- Email: [email protected]
Chopart's joint load during gait
Gait & Posture 27 (2008) 216–222 www.elsevier.com/locate/gaitpost
Chopart’s joint load during gait In vitro study of 10 cadaver specimen in a dynamic model A. Suckel a,*, O. Muller a, P. Langenstein a, T. Herberts b, P. Reize a, N. Wulker a a b
Orthopaedic Department, Tubingen University Hospital, Hoppe-Seyler Strasse 3, 72076 Tubingen, Germany Department of Medical Biometry, University of Tubingen, Westbahnhofstrasse 55, 72070 Tubingen, Germany Received 31 October 2006; received in revised form 18 March 2007; accepted 21 March 2007
Abstract Background: Chopart’s joint is fundamental to foot function. Until today, intra-articular force and peak pressure has not been investigated under dynamic conditions. Methods: The study used a cadaver model to measure intra-articular force and peak pressure with electronic sensors. Force was applied to extrinsic tendons via cables attached to computer-regulated hydraulic cylinders. A ground reaction force of 350 N was simulated in a tilting angle- and force-controlled translation stage. Results: We observed a characteristic rising curve with a peak during push-off for intra-articular force and peak pressure. The increase of intra-articular force at the talonavicular and calcaneocuboid joint from a low level at heel-on varies up to a maximum of 174 N/149 N and a peak pressure of 3877 kPa/3396 kPa, respectively, at push-off. We observed highest loading at the dorsal aspect of the talonavicular joint and the plantar aspect of calcaneocuboid joint. Conclusion: The highest loading on Chopart’s joint is attained during push-off. We observe higher force and peak pressure on the medial column of the foot compared to the lateral column. The higher load of the dorsal aspect of talonavicular joint and plantar aspect of calcaneocuboid joint confirms the theory of a previous described locking mechanism for forceful push-off. # 2007 Elsevier B.V. All rights reserved. Keywords: Chopart joint; Talonavicular joint; Calcaneocuboid joint; Dynamic foot model; Intra-articular loading
1. Introduction Chopart’s joint is the central articulation within the tarsus, i.e. the hindfoot. It enables movement of the midfoot and forefoot in three dimensions. Chopart’s joint consists of the articulation between the talus and the navicular bone at the medial column of the foot (TN joint) and laterally between the calcaneus and the cuboid bone (CC joint). Both are anatomically separate joints with distinct articular spaces and capsules, forming a single functional entity [1,2]. Located at the center of the foot, Chopart’s joint is of fundamental importance for basic foot function during gait. * Corresponding author at: University Hospital Tubingen, Orthopaedic Department, Hoppe-Seyler Strasse 3, 72076 Tubingen, Germany. Tel.: +49 7071 29 86688; fax: +49 7071 29 4091. E-mail address: [email protected] (A. Suckel). 0966-6362/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2007.03.010
Ambulation requires a flexible foot from heel contact until push-off. During this time, the deceleration energy of the downward movement of the body must be absorbed and the foot has to adjust to potentially uneven ground. In contrast, push-off requires a rigid foot that is able to transmit the force for upward acceleration of the body in relation to the ground. It has been postulated that the change from a flexible to a rigid foot is largely effected by Chopart’s joint [1,2]. The talonavicular joint has been identified as the key articulation of the triple joint complex of the hindfoot in previous studies using kinematic approaches [3,4]. So far, no valid biomechanical data on the intra-articular loads in Chopart’s joint exists. The aim of the study is to record intra-articular force and peak pressure during the stance phase of the gait, the load distribution between the medial and the lateral column of the foot, and an analysis of different load distribution within Chopart’s joint. The
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
217
Fig. 1. Simulated gait cycle left cadaver foot with sensor layers in situ, Steinmann pins for later simulated hind foot arthrodesis already placed. Images at 5 s, 15 s, 25 s, 35 s, 45 s, 55 s of the 60 s lasting gaitcycle. Normal foot motion was observed from heel contact to toe-off.
present study uses a validated dynamic foot model to simulate physiological gait [5,6]. 2. Materials and methods 2.1. Foot model The foot model simulated the stance phase of human walking inversely by hydraulically controlled movement of a base plate with respect to the fixed foot and tibia, which was described extensively [5]. Foot roll-off was simulated by a tilting, angle and force-controlled translation stage (base plate). A pressure-sensitive force plate (EMED SF, Novel GmbH, Munich, Germany) was mounted onto the base plate to record ground reaction force and the gait line of foot roll-off. Time-dependent forces were applied to the tendons of the foot flexor and extensor muscles via cables attached to an additional set of six force-controlled hydraulic cylinders. Force was directed to the tendon clamps via a pulley system. Tibial rotation was additionally achieved by an electrical servo motor. The position-controlled movement of the base plate and the force-controlled movement of the hydraulic cylinders were controlled by customprogrammed computer software [5]. Input data for the ground reaction force and tibia—ground angle was based on previous experiments [5,6]. Thus, the course of the ground reaction force over time represented a typical M-curve for half of body weight (350 N) in order to prevent damage to the specimens, and to ensure that the simulation could be repeated several times.
Gait was simulated over a standardized time of 60 s from heel strike (0%) to toe-off (100%) (Fig. 1). Based on previous experiments, the muscle force for each tendon was set as follows: Achilles tendon: 780 N at 70% of stance phase; tibialis posterior: 225 N at 75%; peroneal tendons: 132 N at 70%; flexor hallucis: 90 N at 85%; tibialis anterior: 66 N at heel contact; and extensors of toes: 19 N at 90% [5]. In each specimen, the force of simulated muscle was adjusted until the physiological kinetics and kinematics of the motion were reproduced as closely as possible, judged by the plantar pressure gait line (Fig. 2) [6,7] and by visual observation. 2.2. Specimens Lower limb specimens were obtained from ten macroscopically and radiologically normal fresh frozen foot specimens by transection at approximately mid-tibial length. Soft tissues were removed to roughly 4 cm above the ankle joint without destroying the
Fig. 2. Plantar pressure distribution during stance phase for a selected specimen, as well as mean curve for centre of pressure (gait line) with standard deviation.
