- Email: [email protected]
A comparison of multi-segment foot kinematics during level overground and treadmill walking
Gait & Posture 31 (2010) 104–108
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
A comparison of multi-segment foot kinematics during level overground and treadmill walking Kirsten Tulchin *, Michael Orendurff, Lori Karol Texas Scottish Rite Hospital for Children, Movement Science Laboratory, 2222 Welborn Street, Dallas, TX 75219, United States
A R T I C L E I N F O
A B S T R A C T
Article history: Received 26 February 2009 Received in revised form 21 August 2009 Accepted 12 September 2009
Previous work comparing treadmill and overground walking has focused on lower extremity motion and kinetics, with few identified differences. However, a comparison of multi-segment foot kinematics between these conditions has not been previously reported. Sagittal ankle motion using a single rigid body foot model and three-dimensional hindfoot and forefoot kinematics were compared during barefoot, level, overground walking at a self-selected speed and treadmill walking at a similar speed for 20 healthy adults. Slight differences were seen in ankle plantarflexion and hindfoot plantarflexion during first rocker, as well as peak forefoot eversion and abduction, however all changes were less than 38, and most were within the day-to-day repeatability. These results indicate that foot mechanics as determined using a multi-segment foot model were similar between overground and treadmill walking at similar speeds in healthy adults. Treadmill protocols may provide a controlled method to analyze a patient’s ability to adapt to walking at different speeds and surface slopes, which are encountered often during ambulation of daily living. ß 2009 Elsevier B.V. All rights reserved.
Keywords: Gait Multi-segment foot kinematics Treadmill Overground walking
1. Introduction Instrumented lower extremity gait analysis in the clinical setting typically focuses on overground walking at a self-selected speed. Ambulation of daily living in the community however often requires more demanding tasks such as speed variation, turning and negotiating uneven terrain [1,2]. The use of a treadmill provides clinicians a means to design controlled protocols to evaluate a patient’s ability to adapt to changes in speed and surface slope. The advantages of gait analysis using a treadmill protocol also include the ability to collect multiple, consecutive strides within a small capture volume, which allows data collection to occur within a shorter period of time and with less variability in speed between trials [3]. Previous work comparing overground level walking and treadmill walking has focused on differences in lower extremity kinematics and kinetics. Alton et al. [4] found greater hip range of motion, greater maximal hip flexion angle and higher cadence during treadmill walking. They also observed a slightly decreased stance time compared to overground walking. Similar findings of a higher cadence during treadmill walking was also reported by Murray et al. [5]; however these results were not duplicated by
* Corresponding author. Tel.: +1 214 559 7580. E-mail address: [email protected] (K. Tulchin). URL: http://www.tsrhc.org 0966-6362/$ – see front matter ß 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2009.09.007
Riley et al. [6]. In their study, Riley et al. reported on both kinematic and kinetic changes using a forceplate instrumented treadmill. In evaluating the minima and maxima values of each kinematic variable, they found that differences between conditions were less than 38. In addition all differences seen were within the coefficients of repeatability of overground walking tests, with similar findings seen in the kinetic variables. Of particular interest to this study, they found no significant differences in maximum ankle dorsiflexion or plantarflexion motion, moments or power [6]. The use of multi-segment foot models to evaluate the pathological foot has expanded greatly over the last decade. These models provide a means to assess foot mechanics in more detail compared to the traditional lower extremity single rigid body foot model. By dividing the foot into multiple ‘rigid’ segments, three-dimensional intersegmental motion can be assessed for the hindfoot and forefoot, for example. Benedetti et al. [7] stated that approximately 16 studies utilizing multi-segment foot models were reported in literature between 1999 and 2008, on topics ranging from posterior tibialis dysfunction to talipes equinovarus deformity. Further analysis of this literature indicates that 13 of these 16 studies were published between 2004 and 2008, illustrating the emerging interest in the use of multi-segment foot models as an outcome tool in pathological gait evaluations. Fifteen of the 16 studies were conducted during overground ambulation at subjects’ self-selected walking speed [8–21]. Rao et al. [22] conducted testing with all subjects walking overground at a speedof0.89 m/s, and only three reported monitoring speed to within 5% of each subject’s self-selected velocity [11,14,19].
K. Tulchin et al. / Gait & Posture 31 (2010) 104–108
105
Tulchin et al. [23] have recently reported that walking speed can significantly affect sagittal plane multi-segment foot kinematics in healthy adults. Changes in inter-segment foot kinematics were similar to that seen at the ankle, with a shift towards plantarflexion and reduced stance duration as a percent of the gait cycle at faster speeds. The treadmill can provide a structured protocol to assess patients with pathological foot conditions at various speeds and inclines, which these patients may be subjected to during ambulation of daily living. However, it is important to first determine if there are any underlying differences in foot biomechanics, as measured using a multi-segment foot model, between overground and treadmill walking at similar speeds. The hypothesis of the current study was there would be no differences seen in multi-segment foot kinematics during level overground walking at a self-selected speed and level treadmill ambulation at a similar speed in a group of healthy adult subjects. 2. Methods Twenty adults, six males and 14 females, underwent gait analysis with IRB informed consent (Table 1). Subjects were excluded if they had a history of lower extremity surgery, or recent lower extremity injury within the last 6 months. Subjects were instrumented with a modified Helen Hayes marker set [24], as well as a multi-segment foot marker set bilaterally [23,25]. For the foot model, two of the Helen Hayes markers placed on the foot (toe and heel) were used in conjunction with five additional markers to model the hindfoot and forefoot. Marker trajectories were collected using a 12 camera VICON Mx motion capture system operating at 120 Hz (VICON, Denver, CO). All data were filtered with a smoothing spline (Woltring) routine using a mean square error of 10 [26]. Subjects performed level, overground barefoot walking at a self-selected speed. Multiple trials were collected for each leg, however a single side was randomly chosen for analysis for each subject, which resulted in equal distribution between right and left legs across all subjects. Subjects were then asked to complete a treadmill protocol, walking at speeds beginning at 0.89 m/s and increasing in increments of 0.23 m/s up to 1.79 m/ s. The speed which most closely matched their self-selected walking speed during the overground condition was used for comparison. A single representative trial was chosen for each condition for comparison. Maximum and minimum values of sagittal plane ankle and three-dimensional hindfoot and forefoot motion were determined for each subject. Sagittal plane ankle motion as defined using a single rigid body foot model (Plug-In Gait, VICON) was determined. Hindfoot and forefoot angles were calculated using a custom written multi-segment foot kinematic model (Bodybuilder for Biomechanics, VICON). Sagittal plane motion of ankle and hindfoot motion were examined specifically during each of the three ankle rockers: (1) plantarflexion during weight acceptance, (2) dorsiflexion during single limb stance, and (3) plantarflexion at terminal stance [27]. The mean differences between conditions for all variables were determined for each individual subject, and one sample t-tests were conducted for the group with a hypothesized mean of zero, indicating that no differences between conditions were expected. In addition, mean differences between overground and treadmill walking were compared to the dayto-day repeatability of each variable based on previous work comparing two sessions of data collected 2 weeks apart on 10 healthy adults [25]. 2.1. Multi-segment foot model The multi-segment foot model employed for this work was a three segment model of the shank, hindfoot, and forefoot [23,25]. Marker placement on the foot is shown in Fig. 1. A marker was also placed on the medial malleoli during the static trial. For the shank segment: the origin was located at the midpoint of the lateral and medial malleoli (ankle joint center); the X-axis was defined as the line joining the lateral and medial malleoli; the Y-axis was perpendicular to a line joining the knee and ankle joint centers, and the X-axis; and the long axis, Z, was therefore defined as being orthogonal to the X and Y axes. The shank segment was offset in the coronal plane to be parallel to the plane of the floor during the static trial, thus eliminating offsets in hindfoot motion based on bimalleolar axis variations. Individual markers were placed on the posterior, medial and lateral calcaneus to define the hindfoot segment The foot was marked non-weight bearing such that, with the subject lying prone, the clinician palpated the medial and lateral borders of the Table 1 Subject demographics, including age, height and weight.
