- Email: [email protected]
Sprint kinematics of athletes with lower-limb amputations
501
Sprint Kinematics of Athletes With Lower-Limb
Amputations
John G. Buckley, MSc ABSTRACT. Buckley JG. Sprint kinematics of athletes with lower-limb amputations. Arch Phys Med Rehabil 1999;80:501508.
Objective: To determine and compare the kinematics of the sound and prosthetic limb in five of the world’s best unilateral amputee sprinters. Subjects: Five men, all unilateral lower-limb amputee (one transfemoral, four transtibial) athletes. The individual with transfemoral amputation used a Endolite Hi-activity prosthesis incorporating a CaTech hydraulic swing and stance control unit, a Flex-Foot Modular III, and an ischial containment total contact socket. Those with transtibial amputations used prostheses incorporating a Flex-Foot Modular III and patellar tendonbearing socket, with silicone sheath liner (Iceross) and lanyard suspension. Design: Case series. Subjects were videotaped sprinting through a performance area. Sagittal plane lower-limb kinematics derived from manual digitization (at 50Hz) of the video were determined for three sprint trials of the prosthetic and sound limb. Hip, knee, and ankle kinematics of each subject’s sound and prosthetic limb were compared to highlight kinematic alterations resulting from the use of individual prostheses. Comparisons were also made with mean data from five able-bodied men who had similar sprinting ability. Results: Sound limb hip and knee kinematics in all subjects with amputation were comparable to those in able-bodied subjects. The prosthetic knee of the transfemoral amputee athlete fully extended early in swing and remained so through stance. In the transtibial amputee athletes, as in able-bodied subjects, a pattern of stance flexion-extension was evident for both limbs. During stance, prosthetic ankle angles of the transtibial amputee subjects were similar to those of the sound side and those of able-bodied subjects. Conclusion: Prosthetic limb kinematics in transtibial amputee subjects were similar to those for the sound limb, and individuals achieved an “up-on-the-toes” gait typical of ablebodied sprinting. Kinematics for the prosthetic limb of the transfemoral amputee subject were more typical of those seen for walking. This resulted in a sprinting gait with large kinematic asymmetries between contralateral limbs. 0 1999 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation
From the Department of Exercise and Sport Science, The Manchester Metropolitan University, Stoke-on-Trent, United Kingdom. Submitted for publication June 1, 1998. Accepted in revised form October 1, 1998. Supported in part by Chas. A. Blatchford and Sons Ltd., Basingstoke, Hampshire, United Kingdom. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author or upon any organization with which the author is associated. Reprint requests to J.G. Buckley, The Manchester Metropolitan University. Department of Exercise and Sport Science, Hassall Road; Stoke-on-Trent ST7 ZHL, United Kingdom. B 1999 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation 0003.9993/99/8005-5050$3.00/O
A
FTER LOWER LIMB amputation, the ultimate rehabilitation goal for many patients is to attain a full and active lifestyle. For some, this means being involved in competitive sport. Because many sports require the ability to run or sprint, designing a prosthesis that will accommodate a near-to-normal leg action is a priority in prosthetic evolution. From studies of able-bodied athletes, it is apparent that the biomechanics of running and sprinting are noticeably different. The duration of stance, as a percentage of stride time, decreases from 31% for running to 22% for sprinting.t In sprinting the center of mass of the body is lowered by increasing hip, knee, and ankle flexion, and during the stance phase of sprinting, only knee flexion occurs, whereas both flexion and extension occur during running. l-3 Whether lower limb amputee persons achieve kinematics in sprinting comparable to those achieved by able-bodied athletes is the focus of this article. This work is part of an ongoing study4z5attempting to quantify the biomechanical parameters and prosthetic requirements of sprinting by amputee runners. Previously, the only amputee athletic gait that has been analyzed biomechanically in detail is transtibial amputee running at a moderate pace.6 i3 Findings suggest that for individuals with this level of amputation, running is characterized by several compensatory mechanisms. During stance there is prolonged and excessive knee extension with only a small change in the thigh angle of the prosthetic limb.7 Two to three times more total work is done by the hip of the amputee runner’s prosthetic limb than by the intact limb.9:13 Typically, there is an absence of an impact ground reaction force peak for the prosthetic limb. t2 During swing the total work done by the prosthetic limb is similar to that in able-bodied subjects9 In contrast, the sound limb exhibits a 69% increase in the total work done, with an accompanying increase in the energy transferred from this limb to the trunk.11 Because the number of persons with amputations who participate in sporting activities has been increasing, and because many of these activities require the ability to sprint, a similar understanding of the biomechanical adaptations adopted during sprinting is required. Not only will this knowledge allow specific rehabilitation and training strategies to be developed, but it also will allow developments in prosthetic hardware (made to encourage and facilitate a more active lifestyle) to be objectively evaluated. This study was conducted to determine how amputee athletes use their prostheses during maximal sprinting. The specific aim was to measure and compare the kinematics of the sound and prosthetic limbs in five of the world’s best unilateral amputee sprinters. In an attempt to highlight kinematic alterations resulting from the use of an individual prosthesis, the kinematics of the sound and prosthetic limb of each amputee athlete were also compared with those of a group of able-bodied athletes who had similar personal best 100m times. METHODS Subject Data Five men with unilateral lower-limb amputations (one transfemoral and four transtibial) and five able-bodied men volunteered to participate. The five amputee athletes (mean age, 25.2 i 5.lyr.s; mean body mass, 70.4 i 5.9kg) all undertook Arch
Phys
Med
Rehabil
Vol 80, May
1999
502
AMPUTEE
SPRINTING,
athletic training on a weekly basis. All were experienced competitors with personal bests for the 100m sprint ranging from 11.7sec to 14.0sec (transtibial) and 15. lsec (transfemoral). The five able-bodied athletes (mean age, 18.2 +- 2.4yrs; mean body mass, 75.0 5 7.7kg) were members of a local athletic club and had personal bests for the 1OOmsprint ranging from 11.3sec to 12.6sec. All subjects were informed of the purpose of the study, and written informed consent was obtained from each subject before testing. The study met with local bioethics committee approval. The transfemoral amputee patient (subject 5) used an Endolite Hi activity prosthesisa incorporating a CaTech hydraulic swing and stance control unit,a a Flex-Foot Modular III,b and an ischial containment total contact socket. This limb was also worn for “everyday” use. All transtibial amputee athletes used prostheses incorporating a Flex-Foot Modular IIIb and patellartendon-bearing suction socket, with silicone sheath 1inerCand lanyard suspension. Subject 1 had a dedicated sprint prosthesis, which was adjusted so its length was the same as the sound leg when the wearer stood on tip-toe. The other three subjects used prostheses that had been fitted and aligned for “everyday” use; however, the prosthesis of subject 3 was modified by anteriorally tilting the shin by 10” at the socket mount to place the foot into a more plan&r-flexed position. The residual knee of subject 4 was unable to flex beyond 90”, because it had been pinned after the trauma that led to the transtibial amputation.