218
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
superior extensor retinaculum. The tendons of the lower-leg muscles were detached from the leg by blunt dissection to allow attachment via clamps to cables by means of which force was applied. Nine tendons of the foot were simulated with six hydraulic cylinders: the triceps surae, tibialis posterior, flexor hallucis longus and flexor digitorum longus in combination, tibialis anterior, peroneus longus and brevis in combination, and extensor digitorum longus and extensor hallucis longus in combination. The proximal ends of the specimen tibia and fibula were dissected to the surface of the bone and embedded in their relative anatomical positions in a cold-setting methyl methacrylate resin (Technovit 4071, Heraeus Kulzer, GmbH, Wehrheim, Germany). The embedded tibia and fibula were mounted concentrically and vertically in an aluminum tibia mounting cylinder using heavy centering screws. Neutral rotation of the tibia was defined prior to testing by turning the mounting cylinder to orient the foot to 128 outward rotation with respect to the long axis of the pressure measurement platform [6]. 2.3. Chopart’s joint pressure measurements Intra-articular pressure was measured with resistive pressure sensors (K-scan sensor, map 4201, Tekscan Inc., Boston, MA). The 0.3 mm thick sensor sheets consist of 264 different sensors arranged in an 11 24 matrix, with a spatial resolution of 1.8 mm and a maximum pressure range of 13.8 MPa. The pressure values were recorded using the manufacturer’s software (Iscan) at a recording rate of 2 Hz for a time span of 60 s to achieve 120 pressure frames. Raw data were then exported as ASCII files and further analyzed at MatLab (TheMathwork Inc, Natik, MA). Separate sheets were used for the TN and the CC joints, which were measured simultaneously during the experiment. As such, the first sheet was cut along its electronic pathways to obtain sensor areas of 9 mm 24 mm and 11 mm 24 mm, corresponding to the anatomical natures of the joints. The cut sheets were sealed carefully to exclude moisture, and were fixed inside the joint with sutures to the capsular tissue. Each sensor part was calibrated separately in a material testing machine (ZwickRoell, Ulm Germany) with a piece of rubber (area 1 cm2) at up to 300 N. During the experiments, the position of the sheets was closely observed by the investigators in order to rule out displacement during movement of the foot. Each experiment was performed with five repetitions.
Each sensor part was calibrated by one-point calibration (300 N) using the Matlab program developed to do all the data analyses. Different regions of the joint – plantar or dorsal, for instance – could be evaluated using appropriate submasks of the sensor matrix. Maximum peak pressure (MPP) values were determined for each pressure frame to get the course of MPP over time along the stance phase as well as the overall force of the submask of interest. Average courses of force and MPP over time (n = 5) were used for subsequent calculations and further analysis. The standard deviation (S.D.) at each time-stamp characterizes the variance of the measurements. In order to summarize all data for the specimens, each course over time (of force and MPP) was normalized to its maximum value (100%) and averaged over all specimens. Results for the TN joint of specimen #3 were excluded because the sensor sheets dislocated during movement of the foot. Furthermore, data from the TN joint in specimen #9 and from the CC joint in specimen #10 had to be excluded because of incorrect placement of the pressure layer in the joints that resulted in only one small edge of the sensor being loaded. To examine whether foot length has an effect on maximum force and/or maximum peak pressure, we performed analyses of variance, taking the joint and the foot length as fixed nominal factors and the foot as a random factor.
3. Results All specimens showed the same characteristic course of intra-articular force and peak pressure over time. Low pressure values were present following heel contact and for a variable period of up to 20% of the gait cycle. This was followed by a gradual pressure increase, which lasted until 60–80% of the gait cycle. Finally, pressure receded at 90– 100% of the gait cycle. The mean of five replications of maximum intra-articular force and MPP values recorded in each specimen are shown in Table 1. By way of example, we show the course of intraarticular force and peak pressure over time for two specimens for the TN joint (Fig. 3a) and for the CC joint (Fig. 3b). The summarized data (MPP) for all specimens are shown in Fig. 4a for the TN joint and in Fig. 4b for the CC joint.
Table 1 Maximum force and maximum peak pressure (MPP) during stance phase in talonavicular joint (TN) and calcaneocuboid joint (CC)—mean, minimum and maximum of five replications per foot Foot (footlength in cm)
TN
CC
Maximum force (N) (min–max)
MPP (kPa) (min–max)
Maximum force (N) (min–max)
MPP (kPa) (min–max)
1 (26) 2 (26) 3 (24) 4 (25) 5 (27) 6 (26) 7 (24) 8 (24) 9 (23) 10 (25)
234 247 – 160 410 74 98 188 – 133
3519 8166 – 4389 5834 2803 2159 3793 – 3963
149 41 149 33 240 174 140 166 323 –
3816 3139 3563 1264 5867 3131 2278 3396 4950 –
Mean
170
(209–262) (179–293) (155–167) (409–412) (72–75) (96–99) (186–188) (123–145)
4328
(3118–3693) (7931–8284) (4358–4436) (5575–5968) (2709–2866) (2081–2277) (3730–3887) (3901–4106)
158
(126–177) (40–44) (139–155) (31–34) (224–247) (171–175) (138–144) (160–175) (311–329)
3489
(3361–4316) (3058–3259) (3335–3806) (1207–1368) (5674–5996) (3098–3179) (2253–2334) (3340–3420) (4829–5110)
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
219
Fig. 3. (a) Exemplary time course of force (left column, in N) and MPP (right column, in kPa) at the TN joint for two exemplary specimens (#5 and #8) during stance phase (mean of five cycles, standard deviation). (b) Exemplary time course of force (left column, in N) and MPP (right column, in kPa) at the CC joint for two selected specimens (#5 and #8) during stance phase (mean of five cycles, standard deviation).
220
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
Fig. 4. (a) Summarized time course of MPP during stancephase in gait in the TN joint for eight specimen (with standard deviation of the normalized values). (b) Summarized time course of MPP during stancephase in gait in the CC joint for nine specimen (with standard deviation of the normalized values).