calcaneus and drew a midline bisection along the posterior calcaneus. During quiet standing, a custom hindfoot alignment device with a cross-hair laser, based on the novel device described by Wervey and Schwartz [28], was used to ensure proper placement of the medial and lateral calcaneus markers in the coronal plane, so that they were perpendicular to the posterior midline of the calcaneus. This method was used to assure accurate measurement of hindfoot inversion/eversion. The third marker was placed along the midline of the posterior calcaneus. Calipers were then used to ensure that the distances from the medial and lateral calcaneal markers to the posterior calcaneal marker were equal, decreasing erroneous hindfoot rotation. For the hindfoot: the origin was the midpoint of the lateral and medial calcaneal markers (mid-calcaneus); The X-axis was defined as the line joining the medial and lateral calcaneal markers; the Z-axis was defined perpendicular to the line joining the posterior calcaneus marker and the mid-calcaneus, and the X-axis; and the Y-axis was defined as orthogonal to the X and Z axes. The hindfoot segment was offset in the sagittal plane to be parallel to the plane of the floor when the foot was plantigrade. Palpation was used to place markers on the first and fifth metatarsal heads and the fifth metatarsal base. For the forefoot: the origin was defined as a point at the perpendicular bisector of the fifth metatarsal base to a line joining the second/third metatarsal heads to the posterior calcaneus (long axis of the foot); the X axis joined the heads of the first and fifth metatarsals; the Z axis was perpendicular to a line joining the fifth metatarsal base and the fifth metatarsal head, and the X axis; and the Y axis was orthogonal to the X and Z axes. Relative inter-segmental motion of the following segments were examined: hindfoot relative to the shank (termed hindfoot motion) and forefoot relative to the hindfoot (forefoot motion). Hindfoot motion represented motion at the ankle mortise and the subtalar joint. Forefoot motion included the Chopart and Lisfranc joints. All angles were described by Euler rotations about the proximal segment axes, using a XYZ (sagittal, coronal, transverse) rotation order.
3. Results
Subjects
N
Age (years)
Height (cm)
Weight (kg)
Female Male All subjects
14 6 20
25.0 4.3 25.3 6.2 25.1 4.8
166.8 4.3 182.0 7.2 171.4 8.8
62.3 10.0 84.9 16.6 69.1 15.9
Mean (standard deviation).
Fig. 1. Skin marker placement for the TSRH multi-segment foot model. Markers on the second/third metatarsal, lateral malleolus and posterior calcaneus, shown in gray, are used for the single rigid body foot model, as well as the multi-segment foot model. Markers on the heads of the first and fifth metatarsals, the fifth metatarsal base, medial and lateral calcaneus markers, shown in black, in addition to the posterior calcaneus marker, are used to define the hindfoot and forefoot segments in the multi-segment foot model.
The 1.12 m/s (2.5 mph) treadmill speed was selected for 11 subjects (average overground speed 1.17 0.05 m/s) and 1.34 m/s (3.0 mph) was used for the nine remaining subjects (average overground speed 1.31 0.07 m/s). Mean differences between
K. Tulchin et al. / Gait & Posture 31 (2010) 104–108
106
Table 2 Mean differences between overground and treadmill walking of temporal parameters, ankle motion using the single rigid body foot model and hindfoot and forefoot motion using the multi-segment foot model. Variable
Mean difference
Std dev.
t-test
Day-to-day repeatability
Temporal/spatial Cadence (steps/min) Stance time (%GC)
0.03 0.00
6.23 0.04
0.98 0.70
0.04 0.87
Ankle Max PF angle—first rocker (8) Max DF angle—second rocker (8) Max PF angle—third rocker (8)
1.5 1.2 3.0
2.2 3.4 7.5
0.006** 0.124 0.09
0.8 1.0 2.2
Hindfoot Max PF angle—first rocker (8) Max DF angle—second rocker (8) Max PF angle—third rocker (8) Total sagittal ROM (8) Max inversion angle (8) Max eversion angle (8) Total coronal ROM (8) Max internal rotation angle (8) Max external rotation angle (8) Total transverse ROM (8)
1.3 0.4 1.3 0.8 0.7 0.1 0.7 0.7 0.6 1.3
2.5 2.3 6.8 5.5 2.2 1.5 2.7 2.6 2.8 2.9
0.027** 0.43 0.42 0.511 0.20 0.80 0.24 0.20 0.39 0.06
3.1 3.1 2.9 1.9 1.3 2.7 2.8 4.5 3.7 1.8
Forefoot Max DF angle—second rocker (8) Max PF angle—third rocker (8) Total sagittal ROM (8) Max inversion (8) Max eversion (8) Total coronal ROM (8) Max adduction (8) Max abduction (8) Total transverse ROM (8)
0.3 0.3 0.02 0.04 1.5 1.4 0.9 1.2 2.2
2.3 2.2 3.4 2.0 2.1 3.1 1.9 1.6 2.1
0.60 0.55 0.98 0.94 0.005** 0.05 0.039** 0.003** <0.001**
3.3 3.7 2.7 2.1 2.5 1.6 2.9 3.0 1.5
Positive mean differences indicate that overground values were greater than treadmill values. The day-to-day repeatability based on previous work [25] examining the multisegment foot model is shown for comparison. The variables that had statistically significant differences between conditions and were greater than the day-to-day repeatability errors are highlighted in bold italics. %GC = percentage gait cycle, DF: dorsiflexion, PF: plantarflexion, ROM: range of motion.
