- Email: [email protected]
Rectus femoris transfer and musculo-skeletal modeling: Effect of surgical treatment on gait and on rectus femoris kinematics
Gait & Posture 34 (2011) 519–523
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Rectus femoris transfer and musculo-skeletal modeling: Effect of surgical treatment on gait and on rectus femoris kinematics Eric Desailly a,*, Nejib Khouri b, Philippe Sardain c, Daniel Yepremian a, Patrick Lacouture c a
Fondation Ellen Poidatz, 77310 St Fargeau-Ponthierry, France Hopital Armand Trousseau, AP-HP, 75571 Paris Cedex 12, France c Institut PPrime UPR 3346 CNRS – Universite´ de Poitiers – ENSMA, De´partement ge´nie me´canique et syste`mes Complexes axe RoBioSS ‘‘Robotique Biome´canique Sport Sante´’’, SP2MI, BP 30179, 86962 Futuroscope Cedex, France b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 19 July 2010 Received in revised form 30 June 2011 Accepted 3 July 2011
Spasticity of the rectus femoris (RF) is one of the possible causes of stiff knee gait (SKG) in cerebral palsy. Musculoskeletal studies have shown that in SKG, length and speed of the RF are affected. No evaluation had been made to quantify the modifications of those parameters after surgery. The effect of this operation on gait quality and on RF kinematics was assessed in this study in order to identify kinematic patterns that may aid its diagnosis. For 26 transfers, clinical gait analysis pre- and post-surgery was used to compute the Gait Deviation Index (GDI) and Goldberg’s index. The kinematics of the Original RF path (ORFp) was studied before and after surgery. The expression ORFp was chosen to avoid any confusion between this modeling parameter, whose computation was unchanged, and the actual anatomical path that was modified by surgery. The gait quality was improved (+18 12GDI) and there was an inverse relation between the preoperative GDI and its improvement. The Golberg’s index was improved (88% of the cases). The operation had a significant effect on the normalization of the timings of maximum length and speed of the ORFp. The improvement of SKG was correlated with the normalization of the timing of the ORFp’s maximum length. The global improvement of the gait quality and of the SKG was demonstrated. Some parameters of muscular kinematics (RF length and velocity) have been standardized, showing an effect of the transfer not only during the swing, but also during stance. The premature timing of the ORFp peak length has been identified as a prognostic factor of a successful surgical outcome. ß 2011 Elsevier B.V. All rights reserved.
Keywords: Cerebral palsy Musculoskeletal modeling Rectus femoris Stiff knee Tendon transfer
1. Introduction In normal gait, the rectus femoris muscle (RF) is active during the transition between stance and swing phases to restrain knee flexion [1]. In children with cerebral palsy, over-activity of the RF may lead to a prolonged contraction during the swing phase. This abnormal activity is recognized as possibly restricting knee flexion during oscillation [1–5]. This gait abnormality, characterized by a defect of knee flexion during the swing phase, is called ‘‘stiff knee gait’’. Treatments of stiff knee gait include botulinic toxin injections, tendon transfer or distal release of the RF. The conventional treatment is to transfer the RF on to the tendon of the gracilis or semi-tendinosus muscle. The transferred muscle, although always spastic, remains a flexor of the hip, and is presumed to become a flexor rather than an extensor of the knee. This enables spasticity to functionally improve gait [1,2].
* Corresponding author. Tel.: +33 160658282; fax: +33 164381560. E-mail address: [email protected] (E. Desailly). 0966-6362/$ – see front matter ß 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2011.07.005
Spasticity of the RF is clinically assessed by the Duncan/Ely test, but this does not supply information about the functional consequences of the spasticity during walking. For this reason, surface electromyography (sEMG) is carried out to identify prolonged activity of the RF during the swing phase [1,3]. Nevertheless it has been shown that there is no correlation between a positive Duncan/Ely test and pathological sEMG activity. Also, the presence of abnormal sEMG activity of the RF does not influence the result of the surgical treatment [6,7]. In this context, clinical gait analysis and musculoskeletal modeling studies have been used to evaluate and to understand stiff knee gait, as well as the effects and results of the surgery [5,8]. 1.1. Causes of stiff knee Simulations have shown that the increase of activity of the RF leads to reduced knee flexion [9,10]. In addition, the improvement in knee movement after surgery has been shown to be associated with an increase in the knee flexion velocity at toe-off, as well as with a reduction in the knee extension moment during the doublesupport phase preceding the oscillation [11]. This confirms the
E. Desailly et al. / Gait & Posture 34 (2011) 519–523
520
important influence of pre-swing knee flexion velocity on the oscillation [8,12]. Others have shown that a decrease in the strength of the iliopsoas or gastrocnemius muscles, as well as an increase in the strength of the quadriceps muscles, might decrease the knee flexion range during the swing phase [13].
to study the overall effect of RF transfer; to study the effect of this surgery on the ORFp kinematics; to look for possible kinematic patterns of the ORFp, which would be predictive of a good surgical outcome, or, on the other hand, which could help to eliminate incorrect indications for surgery.
1.2. Effect of transfer
To answer these questions, a retrospective longitudinal study was carried out, whereby various parameters relating to the preand post-surgery conditions were compared.