Buckley
joints, hips, knees, ankles, and distal end of feet). Markers were placed on the prosthetic foot at the same height as the lateral malleolus of the intact limb during normal standing and on the prosthetic socket, at the position corresponding to the underlying knee center (transtibial) or hip center (transfemoral) (fig 1 inset). All other points were located by assuming joint centers and segment endpoints to lie on the midlines of body segments.15 The raw two-dimensional marker coordinate data were then smoothed and differentiated using cross-validatedquintic splines. The smoothed data were used to define a 14 linked-segment model (fig 1) of the performer in the sagittal plane, from which all kinematic parameters were determined. Knee and ankle angles were measured as the absolute angles between the thigh and shank segment and shank and foot segment, respectively; thus, a fully extended knee, for example, would have an angle of 180”. The ankle angle at the instance when the shank was fully vertical during stance in each subject was used as a zero reference; thus, positive angles indicated plantarflexion and negative angles dorsiflexion. The thigh angle was measured as the included angle of the thigh segment to the vertical; positive angles thus indicated hip flexion and negative angles hip extension. Sprint speed was the average speed of the subject’s center of mass during a complete stride. Data Analysis Time series data were normalized to a percentage of each subject’s cycle (stride) time (the period between consecutive
Data Collection Subjects were given time to warm up and to become familiar with sprinting through a “performance” area on an outdoors cinder track. To ensure that subjects could achieve their maximum sprinting speed, starting points were set approximately 20m from the performance area; these were adjusted for each subject to ensure that stance of one limb occurred in approximately the middle of the performance area without any overstriding or understriding. A 20m run-off allowed subjects to safely reduce their speed. Subjects were videotaped using two Panasonic F15 video cameras,dmounted on tripods approximately 15m from, perpendicular to, and on either side of the performance area. Amputee subjects performed 8 to 10 sprints until three trials in which the prosthetic foot and three trials in which the sound foot was planted in the middle of the performance area were recorded. These were termed successful trials. Because asymmetry between contralateral sides was not investigated in the ablebodied sprinters, they were required only to complete three successful trials in which either their left or right foot was planted in the middle of the performance area. Regulating foot placement in this manner meant a full stride occurred within the performance area and allowed the size of the camera image to be maximized. Data Processing All data processing was carried out using Bartlett’s biomechanical analysis software. I4 Video recordings from the camera on the side closest to the limb landing in the middle of the performance area were analyzed using an Archimedes 440 microcomputef fitted with a Moving-Image high resolution (640 X 512) color digitizer.f From just before toe-off to just after foot-strike of the ipsilateral limb, each field (50Hz) was manually digitized by locating 18 points on the body (apex of head, C7, shoulders, elbows, wrists, third metacarpophalangeal Arch
Phys
Med
Rehabil
Vol 80, May
1999
4
thigh angle
centreof mass
/
A,
ankle
angle .P=J
Fig 1. Linked-segment model of performer showing how limb angles were defined. Inset shows marker placement used on prosthetic limb; all other points were located by assuming joint centers and segment endpoints to lie on the midlines of body segments.14 The ankle angle at the instance when the shank was fully vertical durina stance was used as a zero reference; thus, oositive anales
AMPUTEE
SPRINTING,
foot-strikes of the ipsilateral limb). The mean of the three successful trials was then determined. For the able-bodied subjects individual means were then added to give a group mean. Because sprinting technique is highly individualized it is difficult to make reference to ideal or optimal technique. One could argue, therefore, that an able-bodied group mean has limited meaning; certainly this would be the case if the data were used to represent a control group from which to make statistical comparison. The kinematics for the able-bodied group were determined in an attempt to highlight, qualitatively, the kinematic alterations resulting from the use of individual prostheses. Without such a comparison it would be difficult to determine whether sound limb kinematics were typical of those seen in able-bodied sprinting. Findings for the able-bodied subjects demonstrated low intersubject variability, suggesting subjects performed kinematically in similar fashion. The low intersubject variability meant the concerns regarding determining a group mean (for this particular group) were allayed. Digitizing repeatability was assessedby digitizing one trial for one subject 10 times over two separate digitizing sessions (days). After smoothing the coordinate and angular displacement data, the variability (residual mean square within subjects [S,]) in the angular displacement at the knee between the 10 trials was evaluated over the complete stride period using analysis of variance (ANOVA). The 95% agreement limitsi were then determined as 22 d/s,. To give an indication of intrasubject consistency, the alpha (stability) coeflicients17 for the three repeated trials (for each subject) were determined using ANOVA. RESULTS Error Estimates Statistical analysis (ANOVA) indicated that repeatability in determining knee angular displacements within 95% agreement limits was 23.0”. Thus kinematic differences between prosthetic and sound limb, or either limb and the able-bodied mean, were only considered of consequence if they were greater than 3”. The intrasubject consistency in angular displacements at the knee for the three trials was 2.98 in all subjects. The high consistency values likely reflect that subjects, who undertook athletic training on a regular basis, were highly skilled amputee sprinters. Data Interpretation Because of differences in prosthetic alignment, amputation level, stump length/strength, etc, it would have been inappropriate to group the amputee subjects to make statistical comparisons with the able-bodied group. The results are thus presented as a series of individual case reports. In each case, sound and prosthetic limb kinematics were compared with each other and with the able-bodied mean. It is evident from the ankle, knee, and thigh angular displacement data for each subject (figs 2 through 6, parts a through c in each fig) that there were considerable kinematic differences between prosthetic and sound limbs and between both limbs and the able-bodied mean. It is apparent that these differences were greatest for the ankle and least for the thigh and were, with the exception of the transfemoral amputee athlete’s prosthetic limb, related to the magnitude of angular displacement, with the temporal occurrence of displacement changes remaining constant. The ankle angle graphs show there were small angular displacement changes during swing. This apparent motion was likely the result of out-of-plane movement of the foot during swing, appearing to the camera as dorsiflex-
Buckley
%Of Stride (foot-strike to foot-strike, b) Knee angular displacement
a) Ankle angular displacement
;o
0 20 40 60 thighangle,cieg,
d) Thigh-knee angle-angle diagram
Fig 2. Subject 1 (transtibial): (a) ankle, (b) knee, and (c) thigh graphs and (d) thigh-knee angle-angle diagram, for prosthetic (dotted line) and sound limb (solid bold line), with able-bodied (solid thin line). 0, toe-off; 0, foot-strike.
angle limb mean
ion or plantarflexion. Because the limb is more in line with the plane of progression and thus perpendicular to the camera during stance, this problem was considered inconsequential to data (presented) for the stance phase. The results were also presented as thigh-knee angle-angle diagrams (figs 2 through 6, part d in each fig). Because there is no requirement to normalize data to percentage of cycle time, such diagrams allow rapid assessmentof limb function, albeit in
%Of Stride (foot-strike to ‘c.ot-*trike)
%o‘stride (foot-strike to foot-strike, b) Knee angular displacement
a) Ankle angular displacement
c) Thigh angular displacement
d) Thigh-knee angle-angle diagram
Fig 3. Subject 2 (transtibial): (a) ankle, graphs and (d) thigh-knee angle-angle (dotted line) and sound limb (solid bold (solid thin line). 0, toe-off; 0, foot-strike.