conditions of temporal and spatial parameters as well as ankle and multi-segment kinematic foot variables are shown in Table 2, and kinematic curves for both conditions are shown in Fig. 2. There were no statistical differences in cadence or stance time (as a percentage of the gait cycle) between overground and treadmill walking. Using the single rigid body foot model, there was a statistically significant difference in the maximal plantarflexion angle during first rocker between conditions (p = 0.006). However, the average difference between conditions across subjects was only 1.5 2.28. Similar results were seen in the hindfoot, with a mean difference between conditions during first rocker of 1.3 2.58. At the forefoot, statistical differences were noted in the coronal and transverse planes relative to the hindfoot, with slightly decreased eversion (1.5 2.18), slightly reduced maximum adduction (0.9 1.98), maximal abduction (1.2 1.68) and decreased overall range of motion in the transverse plane (2.2 2.18) during treadmill walking compared to the overground condition. These differences averaged less than 2.58 across all subjects. In addition, all differences between conditions were less than the day-to-day repeatability of these measures except for forefoot transverse plane range of motion, which averaged a mean difference of 2.28 and had a repeatability of 1.58. 4. Discussion The current study found several statistically significant differences between overground and treadmill walking at similar speeds in healthy adults. However, most of these differences were less than the day-to-day repeatability of the model, and therefore were not considered to be indicative of clinically significant variations in foot mechanics. Treadmill walking may provide clinicians and researchers the ability to develop more functional protocols that involve controlling and varying the speed and
incline during testing. These modulations in surface slope and speed compared to typical self-selected level walking might help elicit functional deficits in patients with foot pathology which would otherwise not be seen during typical gait analysis. Previous studies [4,5] have reported increased cadence during treadmill walking versus overground ambulation, however this was not found in the current study. While a long accommodation period to treadmill walking following overground walking was not given, all subjects had self-reported experience walking on a treadmill. The results of the current study concur with the conclusions of Riley et al. [5] that the treadmill did not appear to cause differences in temporal parameters during gait. With the focus of the current study on multi-segment foot kinematics, previous studies examining ankle motion, using a single rigid foot segment approach, during overground and treadmill walking were of particular interest [4–6]. Murray et al. [5] reported significant increased dorsiflexion during overground walking compared to treadmill walking at slow and free walking speeds, although these differences were estimated to be only 38. However, other studies have reported no significant differences in peak ankle dorsiflexion or plantarflexion between treadmill and overground conditions [4,6]. The current study evaluated ankle motion in slightly more detail—examining motion specifically within each of the three rockers. Peak ankle plantarflexion during first rocker was found to be significant less during treadmill walking, however this change was only 1.58. Visual examination of average curves indicate a slight shift towards dorsiflexion during treadmill walking, which is similar to the supplemental data reported by Riley et al. [6], particularly during stance phase. However, peak dorsiflexion during second rocker and peak plantarflexion during third rocker were not shown to been significantly different between conditions in either study.
K. Tulchin et al. / Gait & Posture 31 (2010) 104–108
107
Fig. 2. Kinematic results for N = 20 healthy adults. Ankle dorsiflexion/plantarfleixon motion measured using the single rigid body foot model; hindfoot relative to tibia, and forefoot relative to hindfoot motion based on the multi-segment foot model. Mean std. dev.: gray bar represents overground condition. Black solid dashed lines represent the treadmill condition.
The aim of the current study was to expand the analysis of the foot by the utilization of a multi-segment foot model. Analysis indicated that sagittal plane forefoot motion was not different between overground and treadmill ambulation, however peak hindfoot plantarflexion during first rocker was slightly reduced during treadmill walking. One advantage of using a multisegment foot model was the ability to assess coronal and transverse plane motion. Overall there were no significant variations in hindfoot motion between conditions in the coronal or transverse planes. The forefoot however did show a decrease in maximal eversion (p = 0.005) which lead to a slight decrease in coronal plane ROM. Similar changes were also demonstrated in forefoot rotation, with an decrease in maximal abduction seen during treadmill walking. Most significant changes reported in the current study were less 2.58 and within the day-to-day repeatability of the model in healthy adults, and were not deemed to be of clinical significance. There are several factors which may influence the results of the current study. First, all walking trials were conducted during barefoot ambulation. While subjects self-reported that they were experienced treadmill walkers, none had experienced barefoot ambulation on a treadmill. There were no reported problems during testing, however it is unclear whether possible increased foot-to-belt sensations had an impact on foot mechanics. As previously stated there were no accommodation periods given to the subjects prior to treadmill data collection. As part of a larger protocol, multiple walking trials were recorded for at least one or two speeds prior to the speed selected for the current analysis. This
would allow subjects to have at least a 1–2 min accommodation to treadmill walking prior to the data used for this study. Previous work by Matsas et al. [29] indicated that 6 min of walking was required for accommodation to a treadmill. In addition, overground self-selected walking speed was not directly matched for each subject during treadmill ambulation. Subjects walked at five speeds ranging from 0.89 to 1.79 m/s, in increments of 0.23 m/s, therefore the maximum possible difference between overground and treadmill walking was 0.12 m/s. Lastly, during treadmill walking, all subjects were instructed to rest their hands on the side treadmill bars. It is unclear what effect this may have had foot mechanics and propulsion mechanisms. In summary, the current study found multi-segment foot kinematics during overground and treadmill walking to be similar in healthy adults, with only a slight decrease (2.28) in forefoot transverse plane ROM during treadmill walking. In addition ankle plantarflexion using a single rigid body foot model was shown to be significantly decreased by 1.58 during treadmill walking. Further analysis is needed to evaluate these walking conditions in patients with foot pathology. However, a treadmill can provide a means to assess a patient’s ability to accommodate to increased speed or incline. These modulations in surface slope and speed may provide clinicians a more detailed indication of a subject’s functional deficits during community ambulation. Conflict of interest statement The authors report no conflicts of interest.