The effect of the RF transfer in terms of the moment arm has been validated and the numerical model has allowed the amount of change of this moment arm to be predicted [14]. However, it has been found in vivo that the transferred RF generates a knee extension moment under electrical stimulation [15]. Others [16,17] have noted the presence of scar tissue situated between the RF and the adjacent structures. These observations led the authors to hypothesize that the transfer has no direct action on the flexion of the knee. Further, it seems to enable oscillation in much the same way as distal lengthening. However, it allows better hip flexion than would be achieved by simple distal elongation. It should however be noted that better results have been demonstrated with transfer, compared to distal release [3,6]. This is the case particularly when the range of motion of the knee is significantly limited [5,18]. 1.3. Rectus femoris kinematics
V RF ðtÞ
2.1. Population, materials and methods Sixteen children took part in this study (Electronic addenda). Children and their parents gave their assent for the use of their personal data at the end of the clinical research. All subjects were clinically examined before surgery and underwent gait analysis before and 1.8 0.9 years after surgery. The decision criteria leading up to surgery by transfer of the RF were based on clinical examination (Duncan/Ely test), on EMG (pathological activity during swing), and on kinematic examination focusing especially on the criterion of delayed and reduced peak knee flexion during swing. The gait analyses were undertaken using the modified Helen Hayes protocol, the medial femoral condyle and medial malleolus as anatomical landmarks and eight cameras (Vicon MX20). The EMG tests were undertaken using one of two devices (MotionLabs or Aurion Zero-Wire). All patients were asked to perform active knee flexion-extension in order to assess if the vasti and the RF were clearly differentiated. 2.2. Surgical protocol
In parallel with the above work, some studies have been carried out on the deviations of RF kinematics (length and velocity) in the case of stiff knee [19,20]. These showed reduced maximal length and a premature appearance of this during the swing phase, compared with the normal population. In addition, the elongation speed of the RF was significantly reduced and shifted from the double stance phase to that of the single stance. This would indicate that the RF does not reach maximal length at the time of maximal knee flexion: the first occurs earlier and the second is delayed in the case of stiff knee. This early timing of maximum RF lengthening velocity confirms the importance of stance in stiff knee. In such modeling studies, the purpose is to understand the human motion at the musculoskeletal level and the modeled parameters are only a representation of actual anatomy. Their modeling function depends on the origin A and termination (distal attachment) B of the muscle, the crossed joints angles qi(t), the modeling of the underlying bones Sj and on the muscle path identification algorithm: LRF ðtÞ ¼ f ðqi ðtÞ; S j Þ
2. Methods
dLRF ðtÞ dt
To date, the changes of the RF length and velocity after transfer have not been reported in the literature. It is therefore unknown if the RF kinematic parameters are modified by surgery and if they could help in decision-making relating to treatment. Studying these RF kinematic parameters should require taking into account the length LRF(t) and velocity VRF(t) of the Original RF path (from the Anterior Iliac Spine to the Tibial Tuberosity) before and after surgery. Thus, the studied postoperative parameter should not be the path of the transferred RF because this would lead to a comparison of variations resulting from two different modeling functions. The Original RF path (ORFp) has no postoperative anatomical existence, but its length and velocity parameters can be modified by the surgery. This should enable the modifications of these kinematic parameters to be studied. Therefore, if the disturbances described above are shown to contribute to the causes of the stiff knee gait, then the surgical treatment should improve those disorders. There are thus multiple objectives:
Twenty-six transfers of RFs on gracilis were undertaken. All patients in the group were operated on by the same surgeon with a technique comparable to the one described in [21]. Some subjects underwent additional procedures at the time of the RF transfer (26 intramuscular hamstring lengthening, 8 patella advancement, 10 femoral derotation, 11 adductor and/or psoas lengthening and 24 tibial osteotomy and/or foot deformity correction). 2.3. Computer data processing For this study all 16 medical files were examined and the length of the ORFp before and after surgery was simulated. The model used [22] is based on CT-Scan data provided by the European project VAKHUM (contract #IST-1999-10954subject 006). This data set is based on a single unimpaired adult. Geometric customization of the model was performed before and after surgery. Every skeletal segment was scaled to the subject by a homothetic transformation based on the ratio, subject/VAKHUM segment lengths. The morphological adaptation of the axial torsion of the femur and tibia was based on the clinical evaluation of the femoral anteversion or tibial torsion a [23–25]. As in previously published algorithms [26– 29] the adaptation was achieved by two successive transformations. First, each vertex of the considered bone was progressively rotated along its shaft by a value varying linearly from 0 to u = b a (b being the torsion of the generic bone). Second, the bone reference frame was inversely rotated by u in order to restore joint contact. Each time, after an iterative solidification of segment lengths, the position and the orientation of each bone was determined by the homogeneous matrix of the segments to which it belonged. Finally, a path was defined [30] from the anterior iliac spine to the tibial tuberosity passing by the patella. As explained in the introduction, this path corresponded to the Original RF path before surgery, although its kinematic parameters being studied before and after surgery. The length of the ORFp was normalized with regard to its length in the anatomical reference position. Lengthening speeds were calculated by finite time differentiation of the lengths. The same type of analysis was carried out on 14 healthy subjects (7 children and 7 adults), allowing a standard reference to be established for the kinematics of the ORFp. The ORFp was considered to be short if its maximum length was lower than the normal average maximum length minus two standard deviations. The timing of the maximum length peak was considered to be early if it occurred earlier than the normal average length peak minus two standard deviations. The ORFp lengthening rate was considered to be slow if its maximal lengthening speed was lower than the normal average maximal speed minus two standard deviations. Finally, the timing of the peak of maximal speed was considered to be early if it occurred earlier than the normal average peak speed minus two standard deviations. The times of the peaks of maximal length and maximal speed were measured from the beginning of the swing phase.
E. Desailly et al. / Gait & Posture 34 (2011) 519–523
521
In parallel with these measurements, an evaluation of the improvement of the overall quality of the gait was carried out by computing the Gait Deviation Index (GDI) [31] using the normal database previously described. Golberg’s index was calculated from the kinematic characteristics of the knee (peak knee flexion angle, range of knee flexion in early swing measured from toe-off to peak flexion, total range of knee motion and timing of peak knee flexion in swing) in order to determine if the 26 limbs were stiff, borderline or not stiff, using the criteria suggested by the author [11]. Relations of dependence between the pre- or post-surgery conditions and the ORFp variables (ORFp short or not; peak of maximal length premature or not; slow or not; peak of maximal speed premature or not) were tested using multiple contingency tables. The same was undertaken with Goldberg’s index of improvement and the normalization of the ORFp variables. A parameter was considered normalized when its value became included within 2 standard deviations of the control group distribution. The significance of these relations was tested with the Fischer’s exact test (p < 0.05).
3. Results The overall gait quality evaluated by the GDI improved for all except one of the subjects, with a significant improvement between the pre- and post-surgery conditions (Student’s T-test: F = 2.06; p < 0.05). The average GDI of the subjects pre-surgery was 58 12 (min = 40; max = 86) and 74 7 (min = 60; max = 86) postsurgery. The average improvement of the GDI was significant: 18 12 (min = 9; max = 39). An opposite interaction between the pre-surgery GDI and the improvement of this score was observed (Pearson’s coefficient = 0.81; p < 0.05) (Fig. 1). The Goldberg index identified 11 stiff and 15 borderline patients pre-operatively. Post-operatively, 21 were not stiff and 5 were borderline. Twenty-three lower limbs were improved in at least one domain (8 subjects improved in 2 domains). The effect of the surgery on the Goldberg index indicated an 88% improvement (Fisher’s exact test: p < 0.05). The length and velocity of the ORFp during gait, as well as the post-surgery normalization of its maximal length and velocity are illustrated in Figs. 2 and 3. The contingency table (Table 1) shows a significant effect of the surgery on the normalization of the timing of the peak of maximal length and also on the normalization of the timing of the peak for maximal velocity. The interaction between the improvement of stiff knee gait and the normalization of the various parameters was estimated. There appears to be a significant relationship between the normalization of the timing of the peak of maximal length and the improvement of the stiff knee gait (Table 2).
Fig. 1. Inverse correlation between the pre-surgery GDI and the improved GDI.