Arch
Phys
(b) knee, and (c) thigh diagram, for prosthetic line), with able-bodied
Med
Rehabil
Vol80,
May
angle limb mean
1999
AMPUTEE
a) Ankle angular displacement
b) Knee angnlar displacement
%of stride (foot-strike to foot-strike)
(b) knee, and (c) thigh diagram, for prosthetic line), with able-bodied
angle limb mean
a qualitative manner. The overall size of the diagram indicates the range of motion (the larger the diagram the greater the range), and its shape shows the coordinated action of the thigh and knee. The markedly reduced plot size for subject 4’s prosthetic limb (fig 5d) (his knee was unable to flex beyond 90’) illustrates how clearly these diagrams can highlight kinematic differences.
Kof stride (foot-strike to foot-strike) a) Ankle angular displacement
b) Knee angular displacement
c) Thiih angular displacement
Fig 5. Subject 4 (transtibial): (a) ankle, graphs and (d) thigh-knee angle-angle (dotted line) and sound limb (solid bold (solid thin line). 0, toe-off; 0, foot-strike.
Arch
Phys
Med
Rehabil
Vol 80, May
Buckley
a) Ankle angolv displacement
thigh angle (deg) d) Thigh-knee angle-angle diagram
c) Thigh angular displacement
Fig 4. Subject 3 (transtibial): (a) ankle, graphs and (d) thigh-knee angle-angle (dotted line) and sound limb (solid bold (solid thin line). 0, toe-off; 0, foot-strike.
SPRINTING,
d) Thigh-knee angle-angle diagram
(b) knee, and (c) thigh diagram, for prosthetic line), with able-bodied
1999
angle limb mean
d) Tbigh-knee angle-angle diagram
Fig 6. Subject 5 (transfemoral): (a) ankle, (b) knee, and (c) thigh angle graphs and (d) thigh-knee angle-angle diagram, for prosthetic limb (dotted line) and sound limb (solid bold line), with able-bodied mean (solid thin line). 0, toe-off; 0, foot-strike.
To objectively assess the kinematic differences between prosthetic and sound limbs, joint angular displacements and velocities of each limb at key moments in the sprint cycle were determined. Figure 7 highlights these differences for subject 2 (whose limb kinematics were closest to the midrange of all transtibial amputee subjects), and figure 8 the differences for the transfemoral amputee subject (subject 5). Transtibial Amputee Sprinting Results indicate that transtibial amputee subjects all displayed similar limb kinematics. Thus, to avoid repetition, kinematics general to all subjects are described. In addition, the kinematic differences between prosthetic and sound limbs, quantified in figure 7 for subject 2, are referred to (listed in brackets in the text) because they serve to highlight the differences found for all transtibial amputee subjects. The thigh-knee angle-angle diagrams highlight that the kinematics for the sound side compared well with those for the able-bodied subjects. The diagram for the prosthetic side indicates that this limb also achieved a kinematic pattern similar to that of the sound limb, although the amplitude of this pattern indicates reduced angular displacement at the knee. Sound limb dorsiflexion (15’) to midstance was greater than that observed on the prosthetic limb (11”). Plantarflexion at toe-off was also greater on the sound limb (5”) than on the prosthetic limb (-2”). Ankle displacements, however, on the prosthetic limb (during stance) were only marginally less than those of the sound limb. Because these displacements were achieved over a longer stance time (26% of the cycle as opposed to 22%) angular velocities were considerably reduced. In particular, the plantarflexion velocity on the prosthetic limb was less than 40% of that on sound side. Figures 2 through 5 (part a in each fig) indicate there was greater range of ankle displacement on the sound limb than in able-bodied subjects. It is apparent from part b in figures 2 through 5 that with the exception of subject 1 the prosthetic limb, when compared with the sound limb or able bodied subjects, extended at the knee
AMPUTEE
SPRINTING,
Buckley
able-body
I
I
Mid
FS
*
22%,3(,1,, TO
0%
Mid
SWING
STANCE able-body
\
adan= 6.8
l-
Fig 7. Kinematic differences between prosthetic (pros) and sound limbs of subject 2 (transtibia1 amputation). (A) Angular displacement differences at key moments in the sprint cycle; (B) the maximum angular velocities (rad/sec) observed between the key moments. FS,foot-strike; TO, toeoff; ank (or a), ankle; k, knee; h, hip; flex, flexion; ext, extension; dorsi, dorsiflexion; plantar, plantarflexion; V, velocity.
L
lOb% FS
a+=
16.6
sound=
18.4
-I
B
FS
prematurely during late swing from approximately 85% of the cycle onwards. It then either remained in extension or began to flex before foot-strike. After a period of flexion during early stance, the knee then extended to toe-off. Although the pattern of stance flexion-extension was similar for both limbs, the prosthetic knee (160”) was not extended to the same extent as the sound knee (168”) at toe-off (fig 7). Maximum knee flexion angles on the prosthetic side (58”) during swing indicate that this limb was less flexed than the sound limb (53”). Peak knee flexion velocity on the prosthetic limb (16.4rad/sec) during swing was also less than on the sound limb (18.4rad/sec). Sound limb knee kinematics were similar to those achieved by the able-bodied subjects (figs 2 through 5). Figures 2 through 5 (part c in each fig) indicate that hip kinematics for the sound and prosthetic limbs were comparable to those for able-bodied subjects and that there were only slight differences between contralateral sides. At foot-strike the prosthetic limb (34”) was more flexed than the sound limb (2S”) (fig 7). The prosthetic limb (- 14”) was less extended at toe-off
Mid
22% 3_(; ‘f(, TO
100% FS
Mid
STANCE
SWING
than the sound limb (-27”). After toe-off, both sides extended to approximately the same extent, but subsequent swing flexion was less on the prosthetic side (44”) than the sound side (52”) (fig 7). Peak thigh angular velocities for each limb were similar. Transfemoral Amputee Sprinting The shape, size, and position of the thigh-knee angle plot for the transfemoral amputee runner’s prosthetic limb (fig 6d) indicate that the kinematics for this limb were different from those on the sound side or able-bodied subjects. The increased amplitude of this plot in the vertical direction indicate the knee was overextended for a prolonged period. Its decreased width indicates a reduced range of motion of the thigh. Despite such kinematic disruption on the prosthetic side, there was no noticeable evidence, apart from a reduced range of knee flexion, of compensatory alterations in the sound limb kinematics. Data for the ankle (fig 6a) highlight a phase shift in the flexion-extension pattern on the prosthetic limb. At foot-strike there was slight plantarflexion, something not seen in any other Arch
Phys
Med
Rehabil
Vol 80, May
1999
506
AMPUTEE
SPRINTING,
Buckley
pros
-
sound
6% A
FS
Mid
TO
Mid
SWING
STANCE
able-body
pros
-
sound
0% FS
B
Mid
30% jLfo/ii TO
STANCE subject. This meant the instance of maximum dorsiflexion occurred 30% into the cycle compared with at 18% of the cycle observed for the sound limb. Displacement magnitudes, however, were comparable to those on the sound limb or ablebodied subjects. The knee angle graph (fig 6b) shows clearly the knee was fully extended early in swing (from approximately 80% of the cycle) and remained so late into stance. The sound limb knee in contrast, was flexed to 148” at foot-strike then extended to a maximum of 166” at toe-off (fig 8). This kinematic pattern was similar to that observed for able-bodied subjects. During swing there was reduced knee flexion, with angle of maximum llexion of 79” being considerably less than that on the sound side (58”) or that of able-bodied subjects (38” +- 7”) (fig 8). Sound limb thigh angles were similar to those of able-bodied subjects. In contrast, the prosthetic limb was 16” less flexed at foot-strike and, after a period of stance phase extension, began to flex before toe-off, an occurrence not seen in any of the other subjects. During swing the maximum flexion angle on the prosthetic side (39”) was less than the sound side (56”), as was Arch
Phys
Med
Rehabil
Vol 80, May
1999
100%
Mid
SWING
FS
Fig 8. Kinematic differences between prosthetic (pros) and sound (sound) limbs of the transfemoral amputee subject. (A) Angular displacement differences at key moments in the sprint cycle; (B) the maximum angular velocities (radl set) observed between the key moments. FS, foot-strike; TO, toe-off; ank (or a), ankle; k, knee; h, hip; flex, flexion; ext, extension; dorsi, dorsiflexion; plantar, plantarflexion; V, velocity.