On average, we observed higher force and peak pressure in all specimens on the medial column of the feet, at 174 N (range 74–410 N)/3877 kPa (range 2159–8166 kPa), compared to the lateral column, at 149 N (range 33–323 N)/ 3396 kPa (range 1264–5867 kPa) as shown in Table 1. We also observed a higher force load in the TN joint in 5 specimens (#1, #2, #4, #5 and #8) and in the CC joint (#6 and #7) in two specimens. We measured a higher peak pressure in the TN joint in 3 specimens (#2, #4 and #8). In 2 specimens (#5 and #7), this was somewhat evened out and in 2 specimens (#1 and #6), the peak pressure was higher in the CC joint than in the TN joint. We saw a higher mean force on the dorsal aspect of TN joint of 60 N (range 5–142 N) or 3247 kPa (range 1720– 5756 kPa) in relation to the plantar aspect of 45 N (range 1– 105 N) or 3141 kPa (range 440–8119 kPa). In CC joint, the load of the plantar aspect was higher. We recorded 48 N (range 4–113 N) or 3346 kPa (range 459–5762 kPa) on the plantar side compared to 26 N (range 0–57 N) or 2533 kPa (range 129–4209 kPa) on the dorsal half of the joint. Mean force and maximum peak pressure distribution between dorsal and plantar compartment of the TN joint was higher at the dorsal compartment for 5 specimens (#1, #4–6 and #10) out of 8, with highest load during push-off. In one specimen (#7) the load distribution was almost the same, while in two specimens (#2 and #8) we observed higher load on the plantar side of the talonavicular joint. The results are presented in Table 2. In seven specimens (#1–3, #5, #6, #8 and #9) out of 9, higher values were recorded in the plantar part of the CC joint compared to the dorsal part, in one specimen (#7) the load distribution between plantar and dorsal side was similar and in one specimen (#4) the value on the dorsal aspect was higher. While there is a statistically significant correlation ( p < 0.0001) between foot length (see Table 1) and the sum of forces in both parts of the Chopart joint, the force does not increase with foot length, rather first decreases and then increases. There is no significant correlation between foot length and peak pressure ( p = 0.11).
Table 2 Dorsal/plantar load distribution in talonavicular joint (TN) and calcaneocuboid joint (CC) force in N as mean force (time–force integral), maximum peak pressure (MPP) in kPa Foot
1 2 3 4 5 6 7 8 9 10 Mean
TN-force (N)
TN-MPP (kPa)
CC-force (N)
CC-MPP (kPa)
Dorsal
Plantar
Dorsal
Plantar
Dorsal
Plantar
Dorsal
Plantar
81 5 – 100 142 38 21 29 – 62
64 90 – 1 105 3 24 61 – 8
3197 1979 – 4389 5756 2795 1720 2206 – 3932
3154 8119 – 440 5465 856 2120 3761 – 1211
28 0 36 15 21 27 33 18 57 –
37 8 49 4 76 60 31 57 113 –
3137 129 3296 1215 3992 2382 2237 2197 4209 –
3796 3034 3406 459 5762 3131 2197 3388 4942 –
60
45
3247
3141
26
48
2533
3346
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
4. Discussion There has only been scanty experimental data so far on forces and pressure values within the TN and CC joints while walking. Wayne et al. [8] used a static lower leg model and tried to evaluate changes in the talonavicular joint following tibiotalar arthrodesis. They did not find any differences concerning peak pressure compared to the native situation, while they did not present any specific data concerning intra-articular force or peak pressure. Under an axial static load of 445 N, only very little contact pressures were mentioned with the use of Fuji-films. Bertsch et al. [9] investigated pressures in Chopart’s joint with a static model in various positions of dorsiflexion and plantar flexion of the ankle–foot complex and found varying pressure load and forces for different positions of the foot in Chopart’s joint with an increase during plantar flexion. Under axial loading of 740 N intra-articular force and peak pressure values of 159 N and 2300 kPa in the TN joint and of 136 N and 2300 kPa in the CC joint, respectively, were recorded. The values of this study were lower than our data of 174 N and 3877 kPa and 149 N and 3396 kPa at push-off under 350 N axial loading. This is easy to understand, the more so since tendon tension could not be simulated as a dynamic experimental approach in the cited static experiment. Interestingly, the ratio of force distribution between TN joint and CC joint was similar in both studies, there being discretely higher values on the medial side of the foot. Chopart’s joint is responsible for approximately 20% of foot flexion and extension and significant abduction/ adduction and supination/pronation during normal gait, with high interindividual variance [10–12]. During heel contact in the early phase of the gait cycle, 80% of the body weight is transmitted directly to the heel via talus and calcaneus [7]. When the heel lifts off, the entire load is transmitted to the forefoot via Chopart’s joint. Thus, the load at Chopart’s joint varies greatly during the gait cycle. When looked at together with the muscle force profiles, the course of peak pressure and force load can thus be meaningfully understood. Of particular interest is the load distribution between and within the TN and CC joints. Pathological conditions of the foot may lead to a change in distribution of weight-bearing between medial and lateral column of the foot. In an in vitro study, Rosenbaum et al. [13] showed a higher pressure load following calcaneal fracture of the CC joint. The transfer of load from the hindfoot to the forefoot via the medial and lateral column of the foot appears to vary depending on the individual anatomy. In normal gait, this will also largely depend on angulations and unevenness of the ground. Our data suggest, on flat ground, a slightly higher load transfer in the middle via the medial column of the foot in comparison to the lateral column, where a significant deviation from this force distribution can occur individually, as was also demonstrated in our experiments.
221
Pressures at the TN joint mostly increased at its dorsal aspect during the gait cycle. This corresponds to clinical observations. Degeneration at the TN joint usually tends to occur first at the dorsal aspect of the articular surface, often with the formation of prominent osteophytes at the dorsal articular margins of the talus and the navicular bone. The fact that degenerative changes are far more commonly observed at the TN joint than at the CC joint may be due to significantly greater pressure increases in the former compared to the latter. More importantly, loading in the later phase of the gait cycle diverged between the TN joint (dorsal high pressures) and the CC joint (plantar high pressures). This points towards a fundamental mechanism of hindfoot mechanics, which has been known since Elftman [1]: during stance with physiological heel eversion, the axes of TN and CC joints run parallel, leading to an even load distribution and flexible flexion and extension. At push-off with heel inversion, both axes become divergent, which locks Chopart’s joint and creates a rigid foot to elevate the body from the ground [1]. In the first stance phase during heel-on, Chopart’s joint is unlocked for all movements enabling the foot to react to individual ground conditions. As we were able to prove, the joint is not loaded at this phase and therefore can adapt to ground conditions. The locking mechanism is demonstrated in our study by a divergence of forces at the dorsal aspect of the TN joint and plantar aspect of the CC joint during powerful push-off of the foot (Fig. 5). We have to consider the following situation in the in vitro study that was abnormal compared to an in vivo-situation. The model allowed a weight application of only 350 N, i.e. about half the body weight. There is no proof that allows us to presume that the values in Chopart’s joint will double
Fig. 5. Exemplary time course of force in the TN joint (blue) and in the CC joint (red) in a selected specimen (#8), diverging active stabilisation of Chpart’s joint caused by a increase of forces in contralateral direction of dorsal part of talonavicular joint and plantar part of calcaneocuboid joint during push-off (mean of five cycles, standard deviation). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of the article.)