4. Discussion The aim of this study was to assess the effects of the RF transfer on gait quality and on ORFp kinematics in order to identify kinematic patterns that may assist in decision making. Nevertheless, the RF transfers were always undertaken in combination with other surgical procedures and were followed by rehabilitation. This study therefore evaluated the effect of the RF transfer in combination with other appropriate lower extremity surgical procedures. The improvement in gait kinematics was significant with an increase in GDI of 18 12. The average improvement was approximately two standard deviations from the mean of a healthy population. This improvement was all the more important given the patients pre-operative gait deviations. These results also indicate that improvements may be limited when the pre-operative GDI is higher than 75 (see linear regression in Fig. 1). All patients were surgically treated by a transfer of the RF after a standardized clinical gait analysis. In relation to this protocol, two different EMG devices were used during the study. This could have
Fig. 2. Example of one patient’s functional original rectus femoris path length before surgery (a) and after surgery (b). Patient data are in blue and are normalized over a gait cycle (one or more gait cycles can be overlaid). The black and shaded curve represents the mean 2S.D. of the same parameter for the control group. Modifications between (a) and (b) illustrate the normalization of the original rectus femoris path maximum length and maximum length timing after transfer. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of the article.)
E. Desailly et al. / Gait & Posture 34 (2011) 519–523
522
Fig. 3. Example of one patient’s functional original rectus femoris path velocity before surgery (a) and after surgery (b). Patient data are in blue and are normalized over a gait cycle (one or more gait cycles can be overlaid). The black and shaded curve represents the mean 2 S.D. of the same parameter for the control group. Modifications between (a) and (b) illustrate the normalization of the original rectus femoris path maximum velocity and maximum velocity timing after transfer. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of the article.)
led to different management decisions. However, the purpose here was not to evaluate the surgical decision but to study the changes in kinematic and musculoskeletal parameters after surgery. Furthermore, the decision to proceed with surgical treatment (or not) was made without knowledge of the stiff/borderline classification given retrospectively by the Goldberg index.
Table 1 Contingency tables assessing the independence between the surgery variable and the kinematic parameters for all the studied original rectus femoris paths. Each table totalizes the 26 transfers. Exact Fisher test
Surgery
Max length Short Not short Max length timing Early Not early Max velocity Slow Not slow Max velocity timing Early Not early
Before
After
17 9
14 12
17 9
6 20
9 17
4 22
23 3
14 12
p < 0.05
p < 0.05
Table 2 Contingency tables assessing the independence between the Goldberg index improvement and the normalization of kinematic parameters of the rectus femoris muscle. Each table only studies the outcome of the pre-operative original rectus femoris paths that presented a deficit of their kinematic parameter as identified in Table 1. Goldberg index improvement Yes Max length Normalized Not normalized Max length timing Normalized Not normalized Max velocity Normalized Not normalized Max velocity timing Normalized Not normalized
Exact Fisher test No
7 8
1 1
12 2
0 3
7 1
0 1
9 11
1 2
p < 0.05
Nevertheless, all of the operated patients were either stiff or borderline. The criteria for surgical decisions were based on clinical examination, EMG and evaluation of the kinematics for the subject investigated. These led to similar conclusions as provided by the retrospective calculation of the Goldberg index. This index, is helpful in decision making, but may not be essential. Nevertheless, it brings an additional element to the planning of surgical treatment. It achieves, along with conventional assessment, an overall evaluation of the sagittal kinematical parameters relating to disorders of the knee. This index was improved by surgery in 88% of the cases. The ORFp length and velocity were computed using a musculoskeletal model aiming to represent the part of the anatomy studied in motion. However, the customization was not entirely precise. As in most clinical musculoskeletal studies, the bone geometry came from a single unimpaired adult and was morphologically fitted to the patient’s bone geometry based on clinical measurements. This limitation should be taken into account when interpreting the results. The surgery significantly normalized the timing of the maximal length peak (65% of the patients were deficient before, against 23% after) and the timing of the maximal speed peak for the ORFp (88% were deficient before, against 54% after). The surgical treatment had a less significant effect on the normalization of the maximal length (65% were deficient before, against 54% after) and on the normalization of its maximal speed (35% were deficient before, against 15% after). The pre-surgery results nevertheless show that the stiff knee gait was correlated for 65% of the cases with a deficit in the timing of the peak of maximal length, for 88% of the cases with a deficit in the timing of the maximal speed peak, and for 65% of the cases with a deficit in the maximal length. Finally, only 35% of the stiff knee gait cases were correlated with a deficit in the maximal speed of elongation. These data taken separately do not allow the definition of a kinematic criterion for the ORFp that would be pathognomonic of a stiff knee gait. Two of these parameters were nevertheless significantly improved by the surgery. In addition, with respect to these parameters, the normalization of the timing of maximal length is correlated to the improvement of stiff knee gait. One can consequently conclude that the presence of premature timing of maximum length of the ORFp is a good indicator of possible improvement of the patient’s stiff knee gait in the event of transfer of the RF.
E. Desailly et al. / Gait & Posture 34 (2011) 519–523
The reported changes in post-operative length and velocity do not represent those of the actual transferred muscle, and therefore cannot be correlated with any changes in the spastic response of this muscle post-operatively. These results, relating as they do to predictors of success of the surgical treatment, should be balanced by the low number of cases in the studied population showing no improvement in stiff knee gait. Nevertheless, this is an additional parameter that could be useful. Previous works relating to outcome prediction revealed that neither the EMG nor the Duncan/Ely test were sufficient [6,32–34]. More recent work based on data-mining techniques combined multiple kinematic and kinetic parameters to predict the transfer outcome [35]. This paper concluded that further studies are required to determine the ultimate set of predictors. The authors believe that ORFp kinematics might be tested as one of these possible predictors. Different indicators were used to evaluate the effect of the rectus femoris transfer in stiff knee gait. Improvement in general gait quality was demonstrated, as well as an improvement in knee stiffness. Some parameters of the RF length and velocity were normalized, showing the effect of the surgery, not only during swing but also during stance. The improvement in stiff knee gait was correlated in particular with the normalization of the timing of the peak of the ORFp’s length. The findings of this study contribute both to the understanding of stiff knee gait and to the prediction of outcome of RF transfer.
[7]
[8] [9]
[10] [11]
[12]
[13]
[14] [15]
[16]
[17]
[18]
Acknowledgments [19]
The authors express their gratitude to the subjects who participated and to all the members of the ‘‘Unite´ d’Analyse du Mouvement’’ (Motion Analysis Unit) at the Fondation Ellen Poidatz, as well as to the ‘‘RoBioSS – Robotique, Biome´canique, Sport et Sante´’’ team at the University of Poitiers. Funding for this work was provided by the University of Poitiers, the Centre National de la Recherche Scientifique (CNRS), the Fondation Ellen Poidatz and the ‘‘Socie´te´ d’Etudes et de Soins pour les Enfants Paralyse´s et polymalforme´s’’.
[20] [21] [22]
[23]
[24]
Conflict of interest
[25]
None declared. [26]
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.gaitpost.2011.07.005.
