e´quin (rare chez les he´miple´giques). On ne retrouvait pas de trouble torsionnel des membres infe´rieurs sur le plan clinique et sur la vide´o. Il n’y avait pas de re´traction du triceps sural dans la population e´tudie´e. Ne´anmoins la flexion dorsale de cheville e´tait significativement moindre du coˆte´ he´miple´gique : il existe, donc, pre´cocement, des modifications des structures capsulo-ligamentaires, musculo-apone´vrotiques et tendineuses. Ces dernie`res, a` l’origine d’une re´sistance passive a` la flexion dorsale, vont concourir a` l’e´volution de l’e´quin et a` sa fixation. Conforme´ment aux donne´es de la litte´rature, les se´quences d’activation normales du GM et du PL se font uniquement pendant la phase d’appui du coˆte´ sain [1,13]. Ces deux muscles n’ont aucune activite´ pendant la phase d’oscillation ce que l’on retrouve bien dans notre e´tude (Tableau 2). Du coˆte´ he´miple´gique, le de´but de la phase d’activation est pre´mature´e (Tableau 2) mais il n’y a pas de diffe´rence significative entre la pre´maturite´ du PL et du GM. Dans la litte´rature la cause principale de l’e´quin varus est due a` une hyperactivite´ du triceps sural essentiellement mais e´galement du TP. Cette activite´ pre´mature´e EMG du PL pendant la PO, i.e. en chaıˆne ouverte, ne correspond pas a` un trouble de l’activation

2.4. Discussion Dans une population d’enfants he´miple´giques, la pre´sence d’un e´quin peut eˆtre explique´e par une activite´ EMG pre´mature´e du GM. Les activations EMG ont bien e´te´ documente´es dans l’e´quin varus concernant les muscles varisants. L’activation pre´mature´e anormale du PL a e´te´ de´crite chez des enfants he´miple´giques avec un e´quin valgus [1]. Par contre Perry [13] avait constate´ l’activation EMG pre´mature´e en PO du PL dans l’e´quin varus dans une population d’enfants PC. Nos re´sultats quantifient cette pre´maturite´ et les constatations faites par Perry.

Fig. 2. Pendant le cycle de marche : en pourcentage (%), le de´but d’activation EMG du GM et du PL du coˆte´ sain versus coˆte´ he´miple´gique. M5 : teˆte du 5e me´tatarsien ; GM : gastrocnemius medialis ; PL : peroneus longus.

192

C. Boulay et al. / Annals of Physical and Rehabilitation Medicine 57 (2014) 185–192

secondaire a` une compensation ou a` un de´se´quilibre (que l’on voit pendant la phase d’appui en chaıˆne ferme´e pour stabiliser [13] le pied et la cheville). Ainsi l’activite´ pre´mature´e du PL et du GM en PO serait plus due a` un trouble de l’activation primaire EMG i.e. de la commande dans le cadre de la PC. Ces activite´s pre´mature´es synergiques du PL et du GM pourraient eˆtre des cocontractions spastiques entre ces 2 muscles. Cette cocontraction spastique est une des formes d’hyperactivite´ musculaire qui entrave la marche chez un sujet atteint de pare´sie spastique [3,4,19,20]. Gracies et al. [5] expliquent cette cocontraction spastique par une mauvaise distribution de la commande descendante comme on le voit dans la PC : cette contraction exage´re´e des fle´chisseurs plantaires de la chevilles (PL et GM) serait de´clenche´e lors d’un effort de flexion dorsale de la cheville tel qu’on l’observe en PO. Ainsi le PL a un roˆle majeur, en synergie avec le GM, dans la gene`se en PO de l’e´quin : cette action de´le´te`re est peut-eˆtre sous-estime´e [18]. Mais il est difficile d’affirmer son roˆle dans la gene`se du varus bien que Thevenon et al. [18] le mentionnent dans un e´quin varus chez un adulte he´miple´gique post-accident vasculaire ce´re´bral. En effet le PL a une action de verrouillage du me´dio-pied pendant le milieu de la phase d’appui pour accompagner le 1er pivot mais le PL n’agit pas sur la tibio-tarsienne [8]. Le PL soutient les arches longitudinales et transversales. En fin de phase d’appui lors du 3e pivot pour la propulsion, le PL a un effet de verrouillage du 1er et du 2e rayon : l’avant-pied est stabilise´ lors du de´collement du talon [8]. Ainsi le PL re´alise une pronation de l’avant-pied pour compenser l’action des muscles varisants (GM et TP essentiellement) [10,14,15]. Cependant, en e´quin, le PL perd sa fonction de stabilisation du 1er rayon en appui [9]. C’est pourquoi toute avance´e physiopathologique passe par l’exploration conjointe PL et TP en termes de se´quence d’activation EMG notamment avec des indices de cocontraction [19,20] entre PL et TP, d’indice de recrutement du TA [19,20] et aussi de force. L’autre composante qui est importante a` prendre en compte est l’aspect constitutionnel, i.e. osseux, qui doit eˆtre e´value´ sur le plan clinique (par exemple supination de l’avant-pied, valgus de l’arrie`re-pied) et radiographique (permettant de suivre l’e´volution de la morphologie du pied pendant la croissance). En effet Westberry et al. [21] ont de´crit des normes radiographiques pe´diatriques chez l’enfant he´miple´giques pour quantifier la morphologie de l’avant-pied (abduction/adduction, planus/cavus), du me´dio-pied (abduction/adduction, pronation/supination) et de l’arrie`re-pied (calcaneus/equinus, valgus/varus). De´claration d’inte´reˆts Les auteurs de´clarent ne pas avoir de conflits d’inte´reˆts en relation avec cet article.