108
K. Tulchin et al. / Gait & Posture 31 (2010) 104–108
References [1] Orendurff MS, Schoen JA, Bernatz GC, Crass SM, Klute GK. How humans walk: bout duration, steps per bout and rest duration. J Rehabil Res Dev 2009;45:1077– 90. [2] Glaister BC, Bernatz GC, Klute GK, Orendurff MS. Video task analysis of turning during activities of daily living. Gait Posture 2007;25:289–94. [3] Dingwll J, Cusumano J, Cavanaugh P, Sternad D. Local dynamic stability versus kinematic variability of continuous overground and treadmill walking. J Biomech Eng 2001;123(1):27. [4] Alton F, Baldey L, Caplan S, Morrissey MC. A kinematic comparison of overground and treadmill walking. J Biomech 1998;13:434–40. [5] Murray MP, Spurr GB, Sepic SB, Gardner GM, Mollinger LA. Treadmill vs. floor walking: kinematics, electromyogram, and heart rate. J Appl Physiol 1985;59(1): 87–91. [6] Riley PO, Paolini G, Della Croce U, Paylo KW, Kerrigan DC. A kinematic and kinetic comparison of overground and treadmill walking in healthy subjects. Gait Posture 2007;26(1):17–24. [7] Benedetti MG, Leardini A, Bianchi L, Berti L, Giannini S. Comparisons of outputs of different models for foot kinematics. In: Proceedings of the international foot and ankle biomechanics congress; 2008. p. 91. [8] Canseco K, Long J, Marks R, Khazzam M, Harris GF. Quantitative characterization of gait kinematics in patients with hallux rigidus using the Milwaukee Foot Model. J Ortho Res 2008;26:419–27. [9] Flemister AS, Neville CG, Houck J. The relationship between ankle, hindfoot, and forefoot position and posterior tibial muscle excursion. Foot Ankle Int 2007;28(4):448–55. [10] Huang UC, Harbst K, Kotajarvi B, Hansen D, Koff M, Kitaoka H, et al. Effects of ankle-foot orthoses on ankle and foot kinematics in patient with ankle osteoarthritis. Arch Phys Med Rehabil 2006;87:710–6. [11] Hunt AE, Smith RM. Mechanics and control of the flat versus normal foot during the stance phase of walking. Clin Biomech 2004;19:391–7. [12] Khazzam M, Long JT, Marks RM, Harris GF. Kinematic changes of the foot and ankle in patients with systemic rheumatoid arthritis and forefoot deformity. J Ortho Res 2007;25:319–29. [13] Levinger P, Gilleard W. Tibia and rearfoot motion and ground reaction forces in subjects with patellofemoral pain syndrome during walking. Gait Posture 2007;25(1):2–8.
[14] Neville C, Flemister A, Tome J, Houck J. Comparison of changes in posterior tibialis muscle length between subjects with posterior tibial tendon dysfunction and healthy controls during walking. JOSPT 2007;37(11):661–9. [15] Ness ME, Long J, Marks R, Harris G. Foot and ankle kinematics in patients with posterior tibial tendon dysfunction. Gait Posture 2008;27:331–9. [16] Rattanaprasert U, Smith R, Sullivan M, Gilleard W. Three-dimensional kinematics of the forefoot, rearfoot and leg with the function of tibialis posterior in comparison with normals during stance phase of walking. Clin Biomech 1999;14:14–23. [17] Ringleb SI, Karvos SJ, Kotajarvi BR, Hansen DK, Kitaoka HB, Kaufman KR. Changes in gait associated with acute stage II posterior tibial tendon dysfunction. Gait Posture 2007;25:555–64. [18] Theologis TN, Harrington ME, Thompson N, Benson KD. Dynamic foot movement in children treated for congenital talipes equinovarus. JBJS [Br] 2003;85-B:572–7. [19] Tome J, Nawoczenski DA, Flemister A, Houck J. Comparison of foot kinematics between subjects with posterior tibialis tendon dysfunction and healthy controls. JOSPT 2006;36(9):635–44. [20] Woodburn J, Nelson KM, Siegel KL, Kepple TM, Gerber LH. Multisegment foot motion during gait: proof of concept in rheumatoid arthritis. J Rheumat 2004;31(10):1918–27. [21] Wu WL, Su FC, Cheng YM, Huang PJ, Chou YL, Chou CK. Gait analysis after ankle arthrodesis. Gait Posture 2000;11:54–61. [22] Rao S, Saltzman C, Yack HJ. Segmental foot mobility in individuals with and without diabetes and neuropathy. Clin Biomech 2007;22:464–71. [23] Tulchin K, Orendurff M, Adolfsen S, Karol L. Effects of speed on multi-segment foot kinematics. J Appl Biomech, in press. [24] Davis III RB, Ounpuu S, Tyburski D, Gage JR. A gait analysis data collection and reduction technique. Hum Mov Sci 1991;10(5):575–87. [25] Tulchin K, Haideri N. Utility of a multi-segment foot model. In: Proceedings of the 8th international symposium on the 3D analysis of human movement; 2004. p. 25–8. [26] Woltring HJ. A FORTAN package for generalized cross-validatory spline smoothing and differentiation. Adv Eng Software 1986;8(2):104–13. [27] .Perry J. Gait analysis; normal and pathological function. Thorofare, NJ: Slack; 1992 [28] Wervey R, Schwartz M. A new device for improving foot marker alignment. In: Proceedings of the 8th annual GCMAS meeting; 2003. [29] Matsas A, Taylor N, McBurney H. Knee joint kinematics from familiarized treadmill walking can be generalized to overground walking in young unimpaired subjects. Gait Posture 2000;11(1):72–9.