222
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
under a weight application of 700 N. Hintermann and Nigg [14] found no changes in the axis of bones of the foot independent of axial loading. We can therefore assume that our load-reduced model is valid with respect to the orientation of the hindfoot bone axis. The body weight of our specimen was not known, i.e. there was a potential individual measurement error for the feet of up to 100% in a small person of 50 kg weight as compared to a person of 100 kg. No other workgroup has described the physiologically realistic simulation of the roll-off process in a dynamic experiment of comparable configuration, and in particular the individual forces for the individual tendon tension and their force profile over time are not clearly scientifically documented. Our dynamic roll-off model was validated by Hurschler et al. [5] and the application of force over time defined. The application of force on the tendons influences the type of roll-off process, and there are interindividually different anatomical and functional principles that made it necessary to adapt the muscle force profiles individually for the specimens. Adjustments in the order of 10% increase or decrease of applied force were made before the actual measurements, and in some cases the peak values were changed by 10% over the course of time. This was necessary above all since the roll-off process either went slightly over the medial foot, or over the lateral edge of the foot. Accordingly, either the force values for the tibialis posterior, tibialis anterior or peroneal tendons were gradually adjusted. Even slight changes in these force profiles led to a change in the roll-off process. The plantar pressure curve changed significantly and, from a visual impression, a physiological roll-off movement could be easily distinguished from an atypical roll-off movement. By nature, the pressure distribution between medial and lateral column of the foot, and thus in the TN and CC joint, changes with every tendon adaptation. These differences in the force and pressure distribution between the two components of the Chopart joint may be assumed to be interindividually physiological given the known great variance in anatomical and functional conditions of the foot. As such, our measured force and pressure curves are only an average value and an indication for estimating the force distribution between medial and lateral columns of the foot, and also peak pressure loads in the joints, and significant variance in the results is possible for individual cases. This variance can be further explained by different foot sizes, body weights and muscular strengths. We see a statistical correlation between the sum of values for force loads in the two parts of the Chopart joint, yet this correlation does not fully explain the high interindividual variance. Body weight could not be evaluated with certainty in all cases for our specimens, so we could not statistically prove any clear influence of this parameter, yet we consider it probable that there is an influence. There also exist individual articular conditions in the Chopart joint in the form of joint area sizes and the inclination of the joint surfaces—the two extremes being high arches and flatfoot.
Presented for the first time are intra-articular force load and pressure measurements in the Chopart joint taken in a physiological simulation in a dynamic cadaver model. We calculated a typical curve for the load of the two joint parts with a punctum maximum after about 70% of the stance phase. With high interindividual variance, we measured higher force and pressure values on the medial column of the foot in the TN joint, at 174 N or 3877 kPa, than on the lateral column in the CC joint, at 149 N or 3396 kPa. What is functionally important is the observation of a pressure increase dorsally in the TN joint and ventrally in the CC joint upon plantar flexion during the push-off phase of the foot. We interpreted this measurement as a supportive effect of the locking of the Chopart joint upon push-off of the foot by the spreading of talus and calcaneus within the convex joint surfaces of the navicular and cuboid.
Acknowledgement The study was supported by Deutsche Arthrose Hilfe, a beneficial organisation.
References [1] Elftman H. The transverse tarsal joint and ist control. Clin Orthop 1960;16:41–6. [2] Mann RA, Coughlin MJ. Surgery of the foot and ankle, 6th ed., St. Louis: Mosby; 1993. [3] Astion DJ, Deland JT, Otis JC, Kenneally S. Motion of the hindfoot after simulated arthrodesis. J Bone Joint Surg [Am] 1997;79a:241–6. [4] Savory KM, Wulker N, Stukenborg C, Alfke D. Biomechanics of the hindfoot joints in response to degenerative hindfoot arthrodeses. Clin Biomech 1998;13:62–70. [5] Hurschler C, Emmerich J, Wulker N. In vitro simulation of stance phase gait. Part I. Model verification. Foot Ankle Int 2003;24:614–22. [6] Wulker N, Hurschler C, Emmerich J. In vitro simulation of stance phase gait. Part II. Simulated anterior tibial tendon dysfunction and potential compensation. Foot Ankle Int 2003;24:623–9. [7] Imamura M, Imamura ST, Salomao O, Pereira CA, De Carvalho Jr AE, Neto RB. Pedobarometric evaluation of the normal adult male foot. Foot Ankle Int 2003;23:804–10. [8] Wayne JS, Lawhorn KW, Davis KE, Prakash K, Adelaar RS. The effect of tibiotalar fixation on foot biomechanics. Foot Ankle Int 1997;18:792–7. [9] Bertsch C, Rosenbaum D, Claes L. Intra-articular and plantar pressure distribution of the ankle joint complex in relation to foot position. Unfallchirurg 2001;104:426–33. [10] Grass R, Rammelt S, Endres T, Zwipp H. Reorientational arthrodesis of the ankle joint using four screws. Orthopade 2005;34:1209–15. [11] Lundberg A, Goldie I, Kalin B, Selvik G. Kinematics of the ankle/foot complex: plantarflexion and dorsiflexion. Foot Ankle 1989;9:194–200. [12] Takakura Y, Tanaka Y, Sugimoto K, Akiyama K, Tamai S. Long-term results of arthrodesis for osteoarthritis of the ankle. Clin Orthop Relat Res 1999;361:178–85. [13] Rosenbaum D, Bauer G, Augat P, Claes L. Calcaneal fractures cause a lateral load shift in Chopart joint contact stress and plantar pressure pattern in vitro. J Biomech 1996;29:1435–43. [14] Hintermann B, Nigg BM. In vitro kinematics of the axially loaded ankle complex in response to dorsiflexion and plantarflexion. Foot Ankle Int 1995;16:514–8.