[27]
[28]
[29]
References [1] Perry J. Gait analysis: normal and pathological function. SLACK incorporated; 1992. [2] Gage JR, Perry J, Hicks RR, Koop S, Werntz JR. Rectus femoris transfer to improve knee function of children with cerebral palsy. Developmental Medicine & Child Neurology 1987;29(2):159–66. [3] Sutherland DH, Santi M, Abel MF. Treatment of stiff-knee gait in cerebral palsy: a comparison by gait analysis of distal rectus femoris transfer versus proximal rectus release. Journal of Pediatric Orthopedics 1990;10(4):433–41. [4] Ounpuu S, Muik E, Davis III RB, Gage JR, DeLuca PA. Rectus femoris surgery in children with cerebral palsy. Part I: The effect of rectus femoris transfer location on knee motion. Journal of Pediatric Orthopedics 1993;13(3):325–30. [5] Ounpuu S, Muik E, Davis III RB, Gage JR, DeLuca PA. Rectus femoris surgery in children with cerebral palsy. Part II: A comparison between the effect of transfer and release of the distal rectus femorison knee motion. Journal of Pediatric Orthopedics 1993;13(3):331–5. [6] Chambers H, Lauer A, Kaufman K, Cardelia JM, Sutherland D. Prediction of outcome after rectus femoris surgery in cerebral palsy: the role of cocontraction
[30]
[31] [32]
[33]
[34]
[35]
523
of the rectus femoris and vastus lateralis. Journal of Pediatric Orthopedics 1998;18(6):703–11. Ounpuu S, DeLuca PA, Bell KJ, Davis RB. Using surface electrodes for the evaluation of the rectus femoris, vastus medialis and vastus lateralis muscles in children with cerebral palsy. Gait and Posture 1997;5(3):211–6. Goldberg SR, Ounpuu S, Delp SL. The importance of swing-phase initial conditions in stiff-knee gait. Journal of Biomechanics 2003;36(8):1111–6. Jonkers I, Stewart S, Spaepen A. The study of muscle action during single support and swing phase of gait: clinical relevance of forward simulation techniques. Gait and Posture 2003;17:97–105. Piazza SJ, Delp SL. The influence of muscles on knee flexion during the swing phase of gait. Journal of Biomechanics 1996;29(6):723–33. Goldberg SR, Ounpuu S, Arnold AS, Gage JR, Delp SL. Kinematic and kinetic factors that correlate with improved knee flexion following treatment for stiffknee gait. Journal of Biomechanics 2006;39(4):689–98. Anderson FC, Goldberg SR, Pandy MG, Delp SL. Contributions of muscle forces and toe-off kinematics to peak knee flexion during the swing phase of normal gait: an induced position analysis. Journal of Biomechanics 2004;37(5): 731–7. Goldberg SR, Anderson FC, Pandy MG, Delp SL. Muscles that influence knee flexion velocity in double support: implications for stiff-knee gait. Journal of Biomechanics 2004;37(8):1189–96. Delp SL. Transfer of the rectus femoris: effects of transfer site on moment arms about the knee and hip. Journal of Biomechanics 1994;27(10):1201–11. Riewald SA, Delp SL. The action of the rectus femoris muscle following distal tendon transfer: does it generate knee flexion moment? Developmental Medicine & Child Neurology 1997;39(2):99–105. Asakawa DS, Blemker SS, Rab GT, Bagley A, Delp SL. Three-dimensional muscle-tendon geometry after rectus femoris tendon transfer. Journal of Bone and Joint Surgery Series A 2004;86(2):348–54. Asakawa DS, Blemker SS, Gold GE, Delp SL. In vivo motion of the rectus femoris muscle after tendon transfer surgery. Journal of Biomechanics 2002;35(8): 1029–37. Presedo A, Mehrafshan M, Laassel E, Topouchian V, Mohammad Y, Gouraud D, et al. Distal rectus femoris release versus transfer to treat gait abnormalities of the knee in children with cerebral palsy. Gait and Posture 2005;21(Supplement 1):S87. Jonkers I, Stewart C, Desloovere K, Molenaers G, Spaepen A. Musculo-tendon length and lengthening velocity of rectus femoris in stiff knee gait. Gait and Posture 2006;23(2):222–9. Stewart C, Farmer S, Forward M, Patrick J. Patterns of behavior of the rectus femoris muscle in cerebral palsy diplegia. Gait and Posture 1999;10:60. Gage JR, Schwartz MH, Koop SE, Novacheck TF. The identification and treatment of gait problems in cerebral palsy. John Wiley and Sons; 2009. Desailly E. Analyse biome´canique 3D de la marche de l’enfant de´ficient moteur – Mode´lisation segmentaire et mode´lisation musculo-squelettique. Ph.D. Thesis; 2008. Staheli LT, Corbett M, Wyss C, King H. Lower-extremity rotational problems in children. Normal values to guide management. The Journal of Bone and Joint Surgery A 1985;67:39–47. Davids JR, Benfanti P, Blackhurst DW, Allen BL. Assessment of femoral anteversion in children with cerebral palsy: accuracy of the trochanteric prominence angle test. Journal of Pediatric Orthopaedics 2002;22:173–8. Jacquemier M, Glard Y, Pomero V, Viehweger E, Jouve JL, Bollini G. Rotational profile of the lower limb in 1319 healthy children. Gait and Posture 2008;28:187–93. Delp SL, Loan JP, Hoy MG, Zajac FE, Topp EL, Rosen JM. An interactive graphicsbased model of the lower extremity to study orthopaedic surgical procedures. IEEE Transactions on Biomedical Engineering 1990;37:757–67. Schutte LM, Hayden SW, Gage JR. Lengths of hamstrings and psoas muscles during crouch gait: effects of femoral anteversion. Journal of Orthopaedic Research 1997;15:615–21. Arnold AS, Salinas S, Asakawa DJ, Delp SL. Accuracy of muscle moment arms estimated from MRI-based musculoskeletal models of the lower extremity. Computer Aided Surgery 2000;5:108–19. Arnold AS, Delp SL. Rotational moment arms of the medial hamstrings and adductors vary with femoral geometry and limb position: implications for the treatment of internally rotated gait. Journal of Biomechanics 2001;34: 437–47. Desailly E, Sardain P, Khouri N, Yepremian D, Lacouture P. The convex wrapping algorithm: a method for identifying muscle paths using the underlying bone mesh. Journal of Biomechanics 2010;43(13):2601–7. Schwartz MH, Rozumalski A. The gait deviation index: a new comprehensive index of gait pathology. Journal of Pediatric Orthopedics 2008;28:351–7. Kay RM, Rethlefsen SA, Kelly JP, Wren TAL. Predictive value of the Duncan-Ely test in distal rectus femoris transfer. Journal of Pediatric Orthopedics 2004;24(1):59–62. Miller F, Dias RC, Lipton GE, Albarracin JP, Dabney KW, Castagno P. The effect of rectus EMG patterns on the outcome of rectus femoris transfers. Journal of Pediatric Orthopedics 1997;17(5):603. Muthusamy K, Seidl AJ, Friesen RM, Carollo JJ, Pan Z, Chang FM. Rectus femoris transfer in children with cerebral palsy: evaluation of transfer site and preoperative indicators. Journal of Pediatric Orthopedics 2008;28(6): 674–8. Reinbolt JA, Fox MD, Schwartz MH, Delp SL. Predicting outcomes of rectus femoris transfer surgery. Gait and Posture 2009;30(1):100–5.