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
A comparison of multi-segment foot kinematics during level overground and treadmill walking Kirsten Tulchin *, Michael Orendurff, Lori Karol Texas Scottish Rite Hospital for Children, Movement Science Laboratory, 2222 Welborn Street, Dallas, TX 75219, United States
A R T I C L E I N F O
A B S T R A C T
Article history: Received 26 February 2009 Received in revised form 21 August 2009 Accepted 12 September 2009
Previous work comparing treadmill and overground walking has focused on lower extremity motion and kinetics, with few identified differences. However, a comparison of multi-segment foot kinematics between these conditions has not been previously reported. Sagittal ankle motion using a single rigid body foot model and three-dimensional hindfoot and forefoot kinematics were compared during barefoot, level, overground walking at a self-selected speed and treadmill walking at a similar speed for 20 healthy adults. Slight differences were seen in ankle plantarflexion and hindfoot plantarflexion during first rocker, as well as peak forefoot eversion and abduction, however all changes were less than 38, and most were within the day-to-day repeatability. These results indicate that foot mechanics as determined using a multi-segment foot model were similar between overground and treadmill walking at similar speeds in healthy adults. Treadmill protocols may provide a controlled method to analyze a patient’s ability to adapt to walking at different speeds and surface slopes, which are encountered often during ambulation of daily living. ß 2009 Elsevier B.V. All rights reserved.
Keywords: Gait Multi-segment foot kinematics Treadmill Overground walking
1. Introduction Instrumented lower extremity gait analysis in the clinical setting typically focuses on overground walking at a self-selected speed. Ambulation of daily living in the community however often requires more demanding tasks such as speed variation, turning and negotiating uneven terrain [1,2]. The use of a treadmill provides clinicians a means to design controlled protocols to evaluate a patient’s ability to adapt to changes in speed and surface slope. The advantages of gait analysis using a treadmill protocol also include the ability to collect multiple, consecutive strides within a small capture volume, which allows data collection to occur within a shorter period of time and with less variability in speed between trials [3]. Previous work comparing overground level walking and treadmill walking has focused on differences in lower extremity kinematics and kinetics. Alton et al. [4] found greater hip range of motion, greater maximal hip flexion angle and higher cadence during treadmill walking. They also observed a slightly decreased stance time compared to overground walking. Similar findings of a higher cadence during treadmill walking was also reported by Murray et al. [5]; however these results were not duplicated by
* Corresponding author. Tel.: +1 214 559 7580. E-mail address: [email protected] (K. Tulchin). URL: http://www.tsrhc.org 0966-6362/$ – see front matter ß 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2009.09.007
Riley et al. [6]. In their study, Riley et al. reported on both kinematic and kinetic changes using a forceplate instrumented treadmill. In evaluating the minima and maxima values of each kinematic variable, they found that differences between conditions were less than 38. In addition all differences seen were within the coefficients of repeatability of overground walking tests, with similar findings seen in the kinetic variables. Of particular interest to this study, they found no significant differences in maximum ankle dorsiflexion or plantarflexion motion, moments or power [6]. The use of multi-segment foot models to evaluate the pathological foot has expanded greatly over the last decade. These models provide a means to assess foot mechanics in more detail compared to the traditional lower extremity single rigid body foot model. By dividing the foot into multiple ‘rigid’ segments, three-dimensional intersegmental motion can be assessed for the hindfoot and forefoot, for example. Benedetti et al. [7] stated that approximately 16 studies utilizing multi-segment foot models were reported in literature between 1999 and 2008, on topics ranging from posterior tibialis dysfunction to talipes equinovarus deformity. Further analysis of this literature indicates that 13 of these 16 studies were published between 2004 and 2008, illustrating the emerging interest in the use of multi-segment foot models as an outcome tool in pathological gait evaluations. Fifteen of the 16 studies were conducted during overground ambulation at subjects’ self-selected walking speed [8–21]. Rao et al. [22] conducted testing with all subjects walking overground at a speedof0.89 m/s, and only three reported monitoring speed to within 5% of each subject’s self-selected velocity [11,14,19].
K. Tulchin et al. / Gait & Posture 31 (2010) 104–108
105
Tulchin et al. [23] have recently reported that walking speed can significantly affect sagittal plane multi-segment foot kinematics in healthy adults. Changes in inter-segment foot kinematics were similar to that seen at the ankle, with a shift towards plantarflexion and reduced stance duration as a percent of the gait cycle at faster speeds. The treadmill can provide a structured protocol to assess patients with pathological foot conditions at various speeds and inclines, which these patients may be subjected to during ambulation of daily living. However, it is important to first determine if there are any underlying differences in foot biomechanics, as measured using a multi-segment foot model, between overground and treadmill walking at similar speeds. The hypothesis of the current study was there would be no differences seen in multi-segment foot kinematics during level overground walking at a self-selected speed and level treadmill ambulation at a similar speed in a group of healthy adult subjects. 2. Methods Twenty adults, six males and 14 females, underwent gait analysis with IRB informed consent (Table 1). Subjects were excluded if they had a history of lower extremity surgery, or recent lower extremity injury within the last 6 months. Subjects were instrumented with a modified Helen Hayes marker set [24], as well as a multi-segment foot marker set bilaterally [23,25]. For the foot model, two of the Helen Hayes markers placed on the foot (toe and heel) were used in conjunction with five additional markers to model the hindfoot and forefoot. Marker trajectories were collected using a 12 camera VICON Mx motion capture system operating at 120 Hz (VICON, Denver, CO). All data were filtered with a smoothing spline (Woltring) routine using a mean square error of 10 [26]. Subjects performed level, overground barefoot walking at a self-selected speed. Multiple trials were collected for each leg, however a single side was randomly chosen for analysis for each subject, which resulted in equal distribution between right and left legs across all subjects. Subjects were then asked to complete a treadmill protocol, walking at speeds beginning at 0.89 m/s and increasing in increments of 0.23 m/s up to 1.79 m/ s. The speed which most closely matched their self-selected walking speed during the overground condition was used for comparison. A single representative trial was chosen for each condition for comparison. Maximum and minimum values of sagittal plane ankle and three-dimensional hindfoot and forefoot motion were determined for each subject. Sagittal plane ankle motion as defined using a single rigid body foot model (Plug-In Gait, VICON) was determined. Hindfoot and forefoot angles were calculated using a custom written multi-segment foot kinematic model (Bodybuilder for Biomechanics, VICON). Sagittal plane motion of ankle and hindfoot motion were examined specifically during each of the three ankle rockers: (1) plantarflexion during weight acceptance, (2) dorsiflexion during single limb stance, and (3) plantarflexion at terminal stance [27]. The mean differences between conditions for all variables were determined for each individual subject, and one sample t-tests were conducted for the group with a hypothesized mean of zero, indicating that no differences between conditions were expected. In addition, mean differences between overground and treadmill walking were compared to the dayto-day repeatability of each variable based on previous work comparing two sessions of data collected 2 weeks apart on 10 healthy adults [25]. 2.1. Multi-segment foot model The multi-segment foot model employed for this work was a three segment model of the shank, hindfoot, and forefoot [23,25]. Marker placement on the foot is shown in Fig. 1. A marker was also placed on the medial malleoli during the static trial. For the shank segment: the origin was located at the midpoint of the lateral and medial malleoli (ankle joint center); the X-axis was defined as the line joining the lateral and medial malleoli; the Y-axis was perpendicular to a line joining the knee and ankle joint centers, and the X-axis; and the long axis, Z, was therefore defined as being orthogonal to the X and Y axes. The shank segment was offset in the coronal plane to be parallel to the plane of the floor during the static trial, thus eliminating offsets in hindfoot motion based on bimalleolar axis variations. Individual markers were placed on the posterior, medial and lateral calcaneus to define the hindfoot segment The foot was marked non-weight bearing such that, with the subject lying prone, the clinician palpated the medial and lateral borders of the Table 1 Subject demographics, including age, height and weight.