the peak flexion angular velocity (fig 8). The stance phase peak extension velocity of 6.0 radlsec was approximately 60% of that achieved on the sound side. DISCUSSION In sprinting, data for elite able-bodied athletes indicate that the knee during stance undergoes flexion only, unlike walking and running where both llexion and extension 0ccur.l This suggests that in sprinting knee function has a limited role in propelling the body forward. In the present study both the prosthetic and sound limbs of the transtibial amputee runners, the sound limb of the transfemoral athlete, and the able-bodied subjects demonstrated knee flexion and extension during stance rather than flexion only. It may well be that stance with knee flexion only occurs at sprinting speeds in excess of the top speed (8.5mlsec) achieved by any of the subjects in the present study. The flexion-extension pattern at the knee of the prosthetic limb, when compared with able-bodied subjects, has been shown to be disrupted in transtibial amputee running with
AMPUTEE
SPRINTING,
prolonged and excessive extension during stance.’ In the present study, the transtibial amputee runners were able to achieve a similar pattern of flexion-extension during stance on the prosthetic limb as was achieved on the sound limb. The disparity in these findings can be explained by the type of prosthesis used. In the study by Enoka and colleagues,’ subjects used prostheses incorporating a SACH or Greissinger foot, whereas in the present study subjects had prostheses that incorporated a Flex-Foot Modular III. Essentially, the Modular III artificial foot is a “Y-shaped carbon fiber leaf spring that deforms during loading. This creates controlled dorsiflexion as the shank rotates forward over the planted foot and in turn allows flexion at the knee. The results presented here indicate that the pattern of dorsiflexion and plantarflexion achieved on the prosthesis was similar to that of the sound limb. The stance flexion-extension pattern at the knee of the prosthetic limb for the transtibial amputee subjects in the present study, therefore, may well be attributable to the dynamic function of the carbon fiber foot. Prosthetic knee kinematics during swing in transtibial amputee athletes were different from those for the sound limb. Apart from one subject, the knee overextended prematurely during late swing and mean peak flexion and extension angular velocities were lower than those attained on the sound limb. Given that the insertion of the hamstrings and quadriceps muscle groups are unaffected by transtibial amputation, it is unclear why this disrupted pattern was seen. It may have been due to the prosthetic socket restricting knee function or to differences in the inertial properties of the prosthetic shank and foot. In either caseprosthetic modification could ameliorate this kinematic alteration. It is evident, as might be expected, that the transfemoral amputee runner had greater kinematic asymmetry between prosthetic and sound limbs. Because the knee was fully extended early in swing, and remained so until late stance, it seemsthe prosthetic limb was used merely as a rigid support. A fully extended limb meant foot-strike was made with the heel keel. Because the keel is designed to flex on loading this explains why plantarflexion following foot-strike was seen in this subject. Such a limb position would also ensure the resultant instantaneous ground reaction force vector would stabilize the knee during the impact-absorbing phase of stance. Because the prosthesis used by the subject was also worn in everyday use and therefore had to facilitate walking, it is not surprising that such limb function, which is more typical of walking, was observed. Thigh kinematics of the sound limb for all amputees compared well with able-bodied subjects. On the prosthetic side the flexion-extension temporal pattern also matched that of the sound limb. However, maximum flexion angle and angular velocity and maximum extension angle and angular velocity during swing and stance respectively were reduced. It is likely that the prematurely and overextended knee was a contributing factor to this reduction, because an overextended knee would increase the moment of inertia of the leg about the hip. This would affect the swing kinematics and in turn could result in a less than optimum limb position at foot-strike, thus affecting stance kinematics. The results presented here cannot be viewed as generalizable to a wider population. They do, however, provide an insight into how some of the world’s best amputee athletes use their prostheses during sprinting. Using a prosthetic foot that acted much like a spring to respectively control and create dorsiflexion and plantarflexion during stance, simulated ankle function was observed. This simulated ankle function meant transtibial
507
Buckley
amputee subjects were able to achieve an “up-on-the-toes” sprinting gait; such digitgrade gait is typical of able-bodied sprinting. Optimizing the spring characteristics (eg, stiffness, natural frequency) through design modifications or alignment adjustments (eg, stiffness could be increased by anteriorally tilting the shin to increase the foot’s plantarflexed position at foot-strike) could maximize limb function in sprinting. The challenge would be to achieve this while maintaining optimal limb function through the range of speedspreceding maximum sprinting speed. The task of improving transfemoral amputee sprinting is more complex. In this study, the prosthesis was fully extended prematurely in swing. It is likely that this early extension was a function of the swing phase control device being incorrectly adjusted for this mode of gait. Prosthetic modification could ensure the knee flexed and extended at the appropriate rate for running. Modification could also ensure the knee was flexed at foot strike and remained so during early stance; this would encourage limb function more typical of able-bodied sprinting. Again, the challenge would be to develop a limb that is able to function optimally through the range of speeds up to and including maximum sprinting. Having a prosthesis that performs with continually changing stance yielding and swing phase control settings could be accomplished by incorporating microprocessor technology into existing hydraulic knee units. SUMMARY This study examined lower extremity motion in several of the world’s best lower-limb amputee sprinters. Prosthetic limb kinematics for transtibial amputee subjects were similar to sound limb kinematics with subjects achieving an “up-on-thetoes” gait typical of able-bodied sprinting. The kinematics of the prosthetic limb of the transfemoral amputee subject was considerably disrupted when compared with the sound limb or able-bodied subjects. The knee was fully extended before and during stance. This meant kinematics for this limb were more typical of those seen in walking and resulted in a sprinting gait that had large kinematic asymmetries between contralateral limbs. Prosthetic developments were discussed. The results presented here cannot be viewed as generalizable to a wider population, but they do provide an insight into how some of the world’s best amputee athletes use their prostheses during sprinting. Because subjects used only a single prosthesis each, the influence of prosthetic componentry could not be assessed.To determine the biomechanical adaptations of the sound and prosthetic limb during this higher mode of gait, future work could adopt a segmental energy flow analysis. Such an approach would also be required to fully assessthe influence of prosthetic design on the dynamics of sprinting. Acknowledgments: This work is part of an ongoing study, in collaborationwith and partly funded by ChasA Blatchford and Sons Ltd., Basingstoke,Hampshire,UK, investigatingthe biomechanicsof running and sprinting in lower-limb amputee subjects.The author thanksAnn DeRidder and Chris Scofieldfor their help during the data collection. References