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Rectus femoris transfer and musculo-skeletal modeling: Effect of surgical treatment on gait and on rectus femoris kinematics Eric Desailly a,*, Nejib Khouri b, Philippe Sardain c, Daniel Yepremian a, Patrick Lacouture c a
Fondation Ellen Poidatz, 77310 St Fargeau-Ponthierry, France Hopital Armand Trousseau, AP-HP, 75571 Paris Cedex 12, France c Institut PPrime UPR 3346 CNRS – Universite´ de Poitiers – ENSMA, De´partement ge´nie me´canique et syste`mes Complexes axe RoBioSS ‘‘Robotique Biome´canique Sport Sante´’’, SP2MI, BP 30179, 86962 Futuroscope Cedex, France b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 19 July 2010 Received in revised form 30 June 2011 Accepted 3 July 2011
Spasticity of the rectus femoris (RF) is one of the possible causes of stiff knee gait (SKG) in cerebral palsy. Musculoskeletal studies have shown that in SKG, length and speed of the RF are affected. No evaluation had been made to quantify the modifications of those parameters after surgery. The effect of this operation on gait quality and on RF kinematics was assessed in this study in order to identify kinematic patterns that may aid its diagnosis. For 26 transfers, clinical gait analysis pre- and post-surgery was used to compute the Gait Deviation Index (GDI) and Goldberg’s index. The kinematics of the Original RF path (ORFp) was studied before and after surgery. The expression ORFp was chosen to avoid any confusion between this modeling parameter, whose computation was unchanged, and the actual anatomical path that was modified by surgery. The gait quality was improved (+18 12GDI) and there was an inverse relation between the preoperative GDI and its improvement. The Golberg’s index was improved (88% of the cases). The operation had a significant effect on the normalization of the timings of maximum length and speed of the ORFp. The improvement of SKG was correlated with the normalization of the timing of the ORFp’s maximum length. The global improvement of the gait quality and of the SKG was demonstrated. Some parameters of muscular kinematics (RF length and velocity) have been standardized, showing an effect of the transfer not only during the swing, but also during stance. The premature timing of the ORFp peak length has been identified as a prognostic factor of a successful surgical outcome. ß 2011 Elsevier B.V. All rights reserved.
Keywords: Cerebral palsy Musculoskeletal modeling Rectus femoris Stiff knee Tendon transfer
1. Introduction In normal gait, the rectus femoris muscle (RF) is active during the transition between stance and swing phases to restrain knee flexion [1]. In children with cerebral palsy, over-activity of the RF may lead to a prolonged contraction during the swing phase. This abnormal activity is recognized as possibly restricting knee flexion during oscillation [1–5]. This gait abnormality, characterized by a defect of knee flexion during the swing phase, is called ‘‘stiff knee gait’’. Treatments of stiff knee gait include botulinic toxin injections, tendon transfer or distal release of the RF. The conventional treatment is to transfer the RF on to the tendon of the gracilis or semi-tendinosus muscle. The transferred muscle, although always spastic, remains a flexor of the hip, and is presumed to become a flexor rather than an extensor of the knee. This enables spasticity to functionally improve gait [1,2].
* Corresponding author. Tel.: +33 160658282; fax: +33 164381560. E-mail address: [email protected] (E. Desailly). 0966-6362/$ – see front matter ß 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2011.07.005
Spasticity of the RF is clinically assessed by the Duncan/Ely test, but this does not supply information about the functional consequences of the spasticity during walking. For this reason, surface electromyography (sEMG) is carried out to identify prolonged activity of the RF during the swing phase [1,3]. Nevertheless it has been shown that there is no correlation between a positive Duncan/Ely test and pathological sEMG activity. Also, the presence of abnormal sEMG activity of the RF does not influence the result of the surgical treatment [6,7]. In this context, clinical gait analysis and musculoskeletal modeling studies have been used to evaluate and to understand stiff knee gait, as well as the effects and results of the surgery [5,8]. 1.1. Causes of stiff knee Simulations have shown that the increase of activity of the RF leads to reduced knee flexion [9,10]. In addition, the improvement in knee movement after surgery has been shown to be associated with an increase in the knee flexion velocity at toe-off, as well as with a reduction in the knee extension moment during the doublesupport phase preceding the oscillation [11]. This confirms the
E. Desailly et al. / Gait & Posture 34 (2011) 519–523
520
important influence of pre-swing knee flexion velocity on the oscillation [8,12]. Others have shown that a decrease in the strength of the iliopsoas or gastrocnemius muscles, as well as an increase in the strength of the quadriceps muscles, might decrease the knee flexion range during the swing phase [13].
to study the overall effect of RF transfer; to study the effect of this surgery on the ORFp kinematics; to look for possible kinematic patterns of the ORFp, which would be predictive of a good surgical outcome, or, on the other hand, which could help to eliminate incorrect indications for surgery.
1.2. Effect of transfer
To answer these questions, a retrospective longitudinal study was carried out, whereby various parameters relating to the preand post-surgery conditions were compared.