calcaneus and drew a midline bisection along the posterior calcaneus. During quiet standing, a custom hindfoot alignment device with a cross-hair laser, based on the novel device described by Wervey and Schwartz [28], was used to ensure proper placement of the medial and lateral calcaneus markers in the coronal plane, so that they were perpendicular to the posterior midline of the calcaneus. This method was used to assure accurate measurement of hindfoot inversion/eversion. The third marker was placed along the midline of the posterior calcaneus. Calipers were then used to ensure that the distances from the medial and lateral calcaneal markers to the posterior calcaneal marker were equal, decreasing erroneous hindfoot rotation. For the hindfoot: the origin was the midpoint of the lateral and medial calcaneal markers (mid-calcaneus); The X-axis was defined as the line joining the medial and lateral calcaneal markers; the Z-axis was defined perpendicular to the line joining the posterior calcaneus marker and the mid-calcaneus, and the X-axis; and the Y-axis was defined as orthogonal to the X and Z axes. The hindfoot segment was offset in the sagittal plane to be parallel to the plane of the floor when the foot was plantigrade. Palpation was used to place markers on the first and fifth metatarsal heads and the fifth metatarsal base. For the forefoot: the origin was defined as a point at the perpendicular bisector of the fifth metatarsal base to a line joining the second/third metatarsal heads to the posterior calcaneus (long axis of the foot); the X axis joined the heads of the first and fifth metatarsals; the Z axis was perpendicular to a line joining the fifth metatarsal base and the fifth metatarsal head, and the X axis; and the Y axis was orthogonal to the X and Z axes. Relative inter-segmental motion of the following segments were examined: hindfoot relative to the shank (termed hindfoot motion) and forefoot relative to the hindfoot (forefoot motion). Hindfoot motion represented motion at the ankle mortise and the subtalar joint. Forefoot motion included the Chopart and Lisfranc joints. All angles were described by Euler rotations about the proximal segment axes, using a XYZ (sagittal, coronal, transverse) rotation order.
3. Results
Subjects
N
Age (years)
Height (cm)
Weight (kg)
Female Male All subjects
14 6 20
25.0 4.3 25.3 6.2 25.1 4.8
166.8 4.3 182.0 7.2 171.4 8.8
62.3 10.0 84.9 16.6 69.1 15.9
Mean (standard deviation).
Fig. 1. Skin marker placement for the TSRH multi-segment foot model. Markers on the second/third metatarsal, lateral malleolus and posterior calcaneus, shown in gray, are used for the single rigid body foot model, as well as the multi-segment foot model. Markers on the heads of the first and fifth metatarsals, the fifth metatarsal base, medial and lateral calcaneus markers, shown in black, in addition to the posterior calcaneus marker, are used to define the hindfoot and forefoot segments in the multi-segment foot model.
The 1.12 m/s (2.5 mph) treadmill speed was selected for 11 subjects (average overground speed 1.17 0.05 m/s) and 1.34 m/s (3.0 mph) was used for the nine remaining subjects (average overground speed 1.31 0.07 m/s). Mean differences between
K. Tulchin et al. / Gait & Posture 31 (2010) 104–108
106
Table 2 Mean differences between overground and treadmill walking of temporal parameters, ankle motion using the single rigid body foot model and hindfoot and forefoot motion using the multi-segment foot model. Variable
Mean difference
Std dev.
t-test
Day-to-day repeatability
Temporal/spatial Cadence (steps/min) Stance time (%GC)
0.03 0.00
6.23 0.04
0.98 0.70
0.04 0.87
Ankle Max PF angle—first rocker (8) Max DF angle—second rocker (8) Max PF angle—third rocker (8)
1.5 1.2 3.0
2.2 3.4 7.5
0.006** 0.124 0.09
0.8 1.0 2.2
Hindfoot Max PF angle—first rocker (8) Max DF angle—second rocker (8) Max PF angle—third rocker (8) Total sagittal ROM (8) Max inversion angle (8) Max eversion angle (8) Total coronal ROM (8) Max internal rotation angle (8) Max external rotation angle (8) Total transverse ROM (8)
1.3 0.4 1.3 0.8 0.7 0.1 0.7 0.7 0.6 1.3
2.5 2.3 6.8 5.5 2.2 1.5 2.7 2.6 2.8 2.9
0.027** 0.43 0.42 0.511 0.20 0.80 0.24 0.20 0.39 0.06
3.1 3.1 2.9 1.9 1.3 2.7 2.8 4.5 3.7 1.8
Forefoot Max DF angle—second rocker (8) Max PF angle—third rocker (8) Total sagittal ROM (8) Max inversion (8) Max eversion (8) Total coronal ROM (8) Max adduction (8) Max abduction (8) Total transverse ROM (8)
0.3 0.3 0.02 0.04 1.5 1.4 0.9 1.2 2.2
2.3 2.2 3.4 2.0 2.1 3.1 1.9 1.6 2.1
0.60 0.55 0.98 0.94 0.005** 0.05 0.039** 0.003** <0.001**
3.3 3.7 2.7 2.1 2.5 1.6 2.9 3.0 1.5
Positive mean differences indicate that overground values were greater than treadmill values. The day-to-day repeatability based on previous work [25] examining the multisegment foot model is shown for comparison. The variables that had statistically significant differences between conditions and were greater than the day-to-day repeatability errors are highlighted in bold italics. %GC = percentage gait cycle, DF: dorsiflexion, PF: plantarflexion, ROM: range of motion.