Chopart’s joint load during gait In vitro study of 10 cadaver specimen in a dynamic model A. Suckel a,*, O. Muller a, P. Langenstein a, T. Herberts b, P. Reize a, N. Wulker a a b
Orthopaedic Department, Tubingen University Hospital, Hoppe-Seyler Strasse 3, 72076 Tubingen, Germany Department of Medical Biometry, University of Tubingen, Westbahnhofstrasse 55, 72070 Tubingen, Germany Received 31 October 2006; received in revised form 18 March 2007; accepted 21 March 2007
Abstract Background: Chopart’s joint is fundamental to foot function. Until today, intra-articular force and peak pressure has not been investigated under dynamic conditions. Methods: The study used a cadaver model to measure intra-articular force and peak pressure with electronic sensors. Force was applied to extrinsic tendons via cables attached to computer-regulated hydraulic cylinders. A ground reaction force of 350 N was simulated in a tilting angle- and force-controlled translation stage. Results: We observed a characteristic rising curve with a peak during push-off for intra-articular force and peak pressure. The increase of intra-articular force at the talonavicular and calcaneocuboid joint from a low level at heel-on varies up to a maximum of 174 N/149 N and a peak pressure of 3877 kPa/3396 kPa, respectively, at push-off. We observed highest loading at the dorsal aspect of the talonavicular joint and the plantar aspect of calcaneocuboid joint. Conclusion: The highest loading on Chopart’s joint is attained during push-off. We observe higher force and peak pressure on the medial column of the foot compared to the lateral column. The higher load of the dorsal aspect of talonavicular joint and plantar aspect of calcaneocuboid joint confirms the theory of a previous described locking mechanism for forceful push-off. # 2007 Elsevier B.V. All rights reserved. Keywords: Chopart joint; Talonavicular joint; Calcaneocuboid joint; Dynamic foot model; Intra-articular loading
1. Introduction Chopart’s joint is the central articulation within the tarsus, i.e. the hindfoot. It enables movement of the midfoot and forefoot in three dimensions. Chopart’s joint consists of the articulation between the talus and the navicular bone at the medial column of the foot (TN joint) and laterally between the calcaneus and the cuboid bone (CC joint). Both are anatomically separate joints with distinct articular spaces and capsules, forming a single functional entity [1,2]. Located at the center of the foot, Chopart’s joint is of fundamental importance for basic foot function during gait. * Corresponding author at: University Hospital Tubingen, Orthopaedic Department, Hoppe-Seyler Strasse 3, 72076 Tubingen, Germany. Tel.: +49 7071 29 86688; fax: +49 7071 29 4091. E-mail address: [email protected] (A. Suckel). 0966-6362/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2007.03.010
Ambulation requires a flexible foot from heel contact until push-off. During this time, the deceleration energy of the downward movement of the body must be absorbed and the foot has to adjust to potentially uneven ground. In contrast, push-off requires a rigid foot that is able to transmit the force for upward acceleration of the body in relation to the ground. It has been postulated that the change from a flexible to a rigid foot is largely effected by Chopart’s joint [1,2]. The talonavicular joint has been identified as the key articulation of the triple joint complex of the hindfoot in previous studies using kinematic approaches [3,4]. So far, no valid biomechanical data on the intra-articular loads in Chopart’s joint exists. The aim of the study is to record intra-articular force and peak pressure during the stance phase of the gait, the load distribution between the medial and the lateral column of the foot, and an analysis of different load distribution within Chopart’s joint. The
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
217
Fig. 1. Simulated gait cycle left cadaver foot with sensor layers in situ, Steinmann pins for later simulated hind foot arthrodesis already placed. Images at 5 s, 15 s, 25 s, 35 s, 45 s, 55 s of the 60 s lasting gaitcycle. Normal foot motion was observed from heel contact to toe-off.
present study uses a validated dynamic foot model to simulate physiological gait [5,6]. 2. Materials and methods 2.1. Foot model The foot model simulated the stance phase of human walking inversely by hydraulically controlled movement of a base plate with respect to the fixed foot and tibia, which was described extensively [5]. Foot roll-off was simulated by a tilting, angle and force-controlled translation stage (base plate). A pressure-sensitive force plate (EMED SF, Novel GmbH, Munich, Germany) was mounted onto the base plate to record ground reaction force and the gait line of foot roll-off. Time-dependent forces were applied to the tendons of the foot flexor and extensor muscles via cables attached to an additional set of six force-controlled hydraulic cylinders. Force was directed to the tendon clamps via a pulley system. Tibial rotation was additionally achieved by an electrical servo motor. The position-controlled movement of the base plate and the force-controlled movement of the hydraulic cylinders were controlled by customprogrammed computer software [5]. Input data for the ground reaction force and tibia—ground angle was based on previous experiments [5,6]. Thus, the course of the ground reaction force over time represented a typical M-curve for half of body weight (350 N) in order to prevent damage to the specimens, and to ensure that the simulation could be repeated several times.
Gait was simulated over a standardized time of 60 s from heel strike (0%) to toe-off (100%) (Fig. 1). Based on previous experiments, the muscle force for each tendon was set as follows: Achilles tendon: 780 N at 70% of stance phase; tibialis posterior: 225 N at 75%; peroneal tendons: 132 N at 70%; flexor hallucis: 90 N at 85%; tibialis anterior: 66 N at heel contact; and extensors of toes: 19 N at 90% [5]. In each specimen, the force of simulated muscle was adjusted until the physiological kinetics and kinematics of the motion were reproduced as closely as possible, judged by the plantar pressure gait line (Fig. 2) [6,7] and by visual observation. 2.2. Specimens Lower limb specimens were obtained from ten macroscopically and radiologically normal fresh frozen foot specimens by transection at approximately mid-tibial length. Soft tissues were removed to roughly 4 cm above the ankle joint without destroying the
Fig. 2. Plantar pressure distribution during stance phase for a selected specimen, as well as mean curve for centre of pressure (gait line) with standard deviation.