The effect of the RF transfer in terms of the moment arm has been validated and the numerical model has allowed the amount of change of this moment arm to be predicted [14]. However, it has been found in vivo that the transferred RF generates a knee extension moment under electrical stimulation [15]. Others [16,17] have noted the presence of scar tissue situated between the RF and the adjacent structures. These observations led the authors to hypothesize that the transfer has no direct action on the flexion of the knee. Further, it seems to enable oscillation in much the same way as distal lengthening. However, it allows better hip flexion than would be achieved by simple distal elongation. It should however be noted that better results have been demonstrated with transfer, compared to distal release [3,6]. This is the case particularly when the range of motion of the knee is significantly limited [5,18]. 1.3. Rectus femoris kinematics
V RF ðtÞ
2.1. Population, materials and methods Sixteen children took part in this study (Electronic addenda). Children and their parents gave their assent for the use of their personal data at the end of the clinical research. All subjects were clinically examined before surgery and underwent gait analysis before and 1.8 0.9 years after surgery. The decision criteria leading up to surgery by transfer of the RF were based on clinical examination (Duncan/Ely test), on EMG (pathological activity during swing), and on kinematic examination focusing especially on the criterion of delayed and reduced peak knee flexion during swing. The gait analyses were undertaken using the modified Helen Hayes protocol, the medial femoral condyle and medial malleolus as anatomical landmarks and eight cameras (Vicon MX20). The EMG tests were undertaken using one of two devices (MotionLabs or Aurion Zero-Wire). All patients were asked to perform active knee flexion-extension in order to assess if the vasti and the RF were clearly differentiated. 2.2. Surgical protocol
In parallel with the above work, some studies have been carried out on the deviations of RF kinematics (length and velocity) in the case of stiff knee [19,20]. These showed reduced maximal length and a premature appearance of this during the swing phase, compared with the normal population. In addition, the elongation speed of the RF was significantly reduced and shifted from the double stance phase to that of the single stance. This would indicate that the RF does not reach maximal length at the time of maximal knee flexion: the first occurs earlier and the second is delayed in the case of stiff knee. This early timing of maximum RF lengthening velocity confirms the importance of stance in stiff knee. In such modeling studies, the purpose is to understand the human motion at the musculoskeletal level and the modeled parameters are only a representation of actual anatomy. Their modeling function depends on the origin A and termination (distal attachment) B of the muscle, the crossed joints angles qi(t), the modeling of the underlying bones Sj and on the muscle path identification algorithm: LRF ðtÞ ¼ f ðqi ðtÞ; S j Þ
2. Methods
dLRF ðtÞ dt
To date, the changes of the RF length and velocity after transfer have not been reported in the literature. It is therefore unknown if the RF kinematic parameters are modified by surgery and if they could help in decision-making relating to treatment. Studying these RF kinematic parameters should require taking into account the length LRF(t) and velocity VRF(t) of the Original RF path (from the Anterior Iliac Spine to the Tibial Tuberosity) before and after surgery. Thus, the studied postoperative parameter should not be the path of the transferred RF because this would lead to a comparison of variations resulting from two different modeling functions. The Original RF path (ORFp) has no postoperative anatomical existence, but its length and velocity parameters can be modified by the surgery. This should enable the modifications of these kinematic parameters to be studied. Therefore, if the disturbances described above are shown to contribute to the causes of the stiff knee gait, then the surgical treatment should improve those disorders. There are thus multiple objectives:
Twenty-six transfers of RFs on gracilis were undertaken. All patients in the group were operated on by the same surgeon with a technique comparable to the one described in [21]. Some subjects underwent additional procedures at the time of the RF transfer (26 intramuscular hamstring lengthening, 8 patella advancement, 10 femoral derotation, 11 adductor and/or psoas lengthening and 24 tibial osteotomy and/or foot deformity correction). 2.3. Computer data processing For this study all 16 medical files were examined and the length of the ORFp before and after surgery was simulated. The model used [22] is based on CT-Scan data provided by the European project VAKHUM (contract #IST-1999-10954subject 006). This data set is based on a single unimpaired adult. Geometric customization of the model was performed before and after surgery. Every skeletal segment was scaled to the subject by a homothetic transformation based on the ratio, subject/VAKHUM segment lengths. The morphological adaptation of the axial torsion of the femur and tibia was based on the clinical evaluation of the femoral anteversion or tibial torsion a [23–25]. As in previously published algorithms [26– 29] the adaptation was achieved by two successive transformations. First, each vertex of the considered bone was progressively rotated along its shaft by a value varying linearly from 0 to u = b a (b being the torsion of the generic bone). Second, the bone reference frame was inversely rotated by u in order to restore joint contact. Each time, after an iterative solidification of segment lengths, the position and the orientation of each bone was determined by the homogeneous matrix of the segments to which it belonged. Finally, a path was defined [30] from the anterior iliac spine to the tibial tuberosity passing by the patella. As explained in the introduction, this path corresponded to the Original RF path before surgery, although its kinematic parameters being studied before and after surgery. The length of the ORFp was normalized with regard to its length in the anatomical reference position. Lengthening speeds were calculated by finite time differentiation of the lengths. The same type of analysis was carried out on 14 healthy subjects (7 children and 7 adults), allowing a standard reference to be established for the kinematics of the ORFp. The ORFp was considered to be short if its maximum length was lower than the normal average maximum length minus two standard deviations. The timing of the maximum length peak was considered to be early if it occurred earlier than the normal average length peak minus two standard deviations. The ORFp lengthening rate was considered to be slow if its maximal lengthening speed was lower than the normal average maximal speed minus two standard deviations. Finally, the timing of the peak of maximal speed was considered to be early if it occurred earlier than the normal average peak speed minus two standard deviations. The times of the peaks of maximal length and maximal speed were measured from the beginning of the swing phase.
E. Desailly et al. / Gait & Posture 34 (2011) 519–523
521
In parallel with these measurements, an evaluation of the improvement of the overall quality of the gait was carried out by computing the Gait Deviation Index (GDI) [31] using the normal database previously described. Golberg’s index was calculated from the kinematic characteristics of the knee (peak knee flexion angle, range of knee flexion in early swing measured from toe-off to peak flexion, total range of knee motion and timing of peak knee flexion in swing) in order to determine if the 26 limbs were stiff, borderline or not stiff, using the criteria suggested by the author [11]. Relations of dependence between the pre- or post-surgery conditions and the ORFp variables (ORFp short or not; peak of maximal length premature or not; slow or not; peak of maximal speed premature or not) were tested using multiple contingency tables. The same was undertaken with Goldberg’s index of improvement and the normalization of the ORFp variables. A parameter was considered normalized when its value became included within 2 standard deviations of the control group distribution. The significance of these relations was tested with the Fischer’s exact test (p < 0.05).
3. Results The overall gait quality evaluated by the GDI improved for all except one of the subjects, with a significant improvement between the pre- and post-surgery conditions (Student’s T-test: F = 2.06; p < 0.05). The average GDI of the subjects pre-surgery was 58 12 (min = 40; max = 86) and 74 7 (min = 60; max = 86) postsurgery. The average improvement of the GDI was significant: 18 12 (min = 9; max = 39). An opposite interaction between the pre-surgery GDI and the improvement of this score was observed (Pearson’s coefficient = 0.81; p < 0.05) (Fig. 1). The Goldberg index identified 11 stiff and 15 borderline patients pre-operatively. Post-operatively, 21 were not stiff and 5 were borderline. Twenty-three lower limbs were improved in at least one domain (8 subjects improved in 2 domains). The effect of the surgery on the Goldberg index indicated an 88% improvement (Fisher’s exact test: p < 0.05). The length and velocity of the ORFp during gait, as well as the post-surgery normalization of its maximal length and velocity are illustrated in Figs. 2 and 3. The contingency table (Table 1) shows a significant effect of the surgery on the normalization of the timing of the peak of maximal length and also on the normalization of the timing of the peak for maximal velocity. The interaction between the improvement of stiff knee gait and the normalization of the various parameters was estimated. There appears to be a significant relationship between the normalization of the timing of the peak of maximal length and the improvement of the stiff knee gait (Table 2).
Fig. 1. Inverse correlation between the pre-surgery GDI and the improved GDI.
4. Discussion The aim of this study was to assess the effects of the RF transfer on gait quality and on ORFp kinematics in order to identify kinematic patterns that may assist in decision making. Nevertheless, the RF transfers were always undertaken in combination with other surgical procedures and were followed by rehabilitation. This study therefore evaluated the effect of the RF transfer in combination with other appropriate lower extremity surgical procedures. The improvement in gait kinematics was significant with an increase in GDI of 18 12. The average improvement was approximately two standard deviations from the mean of a healthy population. This improvement was all the more important given the patients pre-operative gait deviations. These results also indicate that improvements may be limited when the pre-operative GDI is higher than 75 (see linear regression in Fig. 1). All patients were surgically treated by a transfer of the RF after a standardized clinical gait analysis. In relation to this protocol, two different EMG devices were used during the study. This could have
Fig. 2. Example of one patient’s functional original rectus femoris path length before surgery (a) and after surgery (b). Patient data are in blue and are normalized over a gait cycle (one or more gait cycles can be overlaid). The black and shaded curve represents the mean 2S.D. of the same parameter for the control group. Modifications between (a) and (b) illustrate the normalization of the original rectus femoris path maximum length and maximum length timing after transfer. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of the article.)