conditions of temporal and spatial parameters as well as ankle and multi-segment kinematic foot variables are shown in Table 2, and kinematic curves for both conditions are shown in Fig. 2. There were no statistical differences in cadence or stance time (as a percentage of the gait cycle) between overground and treadmill walking. Using the single rigid body foot model, there was a statistically significant difference in the maximal plantarflexion angle during first rocker between conditions (p = 0.006). However, the average difference between conditions across subjects was only 1.5 2.28. Similar results were seen in the hindfoot, with a mean difference between conditions during first rocker of 1.3 2.58. At the forefoot, statistical differences were noted in the coronal and transverse planes relative to the hindfoot, with slightly decreased eversion (1.5 2.18), slightly reduced maximum adduction (0.9 1.98), maximal abduction (1.2 1.68) and decreased overall range of motion in the transverse plane (2.2 2.18) during treadmill walking compared to the overground condition. These differences averaged less than 2.58 across all subjects. In addition, all differences between conditions were less than the day-to-day repeatability of these measures except for forefoot transverse plane range of motion, which averaged a mean difference of 2.28 and had a repeatability of 1.58. 4. Discussion The current study found several statistically significant differences between overground and treadmill walking at similar speeds in healthy adults. However, most of these differences were less than the day-to-day repeatability of the model, and therefore were not considered to be indicative of clinically significant variations in foot mechanics. Treadmill walking may provide clinicians and researchers the ability to develop more functional protocols that involve controlling and varying the speed and
incline during testing. These modulations in surface slope and speed compared to typical self-selected level walking might help elicit functional deficits in patients with foot pathology which would otherwise not be seen during typical gait analysis. Previous studies [4,5] have reported increased cadence during treadmill walking versus overground ambulation, however this was not found in the current study. While a long accommodation period to treadmill walking following overground walking was not given, all subjects had self-reported experience walking on a treadmill. The results of the current study concur with the conclusions of Riley et al. [5] that the treadmill did not appear to cause differences in temporal parameters during gait. With the focus of the current study on multi-segment foot kinematics, previous studies examining ankle motion, using a single rigid foot segment approach, during overground and treadmill walking were of particular interest [4–6]. Murray et al. [5] reported significant increased dorsiflexion during overground walking compared to treadmill walking at slow and free walking speeds, although these differences were estimated to be only 38. However, other studies have reported no significant differences in peak ankle dorsiflexion or plantarflexion between treadmill and overground conditions [4,6]. The current study evaluated ankle motion in slightly more detail—examining motion specifically within each of the three rockers. Peak ankle plantarflexion during first rocker was found to be significant less during treadmill walking, however this change was only 1.58. Visual examination of average curves indicate a slight shift towards dorsiflexion during treadmill walking, which is similar to the supplemental data reported by Riley et al. [6], particularly during stance phase. However, peak dorsiflexion during second rocker and peak plantarflexion during third rocker were not shown to been significantly different between conditions in either study.
K. Tulchin et al. / Gait & Posture 31 (2010) 104–108
107
Fig. 2. Kinematic results for N = 20 healthy adults. Ankle dorsiflexion/plantarfleixon motion measured using the single rigid body foot model; hindfoot relative to tibia, and forefoot relative to hindfoot motion based on the multi-segment foot model. Mean std. dev.: gray bar represents overground condition. Black solid dashed lines represent the treadmill condition.
The aim of the current study was to expand the analysis of the foot by the utilization of a multi-segment foot model. Analysis indicated that sagittal plane forefoot motion was not different between overground and treadmill ambulation, however peak hindfoot plantarflexion during first rocker was slightly reduced during treadmill walking. One advantage of using a multisegment foot model was the ability to assess coronal and transverse plane motion. Overall there were no significant variations in hindfoot motion between conditions in the coronal or transverse planes. The forefoot however did show a decrease in maximal eversion (p = 0.005) which lead to a slight decrease in coronal plane ROM. Similar changes were also demonstrated in forefoot rotation, with an decrease in maximal abduction seen during treadmill walking. Most significant changes reported in the current study were less 2.58 and within the day-to-day repeatability of the model in healthy adults, and were not deemed to be of clinical significance. There are several factors which may influence the results of the current study. First, all walking trials were conducted during barefoot ambulation. While subjects self-reported that they were experienced treadmill walkers, none had experienced barefoot ambulation on a treadmill. There were no reported problems during testing, however it is unclear whether possible increased foot-to-belt sensations had an impact on foot mechanics. As previously stated there were no accommodation periods given to the subjects prior to treadmill data collection. As part of a larger protocol, multiple walking trials were recorded for at least one or two speeds prior to the speed selected for the current analysis. This
would allow subjects to have at least a 1–2 min accommodation to treadmill walking prior to the data used for this study. Previous work by Matsas et al. [29] indicated that 6 min of walking was required for accommodation to a treadmill. In addition, overground self-selected walking speed was not directly matched for each subject during treadmill ambulation. Subjects walked at five speeds ranging from 0.89 to 1.79 m/s, in increments of 0.23 m/s, therefore the maximum possible difference between overground and treadmill walking was 0.12 m/s. Lastly, during treadmill walking, all subjects were instructed to rest their hands on the side treadmill bars. It is unclear what effect this may have had foot mechanics and propulsion mechanisms. In summary, the current study found multi-segment foot kinematics during overground and treadmill walking to be similar in healthy adults, with only a slight decrease (2.28) in forefoot transverse plane ROM during treadmill walking. In addition ankle plantarflexion using a single rigid body foot model was shown to be significantly decreased by 1.58 during treadmill walking. Further analysis is needed to evaluate these walking conditions in patients with foot pathology. However, a treadmill can provide a means to assess a patient’s ability to accommodate to increased speed or incline. These modulations in surface slope and speed may provide clinicians a more detailed indication of a subject’s functional deficits during community ambulation. Conflict of interest statement The authors report no conflicts of interest.