Mann RA, Hagy JL. Biomechanicsof walking, running, and sprinting.Am J SportsMed 1980;8:345-50. MacIntyreDL, RobertsonDGE.EMG profilesof the kneemuscles during treadmill running. In: JonssonB, editor. Biomechanics X-A. Champaign(IL): Human Kinetics;1987.p. 289-94. Ito A, Saito M, SagawaK, Kato K: Ae M, Kobayashi K. Leg movementanalysisof gold and silver medalistsin men’s 1OOmat the III World Championshipsin athletics.In: Proceedingsof the InternationalSocietyof BiomechanicsXlVth Congress;1993 4-8 July;Paris.Paris:ISB; 1993.p. 624-5. Arch
Phys
Med
Rehabil
Vol 80, May
1999
508
AMPUTEE
SPRINTING,
4. Buckley JG, Zahedi MS. Biomechanics of amputee sprinting. In: Van Coppenolle H, Vanlandewijck Y, Simons J, Van de Vliet P, Neerinckx E, editors. Proceedings of the European Conference on Adapted Physical Activity and Sports: a White Paper on Research and Practice: 1994 Dee 18-20: Leuven, Belgium. Leuven. Belgium: Vitgeverigi Acco; 1995. p. 111-5. ” 5. Lewis JV, Buckley JG, Zahedi MS. An insight into paralympic amputee sprinting. Br J Ther Rehabil 1996;3:440-4. 6. Miller DI. Resultant lower extremity joint moments in below-knee amputees during running stance. J Biomech 1987;20:529-41. I. Enoka RM, Miller DI, Burgess EM. Below-knee amputee running gait. Am J Phys Med 1982;62:66-84. 8. Smith AW. A biomechanical analvsis of amuutee athletic gait. Int J Sport Biomech 1990;6:262-82. d I 9. Czerniecki JM, Gitter A. Insights into amputee rnnning. Am J Phys Med Rehabil1992;71:209-18. 10. Czemiecki JM. Gitter A. Munro C. Joint moments and muscle power output ‘characteristics of below-knee amputees during running: the influence of energy storing feet. J Biomech 1991;24: 63-5. 11. Czerniecki JM, Gitter A, Beck JC. Energy transfer mechanisms as a compensatory strategy in below knee amputee runners. J Biomech 1996;29:717-22. 12. Miller DI, Enoka RM, McCulloch RG, Burgess EM, Frankel VH. Vertical ground reaction force-time histories of lower extremity amputee runners. In: Morecki A, Fidelus K, Kedzior K, Wit A, editors. Biomechanics VII-A. Baltimore (MD): University Park Press; 1981. p. 453-60.
Arch
Phys
Med
Rehabil
Vol 80, May
1999
Buckley
13. Sanderson DJ, Martin PE. Joint kinetics in unilateral below-knee amputee patients during running. Arch Phys Med Rehabil 1996;77: 1279-85. 14. Bartlett RM. A biomechanical analysis program package [masters thesis]. London: City University of London; 1990. 15. Bartlett RM, editor. Biomechanical analysis of movement in sport and exercise. Leeds: British Association of Sport and Exercise; 1997. 16. Bland JM. An introduction to medical statistics. Oxford: Oxford University Press; 1995. 17. Thomas JR, Nelson JK. Research methods in physical activity. Champaign (IL): Human Kinetics; 1990. Suppliers a. Llhas. A. Blatchford & Sons Ltd., Lister Road, Basingstoke, Hampshire RG22 4LU, United Kingdom. b. Flex-Foot, Inc., 27412-A Laguna Hills Drive, Aliso Viejo, CA 22656. C. [ceross; OSSUR UK, Synergy House, Manchester Science Park, Suildhall Close, Manchester Ml5 6SY, United Kingdom. d. MatsushitaElectric Industrial Co., Ltd., Central PO Box 288, Osaka 530-8691 Japan. e. 4com Computers Technical Support Department, Acorn House, Vision Park, Histon, Cambridge CB4 4AE, United Kingdom. f. hlam Instruments, Brunel Institute of Bioengineering, Uxbridge, Middlesex, United Kingdom.
Sprint Kinematics of Athletes With Lower-Limb
Amputations
John G. Buckley, MSc ABSTRACT. Buckley JG. Sprint kinematics of athletes with lower-limb amputations. Arch Phys Med Rehabil 1999;80:501508.
Objective: To determine and compare the kinematics of the sound and prosthetic limb in five of the world’s best unilateral amputee sprinters. Subjects: Five men, all unilateral lower-limb amputee (one transfemoral, four transtibial) athletes. The individual with transfemoral amputation used a Endolite Hi-activity prosthesis incorporating a CaTech hydraulic swing and stance control unit, a Flex-Foot Modular III, and an ischial containment total contact socket. Those with transtibial amputations used prostheses incorporating a Flex-Foot Modular III and patellar tendonbearing socket, with silicone sheath liner (Iceross) and lanyard suspension. Design: Case series. Subjects were videotaped sprinting through a performance area. Sagittal plane lower-limb kinematics derived from manual digitization (at 50Hz) of the video were determined for three sprint trials of the prosthetic and sound limb. Hip, knee, and ankle kinematics of each subject’s sound and prosthetic limb were compared to highlight kinematic alterations resulting from the use of individual prostheses. Comparisons were also made with mean data from five able-bodied men who had similar sprinting ability. Results: Sound limb hip and knee kinematics in all subjects with amputation were comparable to those in able-bodied subjects. The prosthetic knee of the transfemoral amputee athlete fully extended early in swing and remained so through stance. In the transtibial amputee athletes, as in able-bodied subjects, a pattern of stance flexion-extension was evident for both limbs. During stance, prosthetic ankle angles of the transtibial amputee subjects were similar to those of the sound side and those of able-bodied subjects. Conclusion: Prosthetic limb kinematics in transtibial amputee subjects were similar to those for the sound limb, and individuals achieved an “up-on-the-toes” gait typical of ablebodied sprinting. Kinematics for the prosthetic limb of the transfemoral amputee subject were more typical of those seen for walking. This resulted in a sprinting gait with large kinematic asymmetries between contralateral limbs. 0 1999 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation
From the Department of Exercise and Sport Science, The Manchester Metropolitan University, Stoke-on-Trent, United Kingdom. Submitted for publication June 1, 1998. Accepted in revised form October 1, 1998. Supported in part by Chas. A. Blatchford and Sons Ltd., Basingstoke, Hampshire, United Kingdom. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author or upon any organization with which the author is associated. Reprint requests to J.G. Buckley, The Manchester Metropolitan University. Department of Exercise and Sport Science, Hassall Road; Stoke-on-Trent ST7 ZHL, United Kingdom. B 1999 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation 0003.9993/99/8005-5050$3.00/O
A