E. Desailly et al. / Gait & Posture 34 (2011) 519–523
522
Fig. 3. Example of one patient’s functional original rectus femoris path velocity before surgery (a) and after surgery (b). Patient data are in blue and are normalized over a gait cycle (one or more gait cycles can be overlaid). The black and shaded curve represents the mean 2 S.D. of the same parameter for the control group. Modifications between (a) and (b) illustrate the normalization of the original rectus femoris path maximum velocity and maximum velocity timing after transfer. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of the article.)
led to different management decisions. However, the purpose here was not to evaluate the surgical decision but to study the changes in kinematic and musculoskeletal parameters after surgery. Furthermore, the decision to proceed with surgical treatment (or not) was made without knowledge of the stiff/borderline classification given retrospectively by the Goldberg index.
Table 1 Contingency tables assessing the independence between the surgery variable and the kinematic parameters for all the studied original rectus femoris paths. Each table totalizes the 26 transfers. Exact Fisher test
Surgery
Max length Short Not short Max length timing Early Not early Max velocity Slow Not slow Max velocity timing Early Not early
Before
After
17 9
14 12
17 9
6 20
9 17
4 22
23 3
14 12
p < 0.05
p < 0.05
Table 2 Contingency tables assessing the independence between the Goldberg index improvement and the normalization of kinematic parameters of the rectus femoris muscle. Each table only studies the outcome of the pre-operative original rectus femoris paths that presented a deficit of their kinematic parameter as identified in Table 1. Goldberg index improvement Yes Max length Normalized Not normalized Max length timing Normalized Not normalized Max velocity Normalized Not normalized Max velocity timing Normalized Not normalized
Exact Fisher test No
7 8
1 1
12 2
0 3
7 1
0 1
9 11
1 2
p < 0.05
Nevertheless, all of the operated patients were either stiff or borderline. The criteria for surgical decisions were based on clinical examination, EMG and evaluation of the kinematics for the subject investigated. These led to similar conclusions as provided by the retrospective calculation of the Goldberg index. This index, is helpful in decision making, but may not be essential. Nevertheless, it brings an additional element to the planning of surgical treatment. It achieves, along with conventional assessment, an overall evaluation of the sagittal kinematical parameters relating to disorders of the knee. This index was improved by surgery in 88% of the cases. The ORFp length and velocity were computed using a musculoskeletal model aiming to represent the part of the anatomy studied in motion. However, the customization was not entirely precise. As in most clinical musculoskeletal studies, the bone geometry came from a single unimpaired adult and was morphologically fitted to the patient’s bone geometry based on clinical measurements. This limitation should be taken into account when interpreting the results. The surgery significantly normalized the timing of the maximal length peak (65% of the patients were deficient before, against 23% after) and the timing of the maximal speed peak for the ORFp (88% were deficient before, against 54% after). The surgical treatment had a less significant effect on the normalization of the maximal length (65% were deficient before, against 54% after) and on the normalization of its maximal speed (35% were deficient before, against 15% after). The pre-surgery results nevertheless show that the stiff knee gait was correlated for 65% of the cases with a deficit in the timing of the peak of maximal length, for 88% of the cases with a deficit in the timing of the maximal speed peak, and for 65% of the cases with a deficit in the maximal length. Finally, only 35% of the stiff knee gait cases were correlated with a deficit in the maximal speed of elongation. These data taken separately do not allow the definition of a kinematic criterion for the ORFp that would be pathognomonic of a stiff knee gait. Two of these parameters were nevertheless significantly improved by the surgery. In addition, with respect to these parameters, the normalization of the timing of maximal length is correlated to the improvement of stiff knee gait. One can consequently conclude that the presence of premature timing of maximum length of the ORFp is a good indicator of possible improvement of the patient’s stiff knee gait in the event of transfer of the RF.
E. Desailly et al. / Gait & Posture 34 (2011) 519–523
The reported changes in post-operative length and velocity do not represent those of the actual transferred muscle, and therefore cannot be correlated with any changes in the spastic response of this muscle post-operatively. These results, relating as they do to predictors of success of the surgical treatment, should be balanced by the low number of cases in the studied population showing no improvement in stiff knee gait. Nevertheless, this is an additional parameter that could be useful. Previous works relating to outcome prediction revealed that neither the EMG nor the Duncan/Ely test were sufficient [6,32–34]. More recent work based on data-mining techniques combined multiple kinematic and kinetic parameters to predict the transfer outcome [35]. This paper concluded that further studies are required to determine the ultimate set of predictors. The authors believe that ORFp kinematics might be tested as one of these possible predictors. Different indicators were used to evaluate the effect of the rectus femoris transfer in stiff knee gait. Improvement in general gait quality was demonstrated, as well as an improvement in knee stiffness. Some parameters of the RF length and velocity were normalized, showing the effect of the surgery, not only during swing but also during stance. The improvement in stiff knee gait was correlated in particular with the normalization of the timing of the peak of the ORFp’s length. The findings of this study contribute both to the understanding of stiff knee gait and to the prediction of outcome of RF transfer.
[7]
[8] [9]
[10] [11]
[12]
[13]
[14] [15]
[16]
[17]
[18]
Acknowledgments [19]
The authors express their gratitude to the subjects who participated and to all the members of the ‘‘Unite´ d’Analyse du Mouvement’’ (Motion Analysis Unit) at the Fondation Ellen Poidatz, as well as to the ‘‘RoBioSS – Robotique, Biome´canique, Sport et Sante´’’ team at the University of Poitiers. Funding for this work was provided by the University of Poitiers, the Centre National de la Recherche Scientifique (CNRS), the Fondation Ellen Poidatz and the ‘‘Socie´te´ d’Etudes et de Soins pour les Enfants Paralyse´s et polymalforme´s’’.
[20] [21] [22]
[23]
[24]
Conflict of interest
[25]
None declared. [26]
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.gaitpost.2011.07.005.