References [1] Boulay C, Pomero V, Viehweger E, Glard Y, Castanier E, Authier G, et al. Dynamic equinus with hindfoot valgus in children with hemiplegia. Gait Posture 2012;36:108–12. [2] Gage J, editor. The treatments of gait problems in cerebral palsy. 2nd ed, Oxford: Mac Keith Press; 2004. [3] Gracies JM. Pathophysiology of spastic paresis. I: paresis and soft tissue changes. Muscle Nerve 2005;31:535–51. [4] Gracies JM. Pathophysiology of spastic paresis. II: emergence of muscle overactivity. Muscle Nerve 2005;31:552–71. [5] Gracies JM, Wilson L, Gandevia SC, Burke D. Stretched position of spastic muscles aggravates their co-contraction in hemiplegic patients. Ann Neurol 1997;42:438–9. [6] Granata K, Padua D, Abel M. Repeatability of surface EMG during gait children. Gait Posture 2005;22:346–50. [7] Hermens H, Freriks B, Merletti R. SENIAM 8: European recommendations for surface electromyography. Results of the SENIAM project Roessingh Research and Development bv; 1999 [ISBN 90-75452-15-2]. [8] Johnson C, Christensen J. Biomechanics of the first ray. Part I. The effects of Peroneus longus function: a three-dimensional kinematic study on a cadaver model. J Foot Ankle Surg 1999;38:313–21. [9] Johnson C, Christensen J. Biomechanics of the first ray part V: the effect of equinus deformity. A 3-dimensional kinematic study on a cadaver model. J Foot Ankle Surg 1999;44:114–20. [10] Jonsson B, Rundgren A. The peroneus longus and brevis muscles. A roentgenologic and electromyographic study. Electromyography 1971;11: 93–103. [11] Kimura J. Electrodiagnosis in diseases of nerve and muscle. Principles and practice. New York: Oxford University Press; 2001. [12] Merletti R, Hermens H. Introduction to the special issue on the SENIAM European Concerted Action. J Electromyogr Kinesiol 2000;10:283–6. [13] Perry J, editor. Gait analysis: normal and pathological function. New York: McGraw Hill; 1992. [14] Perry J, Hoffer MM, Giovan P, Antonelli D, Greenberg R. Gait analysis of the triceps surae in cerebral palsy. A preoperative and postoperative clinical and electromyographic study. J Bone Joint Surg Am 1974;56:511–20. [15] Santilli V, Frascarelli MA, Paoloni M, Frascarelli F, Camerota F, De Natale L, et al. Peroneus longus muscle activation pattern during gait cycle in athletes affected by functional ankle instability: a surface electromyographic study. Am J Sports Med 2005;33:1183–7. [16] Shiavi R, Green N, McFayden B, Frazer M, Chen J. Normative childhood EMG gait patterns. J Orthop Res 1987;5:283–95. [17] Sutherland D, Olshen R, Biden E, Wyatt M. The development of mature walking. London: Mac Keith Press; 1988. [18] Thevenon A, Serafi R, Fontaine C, Grauwin MY, Buisset N, Tiffreau V. An unusual cause of foot clonus: spasticity of fibularis longus muscle. Ann Phys Rehabil Med 2013;56:482–8. [19] Vinti M, Costantino F, Bayle N, Simpson DM, Weisz DJ, Gracies JM. Spastic cocontraction in hemiparesis: effects of botulinum toxin. Muscle Nerve 2012;46:926–31. [20] Vinti M, Couillandre A, Hausselle J, Bayle N, Primerano A, Merlo A, et al. Influence of effort intensity and gastrocnemius stretch on co-contraction and torque production in the healthy and paretic ankle. Clin Neurophysiol 2013;24:528–35. [21] Westberry DE, Davids JR, Roush TF, Pugh LI. Qualitative versus quantitative radiographic analysis of foot deformities in children with hemiplegic cerebral palsy. J Pediatr Orthop 2008;28:359–65. [22] Williams PF. Restoration of muscle balance of the foot by transfer of the tibialis posterior. J Bone Joint Surg Br 1976;58:217–9. [23] Wren TA, Rethlefsen S, Kay RM. Prevalence of specific gait abnormalities in children with cerebral palsy: influence of cerebral palsy subtype, age, and previous surgery. J Pediatr Orthop 2005;25:79–83.