FTER LOWER LIMB amputation, the ultimate rehabilitation goal for many patients is to attain a full and active lifestyle. For some, this means being involved in competitive sport. Because many sports require the ability to run or sprint, designing a prosthesis that will accommodate a near-to-normal leg action is a priority in prosthetic evolution. From studies of able-bodied athletes, it is apparent that the biomechanics of running and sprinting are noticeably different. The duration of stance, as a percentage of stride time, decreases from 31% for running to 22% for sprinting.t In sprinting the center of mass of the body is lowered by increasing hip, knee, and ankle flexion, and during the stance phase of sprinting, only knee flexion occurs, whereas both flexion and extension occur during running. l-3 Whether lower limb amputee persons achieve kinematics in sprinting comparable to those achieved by able-bodied athletes is the focus of this article. This work is part of an ongoing study4z5attempting to quantify the biomechanical parameters and prosthetic requirements of sprinting by amputee runners. Previously, the only amputee athletic gait that has been analyzed biomechanically in detail is transtibial amputee running at a moderate pace.6 i3 Findings suggest that for individuals with this level of amputation, running is characterized by several compensatory mechanisms. During stance there is prolonged and excessive knee extension with only a small change in the thigh angle of the prosthetic limb.7 Two to three times more total work is done by the hip of the amputee runner’s prosthetic limb than by the intact limb.9:13 Typically, there is an absence of an impact ground reaction force peak for the prosthetic limb. t2 During swing the total work done by the prosthetic limb is similar to that in able-bodied subjects9 In contrast, the sound limb exhibits a 69% increase in the total work done, with an accompanying increase in the energy transferred from this limb to the trunk.11 Because the number of persons with amputations who participate in sporting activities has been increasing, and because many of these activities require the ability to sprint, a similar understanding of the biomechanical adaptations adopted during sprinting is required. Not only will this knowledge allow specific rehabilitation and training strategies to be developed, but it also will allow developments in prosthetic hardware (made to encourage and facilitate a more active lifestyle) to be objectively evaluated. This study was conducted to determine how amputee athletes use their prostheses during maximal sprinting. The specific aim was to measure and compare the kinematics of the sound and prosthetic limbs in five of the world’s best unilateral amputee sprinters. In an attempt to highlight kinematic alterations resulting from the use of an individual prosthesis, the kinematics of the sound and prosthetic limb of each amputee athlete were also compared with those of a group of able-bodied athletes who had similar personal best 100m times. METHODS Subject Data Five men with unilateral lower-limb amputations (one transfemoral and four transtibial) and five able-bodied men volunteered to participate. The five amputee athletes (mean age, 25.2 i 5.lyr.s; mean body mass, 70.4 i 5.9kg) all undertook Arch
Phys
Med
Rehabil
Vol 80, May
1999
502
AMPUTEE
SPRINTING,
athletic training on a weekly basis. All were experienced competitors with personal bests for the 100m sprint ranging from 11.7sec to 14.0sec (transtibial) and 15. lsec (transfemoral). The five able-bodied athletes (mean age, 18.2 +- 2.4yrs; mean body mass, 75.0 5 7.7kg) were members of a local athletic club and had personal bests for the 1OOmsprint ranging from 11.3sec to 12.6sec. All subjects were informed of the purpose of the study, and written informed consent was obtained from each subject before testing. The study met with local bioethics committee approval. The transfemoral amputee patient (subject 5) used an Endolite Hi activity prosthesisa incorporating a CaTech hydraulic swing and stance control unit,a a Flex-Foot Modular III,b and an ischial containment total contact socket. This limb was also worn for “everyday” use. All transtibial amputee athletes used prostheses incorporating a Flex-Foot Modular IIIb and patellartendon-bearing suction socket, with silicone sheath 1inerCand lanyard suspension. Subject 1 had a dedicated sprint prosthesis, which was adjusted so its length was the same as the sound leg when the wearer stood on tip-toe. The other three subjects used prostheses that had been fitted and aligned for “everyday” use; however, the prosthesis of subject 3 was modified by anteriorally tilting the shin by 10” at the socket mount to place the foot into a more plan&r-flexed position. The residual knee of subject 4 was unable to flex beyond 90”, because it had been pinned after the trauma that led to the transtibial amputation.
Buckley
joints, hips, knees, ankles, and distal end of feet). Markers were placed on the prosthetic foot at the same height as the lateral malleolus of the intact limb during normal standing and on the prosthetic socket, at the position corresponding to the underlying knee center (transtibial) or hip center (transfemoral) (fig 1 inset). All other points were located by assuming joint centers and segment endpoints to lie on the midlines of body segments.15 The raw two-dimensional marker coordinate data were then smoothed and differentiated using cross-validatedquintic splines. The smoothed data were used to define a 14 linked-segment model (fig 1) of the performer in the sagittal plane, from which all kinematic parameters were determined. Knee and ankle angles were measured as the absolute angles between the thigh and shank segment and shank and foot segment, respectively; thus, a fully extended knee, for example, would have an angle of 180”. The ankle angle at the instance when the shank was fully vertical during stance in each subject was used as a zero reference; thus, positive angles indicated plantarflexion and negative angles dorsiflexion. The thigh angle was measured as the included angle of the thigh segment to the vertical; positive angles thus indicated hip flexion and negative angles hip extension. Sprint speed was the average speed of the subject’s center of mass during a complete stride. Data Analysis Time series data were normalized to a percentage of each subject’s cycle (stride) time (the period between consecutive
Data Collection Subjects were given time to warm up and to become familiar with sprinting through a “performance” area on an outdoors cinder track. To ensure that subjects could achieve their maximum sprinting speed, starting points were set approximately 20m from the performance area; these were adjusted for each subject to ensure that stance of one limb occurred in approximately the middle of the performance area without any overstriding or understriding. A 20m run-off allowed subjects to safely reduce their speed. Subjects were videotaped using two Panasonic F15 video cameras,dmounted on tripods approximately 15m from, perpendicular to, and on either side of the performance area. Amputee subjects performed 8 to 10 sprints until three trials in which the prosthetic foot and three trials in which the sound foot was planted in the middle of the performance area were recorded. These were termed successful trials. Because asymmetry between contralateral sides was not investigated in the ablebodied sprinters, they were required only to complete three successful trials in which either their left or right foot was planted in the middle of the performance area. Regulating foot placement in this manner meant a full stride occurred within the performance area and allowed the size of the camera image to be maximized. Data Processing All data processing was carried out using Bartlett’s biomechanical analysis software. I4 Video recordings from the camera on the side closest to the limb landing in the middle of the performance area were analyzed using an Archimedes 440 microcomputef fitted with a Moving-Image high resolution (640 X 512) color digitizer.f From just before toe-off to just after foot-strike of the ipsilateral limb, each field (50Hz) was manually digitized by locating 18 points on the body (apex of head, C7, shoulders, elbows, wrists, third metacarpophalangeal Arch