108
K. Tulchin et al. / Gait & Posture 31 (2010) 104–108
References [1] Orendurff MS, Schoen JA, Bernatz GC, Crass SM, Klute GK. How humans walk: bout duration, steps per bout and rest duration. J Rehabil Res Dev 2009;45:1077– 90. [2] Glaister BC, Bernatz GC, Klute GK, Orendurff MS. Video task analysis of turning during activities of daily living. Gait Posture 2007;25:289–94. [3] Dingwll J, Cusumano J, Cavanaugh P, Sternad D. Local dynamic stability versus kinematic variability of continuous overground and treadmill walking. J Biomech Eng 2001;123(1):27. [4] Alton F, Baldey L, Caplan S, Morrissey MC. A kinematic comparison of overground and treadmill walking. J Biomech 1998;13:434–40. [5] Murray MP, Spurr GB, Sepic SB, Gardner GM, Mollinger LA. Treadmill vs. floor walking: kinematics, electromyogram, and heart rate. J Appl Physiol 1985;59(1): 87–91. [6] Riley PO, Paolini G, Della Croce U, Paylo KW, Kerrigan DC. A kinematic and kinetic comparison of overground and treadmill walking in healthy subjects. Gait Posture 2007;26(1):17–24. [7] Benedetti MG, Leardini A, Bianchi L, Berti L, Giannini S. Comparisons of outputs of different models for foot kinematics. In: Proceedings of the international foot and ankle biomechanics congress; 2008. p. 91. [8] Canseco K, Long J, Marks R, Khazzam M, Harris GF. Quantitative characterization of gait kinematics in patients with hallux rigidus using the Milwaukee Foot Model. J Ortho Res 2008;26:419–27. [9] Flemister AS, Neville CG, Houck J. The relationship between ankle, hindfoot, and forefoot position and posterior tibial muscle excursion. Foot Ankle Int 2007;28(4):448–55. [10] Huang UC, Harbst K, Kotajarvi B, Hansen D, Koff M, Kitaoka H, et al. Effects of ankle-foot orthoses on ankle and foot kinematics in patient with ankle osteoarthritis. Arch Phys Med Rehabil 2006;87:710–6. [11] Hunt AE, Smith RM. Mechanics and control of the flat versus normal foot during the stance phase of walking. Clin Biomech 2004;19:391–7. [12] Khazzam M, Long JT, Marks RM, Harris GF. Kinematic changes of the foot and ankle in patients with systemic rheumatoid arthritis and forefoot deformity. J Ortho Res 2007;25:319–29. [13] Levinger P, Gilleard W. Tibia and rearfoot motion and ground reaction forces in subjects with patellofemoral pain syndrome during walking. Gait Posture 2007;25(1):2–8.
[14] Neville C, Flemister A, Tome J, Houck J. Comparison of changes in posterior tibialis muscle length between subjects with posterior tibial tendon dysfunction and healthy controls during walking. JOSPT 2007;37(11):661–9. [15] Ness ME, Long J, Marks R, Harris G. Foot and ankle kinematics in patients with posterior tibial tendon dysfunction. Gait Posture 2008;27:331–9. [16] Rattanaprasert U, Smith R, Sullivan M, Gilleard W. Three-dimensional kinematics of the forefoot, rearfoot and leg with the function of tibialis posterior in comparison with normals during stance phase of walking. Clin Biomech 1999;14:14–23. [17] Ringleb SI, Karvos SJ, Kotajarvi BR, Hansen DK, Kitaoka HB, Kaufman KR. Changes in gait associated with acute stage II posterior tibial tendon dysfunction. Gait Posture 2007;25:555–64. [18] Theologis TN, Harrington ME, Thompson N, Benson KD. Dynamic foot movement in children treated for congenital talipes equinovarus. JBJS [Br] 2003;85-B:572–7. [19] Tome J, Nawoczenski DA, Flemister A, Houck J. Comparison of foot kinematics between subjects with posterior tibialis tendon dysfunction and healthy controls. JOSPT 2006;36(9):635–44. [20] Woodburn J, Nelson KM, Siegel KL, Kepple TM, Gerber LH. Multisegment foot motion during gait: proof of concept in rheumatoid arthritis. J Rheumat 2004;31(10):1918–27. [21] Wu WL, Su FC, Cheng YM, Huang PJ, Chou YL, Chou CK. Gait analysis after ankle arthrodesis. Gait Posture 2000;11:54–61. [22] Rao S, Saltzman C, Yack HJ. Segmental foot mobility in individuals with and without diabetes and neuropathy. Clin Biomech 2007;22:464–71. [23] Tulchin K, Orendurff M, Adolfsen S, Karol L. Effects of speed on multi-segment foot kinematics. J Appl Biomech, in press. [24] Davis III RB, Ounpuu S, Tyburski D, Gage JR. A gait analysis data collection and reduction technique. Hum Mov Sci 1991;10(5):575–87. [25] Tulchin K, Haideri N. Utility of a multi-segment foot model. In: Proceedings of the 8th international symposium on the 3D analysis of human movement; 2004. p. 25–8. [26] Woltring HJ. A FORTAN package for generalized cross-validatory spline smoothing and differentiation. Adv Eng Software 1986;8(2):104–13. [27] .Perry J. Gait analysis; normal and pathological function. Thorofare, NJ: Slack; 1992 [28] Wervey R, Schwartz M. A new device for improving foot marker alignment. In: Proceedings of the 8th annual GCMAS meeting; 2003. [29] Matsas A, Taylor N, McBurney H. Knee joint kinematics from familiarized treadmill walking can be generalized to overground walking in young unimpaired subjects. Gait Posture 2000;11(1):72–9.