[27]
[28]
[29]
References [1] Perry J. Gait analysis: normal and pathological function. SLACK incorporated; 1992. [2] Gage JR, Perry J, Hicks RR, Koop S, Werntz JR. Rectus femoris transfer to improve knee function of children with cerebral palsy. Developmental Medicine & Child Neurology 1987;29(2):159–66. [3] Sutherland DH, Santi M, Abel MF. Treatment of stiff-knee gait in cerebral palsy: a comparison by gait analysis of distal rectus femoris transfer versus proximal rectus release. Journal of Pediatric Orthopedics 1990;10(4):433–41. [4] Ounpuu S, Muik E, Davis III RB, Gage JR, DeLuca PA. Rectus femoris surgery in children with cerebral palsy. Part I: The effect of rectus femoris transfer location on knee motion. Journal of Pediatric Orthopedics 1993;13(3):325–30. [5] Ounpuu S, Muik E, Davis III RB, Gage JR, DeLuca PA. Rectus femoris surgery in children with cerebral palsy. Part II: A comparison between the effect of transfer and release of the distal rectus femorison knee motion. Journal of Pediatric Orthopedics 1993;13(3):331–5. [6] Chambers H, Lauer A, Kaufman K, Cardelia JM, Sutherland D. Prediction of outcome after rectus femoris surgery in cerebral palsy: the role of cocontraction
[30]
[31] [32]
[33]
[34]
[35]
523
of the rectus femoris and vastus lateralis. Journal of Pediatric Orthopedics 1998;18(6):703–11. Ounpuu S, DeLuca PA, Bell KJ, Davis RB. Using surface electrodes for the evaluation of the rectus femoris, vastus medialis and vastus lateralis muscles in children with cerebral palsy. Gait and Posture 1997;5(3):211–6. Goldberg SR, Ounpuu S, Delp SL. The importance of swing-phase initial conditions in stiff-knee gait. Journal of Biomechanics 2003;36(8):1111–6. Jonkers I, Stewart S, Spaepen A. The study of muscle action during single support and swing phase of gait: clinical relevance of forward simulation techniques. Gait and Posture 2003;17:97–105. Piazza SJ, Delp SL. The influence of muscles on knee flexion during the swing phase of gait. Journal of Biomechanics 1996;29(6):723–33. Goldberg SR, Ounpuu S, Arnold AS, Gage JR, Delp SL. Kinematic and kinetic factors that correlate with improved knee flexion following treatment for stiffknee gait. Journal of Biomechanics 2006;39(4):689–98. Anderson FC, Goldberg SR, Pandy MG, Delp SL. Contributions of muscle forces and toe-off kinematics to peak knee flexion during the swing phase of normal gait: an induced position analysis. Journal of Biomechanics 2004;37(5): 731–7. Goldberg SR, Anderson FC, Pandy MG, Delp SL. Muscles that influence knee flexion velocity in double support: implications for stiff-knee gait. Journal of Biomechanics 2004;37(8):1189–96. Delp SL. Transfer of the rectus femoris: effects of transfer site on moment arms about the knee and hip. Journal of Biomechanics 1994;27(10):1201–11. Riewald SA, Delp SL. The action of the rectus femoris muscle following distal tendon transfer: does it generate knee flexion moment? Developmental Medicine & Child Neurology 1997;39(2):99–105. Asakawa DS, Blemker SS, Rab GT, Bagley A, Delp SL. Three-dimensional muscle-tendon geometry after rectus femoris tendon transfer. Journal of Bone and Joint Surgery Series A 2004;86(2):348–54. Asakawa DS, Blemker SS, Gold GE, Delp SL. In vivo motion of the rectus femoris muscle after tendon transfer surgery. Journal of Biomechanics 2002;35(8): 1029–37. Presedo A, Mehrafshan M, Laassel E, Topouchian V, Mohammad Y, Gouraud D, et al. Distal rectus femoris release versus transfer to treat gait abnormalities of the knee in children with cerebral palsy. Gait and Posture 2005;21(Supplement 1):S87. Jonkers I, Stewart C, Desloovere K, Molenaers G, Spaepen A. Musculo-tendon length and lengthening velocity of rectus femoris in stiff knee gait. Gait and Posture 2006;23(2):222–9. Stewart C, Farmer S, Forward M, Patrick J. Patterns of behavior of the rectus femoris muscle in cerebral palsy diplegia. Gait and Posture 1999;10:60. Gage JR, Schwartz MH, Koop SE, Novacheck TF. The identification and treatment of gait problems in cerebral palsy. John Wiley and Sons; 2009. Desailly E. Analyse biome´canique 3D de la marche de l’enfant de´ficient moteur – Mode´lisation segmentaire et mode´lisation musculo-squelettique. Ph.D. Thesis; 2008. Staheli LT, Corbett M, Wyss C, King H. Lower-extremity rotational problems in children. Normal values to guide management. The Journal of Bone and Joint Surgery A 1985;67:39–47. Davids JR, Benfanti P, Blackhurst DW, Allen BL. Assessment of femoral anteversion in children with cerebral palsy: accuracy of the trochanteric prominence angle test. Journal of Pediatric Orthopaedics 2002;22:173–8. Jacquemier M, Glard Y, Pomero V, Viehweger E, Jouve JL, Bollini G. Rotational profile of the lower limb in 1319 healthy children. Gait and Posture 2008;28:187–93. Delp SL, Loan JP, Hoy MG, Zajac FE, Topp EL, Rosen JM. An interactive graphicsbased model of the lower extremity to study orthopaedic surgical procedures. IEEE Transactions on Biomedical Engineering 1990;37:757–67. Schutte LM, Hayden SW, Gage JR. Lengths of hamstrings and psoas muscles during crouch gait: effects of femoral anteversion. Journal of Orthopaedic Research 1997;15:615–21. Arnold AS, Salinas S, Asakawa DJ, Delp SL. Accuracy of muscle moment arms estimated from MRI-based musculoskeletal models of the lower extremity. Computer Aided Surgery 2000;5:108–19. Arnold AS, Delp SL. Rotational moment arms of the medial hamstrings and adductors vary with femoral geometry and limb position: implications for the treatment of internally rotated gait. Journal of Biomechanics 2001;34: 437–47. Desailly E, Sardain P, Khouri N, Yepremian D, Lacouture P. The convex wrapping algorithm: a method for identifying muscle paths using the underlying bone mesh. Journal of Biomechanics 2010;43(13):2601–7. Schwartz MH, Rozumalski A. The gait deviation index: a new comprehensive index of gait pathology. Journal of Pediatric Orthopedics 2008;28:351–7. Kay RM, Rethlefsen SA, Kelly JP, Wren TAL. Predictive value of the Duncan-Ely test in distal rectus femoris transfer. Journal of Pediatric Orthopedics 2004;24(1):59–62. Miller F, Dias RC, Lipton GE, Albarracin JP, Dabney KW, Castagno P. The effect of rectus EMG patterns on the outcome of rectus femoris transfers. Journal of Pediatric Orthopedics 1997;17(5):603. Muthusamy K, Seidl AJ, Friesen RM, Carollo JJ, Pan Z, Chang FM. Rectus femoris transfer in children with cerebral palsy: evaluation of transfer site and preoperative indicators. Journal of Pediatric Orthopedics 2008;28(6): 674–8. Reinbolt JA, Fox MD, Schwartz MH, Delp SL. Predicting outcomes of rectus femoris transfer surgery. Gait and Posture 2009;30(1):100–5.