Phys
Med
Rehabil
Vol 80, May
1999
4
thigh angle
centreof mass
/
A,
ankle
angle .P=J
Fig 1. Linked-segment model of performer showing how limb angles were defined. Inset shows marker placement used on prosthetic limb; all other points were located by assuming joint centers and segment endpoints to lie on the midlines of body segments.14 The ankle angle at the instance when the shank was fully vertical durina stance was used as a zero reference; thus, oositive anales
AMPUTEE
SPRINTING,
foot-strikes of the ipsilateral limb). The mean of the three successful trials was then determined. For the able-bodied subjects individual means were then added to give a group mean. Because sprinting technique is highly individualized it is difficult to make reference to ideal or optimal technique. One could argue, therefore, that an able-bodied group mean has limited meaning; certainly this would be the case if the data were used to represent a control group from which to make statistical comparison. The kinematics for the able-bodied group were determined in an attempt to highlight, qualitatively, the kinematic alterations resulting from the use of individual prostheses. Without such a comparison it would be difficult to determine whether sound limb kinematics were typical of those seen in able-bodied sprinting. Findings for the able-bodied subjects demonstrated low intersubject variability, suggesting subjects performed kinematically in similar fashion. The low intersubject variability meant the concerns regarding determining a group mean (for this particular group) were allayed. Digitizing repeatability was assessedby digitizing one trial for one subject 10 times over two separate digitizing sessions (days). After smoothing the coordinate and angular displacement data, the variability (residual mean square within subjects [S,]) in the angular displacement at the knee between the 10 trials was evaluated over the complete stride period using analysis of variance (ANOVA). The 95% agreement limitsi were then determined as 22 d/s,. To give an indication of intrasubject consistency, the alpha (stability) coeflicients17 for the three repeated trials (for each subject) were determined using ANOVA. RESULTS Error Estimates Statistical analysis (ANOVA) indicated that repeatability in determining knee angular displacements within 95% agreement limits was 23.0”. Thus kinematic differences between prosthetic and sound limb, or either limb and the able-bodied mean, were only considered of consequence if they were greater than 3”. The intrasubject consistency in angular displacements at the knee for the three trials was 2.98 in all subjects. The high consistency values likely reflect that subjects, who undertook athletic training on a regular basis, were highly skilled amputee sprinters. Data Interpretation Because of differences in prosthetic alignment, amputation level, stump length/strength, etc, it would have been inappropriate to group the amputee subjects to make statistical comparisons with the able-bodied group. The results are thus presented as a series of individual case reports. In each case, sound and prosthetic limb kinematics were compared with each other and with the able-bodied mean. It is evident from the ankle, knee, and thigh angular displacement data for each subject (figs 2 through 6, parts a through c in each fig) that there were considerable kinematic differences between prosthetic and sound limbs and between both limbs and the able-bodied mean. It is apparent that these differences were greatest for the ankle and least for the thigh and were, with the exception of the transfemoral amputee athlete’s prosthetic limb, related to the magnitude of angular displacement, with the temporal occurrence of displacement changes remaining constant. The ankle angle graphs show there were small angular displacement changes during swing. This apparent motion was likely the result of out-of-plane movement of the foot during swing, appearing to the camera as dorsiflex-
Buckley
%Of Stride (foot-strike to foot-strike, b) Knee angular displacement
a) Ankle angular displacement
;o
0 20 40 60 thighangle,cieg,
d) Thigh-knee angle-angle diagram
Fig 2. Subject 1 (transtibial): (a) ankle, (b) knee, and (c) thigh graphs and (d) thigh-knee angle-angle diagram, for prosthetic (dotted line) and sound limb (solid bold line), with able-bodied (solid thin line). 0, toe-off; 0, foot-strike.
angle limb mean
ion or plantarflexion. Because the limb is more in line with the plane of progression and thus perpendicular to the camera during stance, this problem was considered inconsequential to data (presented) for the stance phase. The results were also presented as thigh-knee angle-angle diagrams (figs 2 through 6, part d in each fig). Because there is no requirement to normalize data to percentage of cycle time, such diagrams allow rapid assessmentof limb function, albeit in
%Of Stride (foot-strike to ‘c.ot-*trike)
%o‘stride (foot-strike to foot-strike, b) Knee angular displacement
a) Ankle angular displacement
c) Thigh angular displacement
d) Thigh-knee angle-angle diagram
Fig 3. Subject 2 (transtibial): (a) ankle, graphs and (d) thigh-knee angle-angle (dotted line) and sound limb (solid bold (solid thin line). 0, toe-off; 0, foot-strike.
Arch
Phys
(b) knee, and (c) thigh diagram, for prosthetic line), with able-bodied
Med
Rehabil
Vol80,
May
angle limb mean
1999
AMPUTEE
a) Ankle angular displacement
b) Knee angnlar displacement
%of stride (foot-strike to foot-strike)
(b) knee, and (c) thigh diagram, for prosthetic line), with able-bodied
angle limb mean
a qualitative manner. The overall size of the diagram indicates the range of motion (the larger the diagram the greater the range), and its shape shows the coordinated action of the thigh and knee. The markedly reduced plot size for subject 4’s prosthetic limb (fig 5d) (his knee was unable to flex beyond 90’) illustrates how clearly these diagrams can highlight kinematic differences.
Kof stride (foot-strike to foot-strike) a) Ankle angular displacement
b) Knee angular displacement
c) Thiih angular displacement
Fig 5. Subject 4 (transtibial): (a) ankle, graphs and (d) thigh-knee angle-angle (dotted line) and sound limb (solid bold (solid thin line). 0, toe-off; 0, foot-strike.
Arch
Phys
Med
Rehabil
Vol 80, May
Buckley
a) Ankle angolv displacement
thigh angle (deg) d) Thigh-knee angle-angle diagram
c) Thigh angular displacement
Fig 4. Subject 3 (transtibial): (a) ankle, graphs and (d) thigh-knee angle-angle (dotted line) and sound limb (solid bold (solid thin line). 0, toe-off; 0, foot-strike.
SPRINTING,
d) Thigh-knee angle-angle diagram
(b) knee, and (c) thigh diagram, for prosthetic line), with able-bodied
1999
angle limb mean
d) Tbigh-knee angle-angle diagram
Fig 6. Subject 5 (transfemoral): (a) ankle, (b) knee, and (c) thigh angle graphs and (d) thigh-knee angle-angle diagram, for prosthetic limb (dotted line) and sound limb (solid bold line), with able-bodied mean (solid thin line). 0, toe-off; 0, foot-strike.