218
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
superior extensor retinaculum. The tendons of the lower-leg muscles were detached from the leg by blunt dissection to allow attachment via clamps to cables by means of which force was applied. Nine tendons of the foot were simulated with six hydraulic cylinders: the triceps surae, tibialis posterior, flexor hallucis longus and flexor digitorum longus in combination, tibialis anterior, peroneus longus and brevis in combination, and extensor digitorum longus and extensor hallucis longus in combination. The proximal ends of the specimen tibia and fibula were dissected to the surface of the bone and embedded in their relative anatomical positions in a cold-setting methyl methacrylate resin (Technovit 4071, Heraeus Kulzer, GmbH, Wehrheim, Germany). The embedded tibia and fibula were mounted concentrically and vertically in an aluminum tibia mounting cylinder using heavy centering screws. Neutral rotation of the tibia was defined prior to testing by turning the mounting cylinder to orient the foot to 128 outward rotation with respect to the long axis of the pressure measurement platform [6]. 2.3. Chopart’s joint pressure measurements Intra-articular pressure was measured with resistive pressure sensors (K-scan sensor, map 4201, Tekscan Inc., Boston, MA). The 0.3 mm thick sensor sheets consist of 264 different sensors arranged in an 11 24 matrix, with a spatial resolution of 1.8 mm and a maximum pressure range of 13.8 MPa. The pressure values were recorded using the manufacturer’s software (Iscan) at a recording rate of 2 Hz for a time span of 60 s to achieve 120 pressure frames. Raw data were then exported as ASCII files and further analyzed at MatLab (TheMathwork Inc, Natik, MA). Separate sheets were used for the TN and the CC joints, which were measured simultaneously during the experiment. As such, the first sheet was cut along its electronic pathways to obtain sensor areas of 9 mm 24 mm and 11 mm 24 mm, corresponding to the anatomical natures of the joints. The cut sheets were sealed carefully to exclude moisture, and were fixed inside the joint with sutures to the capsular tissue. Each sensor part was calibrated separately in a material testing machine (ZwickRoell, Ulm Germany) with a piece of rubber (area 1 cm2) at up to 300 N. During the experiments, the position of the sheets was closely observed by the investigators in order to rule out displacement during movement of the foot. Each experiment was performed with five repetitions.
Each sensor part was calibrated by one-point calibration (300 N) using the Matlab program developed to do all the data analyses. Different regions of the joint – plantar or dorsal, for instance – could be evaluated using appropriate submasks of the sensor matrix. Maximum peak pressure (MPP) values were determined for each pressure frame to get the course of MPP over time along the stance phase as well as the overall force of the submask of interest. Average courses of force and MPP over time (n = 5) were used for subsequent calculations and further analysis. The standard deviation (S.D.) at each time-stamp characterizes the variance of the measurements. In order to summarize all data for the specimens, each course over time (of force and MPP) was normalized to its maximum value (100%) and averaged over all specimens. Results for the TN joint of specimen #3 were excluded because the sensor sheets dislocated during movement of the foot. Furthermore, data from the TN joint in specimen #9 and from the CC joint in specimen #10 had to be excluded because of incorrect placement of the pressure layer in the joints that resulted in only one small edge of the sensor being loaded. To examine whether foot length has an effect on maximum force and/or maximum peak pressure, we performed analyses of variance, taking the joint and the foot length as fixed nominal factors and the foot as a random factor.
3. Results All specimens showed the same characteristic course of intra-articular force and peak pressure over time. Low pressure values were present following heel contact and for a variable period of up to 20% of the gait cycle. This was followed by a gradual pressure increase, which lasted until 60–80% of the gait cycle. Finally, pressure receded at 90– 100% of the gait cycle. The mean of five replications of maximum intra-articular force and MPP values recorded in each specimen are shown in Table 1. By way of example, we show the course of intraarticular force and peak pressure over time for two specimens for the TN joint (Fig. 3a) and for the CC joint (Fig. 3b). The summarized data (MPP) for all specimens are shown in Fig. 4a for the TN joint and in Fig. 4b for the CC joint.
Table 1 Maximum force and maximum peak pressure (MPP) during stance phase in talonavicular joint (TN) and calcaneocuboid joint (CC)—mean, minimum and maximum of five replications per foot Foot (footlength in cm)
TN
CC
Maximum force (N) (min–max)
MPP (kPa) (min–max)
Maximum force (N) (min–max)
MPP (kPa) (min–max)
1 (26) 2 (26) 3 (24) 4 (25) 5 (27) 6 (26) 7 (24) 8 (24) 9 (23) 10 (25)
234 247 – 160 410 74 98 188 – 133
3519 8166 – 4389 5834 2803 2159 3793 – 3963
149 41 149 33 240 174 140 166 323 –
3816 3139 3563 1264 5867 3131 2278 3396 4950 –
Mean
170
(209–262) (179–293) (155–167) (409–412) (72–75) (96–99) (186–188) (123–145)
4328
(3118–3693) (7931–8284) (4358–4436) (5575–5968) (2709–2866) (2081–2277) (3730–3887) (3901–4106)
158
(126–177) (40–44) (139–155) (31–34) (224–247) (171–175) (138–144) (160–175) (311–329)
3489
(3361–4316) (3058–3259) (3335–3806) (1207–1368) (5674–5996) (3098–3179) (2253–2334) (3340–3420) (4829–5110)
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
219
Fig. 3. (a) Exemplary time course of force (left column, in N) and MPP (right column, in kPa) at the TN joint for two exemplary specimens (#5 and #8) during stance phase (mean of five cycles, standard deviation). (b) Exemplary time course of force (left column, in N) and MPP (right column, in kPa) at the CC joint for two selected specimens (#5 and #8) during stance phase (mean of five cycles, standard deviation).
220
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
Fig. 4. (a) Summarized time course of MPP during stancephase in gait in the TN joint for eight specimen (with standard deviation of the normalized values). (b) Summarized time course of MPP during stancephase in gait in the CC joint for nine specimen (with standard deviation of the normalized values).
On average, we observed higher force and peak pressure in all specimens on the medial column of the feet, at 174 N (range 74–410 N)/3877 kPa (range 2159–8166 kPa), compared to the lateral column, at 149 N (range 33–323 N)/ 3396 kPa (range 1264–5867 kPa) as shown in Table 1. We also observed a higher force load in the TN joint in 5 specimens (#1, #2, #4, #5 and #8) and in the CC joint (#6 and #7) in two specimens. We measured a higher peak pressure in the TN joint in 3 specimens (#2, #4 and #8). In 2 specimens (#5 and #7), this was somewhat evened out and in 2 specimens (#1 and #6), the peak pressure was higher in the CC joint than in the TN joint. We saw a higher mean force on the dorsal aspect of TN joint of 60 N (range 5–142 N) or 3247 kPa (range 1720– 5756 kPa) in relation to the plantar aspect of 45 N (range 1– 105 N) or 3141 kPa (range 440–8119 kPa). In CC joint, the load of the plantar aspect was higher. We recorded 48 N (range 4–113 N) or 3346 kPa (range 459–5762 kPa) on the plantar side compared to 26 N (range 0–57 N) or 2533 kPa (range 129–4209 kPa) on the dorsal half of the joint. Mean force and maximum peak pressure distribution between dorsal and plantar compartment of the TN joint was higher at the dorsal compartment for 5 specimens (#1, #4–6 and #10) out of 8, with highest load during push-off. In one specimen (#7) the load distribution was almost the same, while in two specimens (#2 and #8) we observed higher load on the plantar side of the talonavicular joint. The results are presented in Table 2. In seven specimens (#1–3, #5, #6, #8 and #9) out of 9, higher values were recorded in the plantar part of the CC joint compared to the dorsal part, in one specimen (#7) the load distribution between plantar and dorsal side was similar and in one specimen (#4) the value on the dorsal aspect was higher. While there is a statistically significant correlation ( p < 0.0001) between foot length (see Table 1) and the sum of forces in both parts of the Chopart joint, the force does not increase with foot length, rather first decreases and then increases. There is no significant correlation between foot length and peak pressure ( p = 0.11).
Table 2 Dorsal/plantar load distribution in talonavicular joint (TN) and calcaneocuboid joint (CC) force in N as mean force (time–force integral), maximum peak pressure (MPP) in kPa Foot
1 2 3 4 5 6 7 8 9 10 Mean
TN-force (N)
TN-MPP (kPa)
CC-force (N)
CC-MPP (kPa)
Dorsal
Plantar
Dorsal
Plantar
Dorsal
Plantar
Dorsal
Plantar
81 5 – 100 142 38 21 29 – 62
64 90 – 1 105 3 24 61 – 8
3197 1979 – 4389 5756 2795 1720 2206 – 3932
3154 8119 – 440 5465 856 2120 3761 – 1211
28 0 36 15 21 27 33 18 57 –
37 8 49 4 76 60 31 57 113 –
3137 129 3296 1215 3992 2382 2237 2197 4209 –
3796 3034 3406 459 5762 3131 2197 3388 4942 –
60
45
3247
3141
26
48
2533
3346
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
4. Discussion There has only been scanty experimental data so far on forces and pressure values within the TN and CC joints while walking. Wayne et al. [8] used a static lower leg model and tried to evaluate changes in the talonavicular joint following tibiotalar arthrodesis. They did not find any differences concerning peak pressure compared to the native situation, while they did not present any specific data concerning intra-articular force or peak pressure. Under an axial static load of 445 N, only very little contact pressures were mentioned with the use of Fuji-films. Bertsch et al. [9] investigated pressures in Chopart’s joint with a static model in various positions of dorsiflexion and plantar flexion of the ankle–foot complex and found varying pressure load and forces for different positions of the foot in Chopart’s joint with an increase during plantar flexion. Under axial loading of 740 N intra-articular force and peak pressure values of 159 N and 2300 kPa in the TN joint and of 136 N and 2300 kPa in the CC joint, respectively, were recorded. The values of this study were lower than our data of 174 N and 3877 kPa and 149 N and 3396 kPa at push-off under 350 N axial loading. This is easy to understand, the more so since tendon tension could not be simulated as a dynamic experimental approach in the cited static experiment. Interestingly, the ratio of force distribution between TN joint and CC joint was similar in both studies, there being discretely higher values on the medial side of the foot. Chopart’s joint is responsible for approximately 20% of foot flexion and extension and significant abduction/ adduction and supination/pronation during normal gait, with high interindividual variance [10–12]. During heel contact in the early phase of the gait cycle, 80% of the body weight is transmitted directly to the heel via talus and calcaneus [7]. When the heel lifts off, the entire load is transmitted to the forefoot via Chopart’s joint. Thus, the load at Chopart’s joint varies greatly during the gait cycle. When looked at together with the muscle force profiles, the course of peak pressure and force load can thus be meaningfully understood. Of particular interest is the load distribution between and within the TN and CC joints. Pathological conditions of the foot may lead to a change in distribution of weight-bearing between medial and lateral column of the foot. In an in vitro study, Rosenbaum et al. [13] showed a higher pressure load following calcaneal fracture of the CC joint. The transfer of load from the hindfoot to the forefoot via the medial and lateral column of the foot appears to vary depending on the individual anatomy. In normal gait, this will also largely depend on angulations and unevenness of the ground. Our data suggest, on flat ground, a slightly higher load transfer in the middle via the medial column of the foot in comparison to the lateral column, where a significant deviation from this force distribution can occur individually, as was also demonstrated in our experiments.
221
Pressures at the TN joint mostly increased at its dorsal aspect during the gait cycle. This corresponds to clinical observations. Degeneration at the TN joint usually tends to occur first at the dorsal aspect of the articular surface, often with the formation of prominent osteophytes at the dorsal articular margins of the talus and the navicular bone. The fact that degenerative changes are far more commonly observed at the TN joint than at the CC joint may be due to significantly greater pressure increases in the former compared to the latter. More importantly, loading in the later phase of the gait cycle diverged between the TN joint (dorsal high pressures) and the CC joint (plantar high pressures). This points towards a fundamental mechanism of hindfoot mechanics, which has been known since Elftman [1]: during stance with physiological heel eversion, the axes of TN and CC joints run parallel, leading to an even load distribution and flexible flexion and extension. At push-off with heel inversion, both axes become divergent, which locks Chopart’s joint and creates a rigid foot to elevate the body from the ground [1]. In the first stance phase during heel-on, Chopart’s joint is unlocked for all movements enabling the foot to react to individual ground conditions. As we were able to prove, the joint is not loaded at this phase and therefore can adapt to ground conditions. The locking mechanism is demonstrated in our study by a divergence of forces at the dorsal aspect of the TN joint and plantar aspect of the CC joint during powerful push-off of the foot (Fig. 5). We have to consider the following situation in the in vitro study that was abnormal compared to an in vivo-situation. The model allowed a weight application of only 350 N, i.e. about half the body weight. There is no proof that allows us to presume that the values in Chopart’s joint will double
Fig. 5. Exemplary time course of force in the TN joint (blue) and in the CC joint (red) in a selected specimen (#8), diverging active stabilisation of Chpart’s joint caused by a increase of forces in contralateral direction of dorsal part of talonavicular joint and plantar part of calcaneocuboid joint during push-off (mean of five cycles, standard deviation). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of the article.)
222
A. Suckel et al. / Gait & Posture 27 (2008) 216–222
under a weight application of 700 N. Hintermann and Nigg [14] found no changes in the axis of bones of the foot independent of axial loading. We can therefore assume that our load-reduced model is valid with respect to the orientation of the hindfoot bone axis. The body weight of our specimen was not known, i.e. there was a potential individual measurement error for the feet of up to 100% in a small person of 50 kg weight as compared to a person of 100 kg. No other workgroup has described the physiologically realistic simulation of the roll-off process in a dynamic experiment of comparable configuration, and in particular the individual forces for the individual tendon tension and their force profile over time are not clearly scientifically documented. Our dynamic roll-off model was validated by Hurschler et al. [5] and the application of force over time defined. The application of force on the tendons influences the type of roll-off process, and there are interindividually different anatomical and functional principles that made it necessary to adapt the muscle force profiles individually for the specimens. Adjustments in the order of 10% increase or decrease of applied force were made before the actual measurements, and in some cases the peak values were changed by 10% over the course of time. This was necessary above all since the roll-off process either went slightly over the medial foot, or over the lateral edge of the foot. Accordingly, either the force values for the tibialis posterior, tibialis anterior or peroneal tendons were gradually adjusted. Even slight changes in these force profiles led to a change in the roll-off process. The plantar pressure curve changed significantly and, from a visual impression, a physiological roll-off movement could be easily distinguished from an atypical roll-off movement. By nature, the pressure distribution between medial and lateral column of the foot, and thus in the TN and CC joint, changes with every tendon adaptation. These differences in the force and pressure distribution between the two components of the Chopart joint may be assumed to be interindividually physiological given the known great variance in anatomical and functional conditions of the foot. As such, our measured force and pressure curves are only an average value and an indication for estimating the force distribution between medial and lateral columns of the foot, and also peak pressure loads in the joints, and significant variance in the results is possible for individual cases. This variance can be further explained by different foot sizes, body weights and muscular strengths. We see a statistical correlation between the sum of values for force loads in the two parts of the Chopart joint, yet this correlation does not fully explain the high interindividual variance. Body weight could not be evaluated with certainty in all cases for our specimens, so we could not statistically prove any clear influence of this parameter, yet we consider it probable that there is an influence. There also exist individual articular conditions in the Chopart joint in the form of joint area sizes and the inclination of the joint surfaces—the two extremes being high arches and flatfoot.
Presented for the first time are intra-articular force load and pressure measurements in the Chopart joint taken in a physiological simulation in a dynamic cadaver model. We calculated a typical curve for the load of the two joint parts with a punctum maximum after about 70% of the stance phase. With high interindividual variance, we measured higher force and pressure values on the medial column of the foot in the TN joint, at 174 N or 3877 kPa, than on the lateral column in the CC joint, at 149 N or 3396 kPa. What is functionally important is the observation of a pressure increase dorsally in the TN joint and ventrally in the CC joint upon plantar flexion during the push-off phase of the foot. We interpreted this measurement as a supportive effect of the locking of the Chopart joint upon push-off of the foot by the spreading of talus and calcaneus within the convex joint surfaces of the navicular and cuboid.
Acknowledgement The study was supported by Deutsche Arthrose Hilfe, a beneficial organisation.
References [1] Elftman H. The transverse tarsal joint and ist control. Clin Orthop 1960;16:41–6. [2] Mann RA, Coughlin MJ. Surgery of the foot and ankle, 6th ed., St. Louis: Mosby; 1993. [3] Astion DJ, Deland JT, Otis JC, Kenneally S. Motion of the hindfoot after simulated arthrodesis. J Bone Joint Surg [Am] 1997;79a:241–6. [4] Savory KM, Wulker N, Stukenborg C, Alfke D. Biomechanics of the hindfoot joints in response to degenerative hindfoot arthrodeses. Clin Biomech 1998;13:62–70. [5] Hurschler C, Emmerich J, Wulker N. In vitro simulation of stance phase gait. Part I. Model verification. Foot Ankle Int 2003;24:614–22. [6] Wulker N, Hurschler C, Emmerich J. In vitro simulation of stance phase gait. Part II. Simulated anterior tibial tendon dysfunction and potential compensation. Foot Ankle Int 2003;24:623–9. [7] Imamura M, Imamura ST, Salomao O, Pereira CA, De Carvalho Jr AE, Neto RB. Pedobarometric evaluation of the normal adult male foot. Foot Ankle Int 2003;23:804–10. [8] Wayne JS, Lawhorn KW, Davis KE, Prakash K, Adelaar RS. The effect of tibiotalar fixation on foot biomechanics. Foot Ankle Int 1997;18:792–7. [9] Bertsch C, Rosenbaum D, Claes L. Intra-articular and plantar pressure distribution of the ankle joint complex in relation to foot position. Unfallchirurg 2001;104:426–33. [10] Grass R, Rammelt S, Endres T, Zwipp H. Reorientational arthrodesis of the ankle joint using four screws. Orthopade 2005;34:1209–15. [11] Lundberg A, Goldie I, Kalin B, Selvik G. Kinematics of the ankle/foot complex: plantarflexion and dorsiflexion. Foot Ankle 1989;9:194–200. [12] Takakura Y, Tanaka Y, Sugimoto K, Akiyama K, Tamai S. Long-term results of arthrodesis for osteoarthritis of the ankle. Clin Orthop Relat Res 1999;361:178–85. [13] Rosenbaum D, Bauer G, Augat P, Claes L. Calcaneal fractures cause a lateral load shift in Chopart joint contact stress and plantar pressure pattern in vitro. J Biomech 1996;29:1435–43. [14] Hintermann B, Nigg BM. In vitro kinematics of the axially loaded ankle complex in response to dorsiflexion and plantarflexion. Foot Ankle Int 1995;16:514–8.