To objectively assess the kinematic differences between prosthetic and sound limbs, joint angular displacements and velocities of each limb at key moments in the sprint cycle were determined. Figure 7 highlights these differences for subject 2 (whose limb kinematics were closest to the midrange of all transtibial amputee subjects), and figure 8 the differences for the transfemoral amputee subject (subject 5). Transtibial Amputee Sprinting Results indicate that transtibial amputee subjects all displayed similar limb kinematics. Thus, to avoid repetition, kinematics general to all subjects are described. In addition, the kinematic differences between prosthetic and sound limbs, quantified in figure 7 for subject 2, are referred to (listed in brackets in the text) because they serve to highlight the differences found for all transtibial amputee subjects. The thigh-knee angle-angle diagrams highlight that the kinematics for the sound side compared well with those for the able-bodied subjects. The diagram for the prosthetic side indicates that this limb also achieved a kinematic pattern similar to that of the sound limb, although the amplitude of this pattern indicates reduced angular displacement at the knee. Sound limb dorsiflexion (15’) to midstance was greater than that observed on the prosthetic limb (11”). Plantarflexion at toe-off was also greater on the sound limb (5”) than on the prosthetic limb (-2”). Ankle displacements, however, on the prosthetic limb (during stance) were only marginally less than those of the sound limb. Because these displacements were achieved over a longer stance time (26% of the cycle as opposed to 22%) angular velocities were considerably reduced. In particular, the plantarflexion velocity on the prosthetic limb was less than 40% of that on sound side. Figures 2 through 5 (part a in each fig) indicate there was greater range of ankle displacement on the sound limb than in able-bodied subjects. It is apparent from part b in figures 2 through 5 that with the exception of subject 1 the prosthetic limb, when compared with the sound limb or able bodied subjects, extended at the knee
AMPUTEE
SPRINTING,
Buckley
able-body
I
I
Mid
FS
*
22%,3(,1,, TO
0%
Mid
SWING
STANCE able-body
\
adan= 6.8
l-
Fig 7. Kinematic differences between prosthetic (pros) and sound limbs of subject 2 (transtibia1 amputation). (A) Angular displacement differences at key moments in the sprint cycle; (B) the maximum angular velocities (rad/sec) observed between the key moments. FS,foot-strike; TO, toeoff; ank (or a), ankle; k, knee; h, hip; flex, flexion; ext, extension; dorsi, dorsiflexion; plantar, plantarflexion; V, velocity.
L
lOb% FS
a+=
16.6
sound=
18.4
-I
B
FS
prematurely during late swing from approximately 85% of the cycle onwards. It then either remained in extension or began to flex before foot-strike. After a period of flexion during early stance, the knee then extended to toe-off. Although the pattern of stance flexion-extension was similar for both limbs, the prosthetic knee (160”) was not extended to the same extent as the sound knee (168”) at toe-off (fig 7). Maximum knee flexion angles on the prosthetic side (58”) during swing indicate that this limb was less flexed than the sound limb (53”). Peak knee flexion velocity on the prosthetic limb (16.4rad/sec) during swing was also less than on the sound limb (18.4rad/sec). Sound limb knee kinematics were similar to those achieved by the able-bodied subjects (figs 2 through 5). Figures 2 through 5 (part c in each fig) indicate that hip kinematics for the sound and prosthetic limbs were comparable to those for able-bodied subjects and that there were only slight differences between contralateral sides. At foot-strike the prosthetic limb (34”) was more flexed than the sound limb (2S”) (fig 7). The prosthetic limb (- 14”) was less extended at toe-off
Mid
22% 3_(; ‘f(, TO
100% FS
Mid
STANCE
SWING
than the sound limb (-27”). After toe-off, both sides extended to approximately the same extent, but subsequent swing flexion was less on the prosthetic side (44”) than the sound side (52”) (fig 7). Peak thigh angular velocities for each limb were similar. Transfemoral Amputee Sprinting The shape, size, and position of the thigh-knee angle plot for the transfemoral amputee runner’s prosthetic limb (fig 6d) indicate that the kinematics for this limb were different from those on the sound side or able-bodied subjects. The increased amplitude of this plot in the vertical direction indicate the knee was overextended for a prolonged period. Its decreased width indicates a reduced range of motion of the thigh. Despite such kinematic disruption on the prosthetic side, there was no noticeable evidence, apart from a reduced range of knee flexion, of compensatory alterations in the sound limb kinematics. Data for the ankle (fig 6a) highlight a phase shift in the flexion-extension pattern on the prosthetic limb. At foot-strike there was slight plantarflexion, something not seen in any other Arch
Phys
Med
Rehabil
Vol 80, May
1999
506
AMPUTEE
SPRINTING,
Buckley
pros
-
sound
6% A
FS
Mid
TO
Mid
SWING
STANCE
able-body
pros
-
sound
0% FS
B
Mid
30% jLfo/ii TO
STANCE subject. This meant the instance of maximum dorsiflexion occurred 30% into the cycle compared with at 18% of the cycle observed for the sound limb. Displacement magnitudes, however, were comparable to those on the sound limb or ablebodied subjects. The knee angle graph (fig 6b) shows clearly the knee was fully extended early in swing (from approximately 80% of the cycle) and remained so late into stance. The sound limb knee in contrast, was flexed to 148” at foot-strike then extended to a maximum of 166” at toe-off (fig 8). This kinematic pattern was similar to that observed for able-bodied subjects. During swing there was reduced knee flexion, with angle of maximum llexion of 79” being considerably less than that on the sound side (58”) or that of able-bodied subjects (38” +- 7”) (fig 8). Sound limb thigh angles were similar to those of able-bodied subjects. In contrast, the prosthetic limb was 16” less flexed at foot-strike and, after a period of stance phase extension, began to flex before toe-off, an occurrence not seen in any of the other subjects. During swing the maximum flexion angle on the prosthetic side (39”) was less than the sound side (56”), as was Arch
Phys
Med
Rehabil
Vol 80, May
1999
100%
Mid
SWING
FS
Fig 8. Kinematic differences between prosthetic (pros) and sound (sound) limbs of the transfemoral amputee subject. (A) Angular displacement differences at key moments in the sprint cycle; (B) the maximum angular velocities (radl set) observed between the key moments. FS, foot-strike; TO, toe-off; ank (or a), ankle; k, knee; h, hip; flex, flexion; ext, extension; dorsi, dorsiflexion; plantar, plantarflexion; V, velocity.
the peak flexion angular velocity (fig 8). The stance phase peak extension velocity of 6.0 radlsec was approximately 60% of that achieved on the sound side. DISCUSSION In sprinting, data for elite able-bodied athletes indicate that the knee during stance undergoes flexion only, unlike walking and running where both llexion and extension 0ccur.l This suggests that in sprinting knee function has a limited role in propelling the body forward. In the present study both the prosthetic and sound limbs of the transtibial amputee runners, the sound limb of the transfemoral athlete, and the able-bodied subjects demonstrated knee flexion and extension during stance rather than flexion only. It may well be that stance with knee flexion only occurs at sprinting speeds in excess of the top speed (8.5mlsec) achieved by any of the subjects in the present study. The flexion-extension pattern at the knee of the prosthetic limb, when compared with able-bodied subjects, has been shown to be disrupted in transtibial amputee running with
AMPUTEE
SPRINTING,
prolonged and excessive extension during stance.’ In the present study, the transtibial amputee runners were able to achieve a similar pattern of flexion-extension during stance on the prosthetic limb as was achieved on the sound limb. The disparity in these findings can be explained by the type of prosthesis used. In the study by Enoka and colleagues,’ subjects used prostheses incorporating a SACH or Greissinger foot, whereas in the present study subjects had prostheses that incorporated a Flex-Foot Modular III. Essentially, the Modular III artificial foot is a “Y-shaped carbon fiber leaf spring that deforms during loading. This creates controlled dorsiflexion as the shank rotates forward over the planted foot and in turn allows flexion at the knee. The results presented here indicate that the pattern of dorsiflexion and plantarflexion achieved on the prosthesis was similar to that of the sound limb. The stance flexion-extension pattern at the knee of the prosthetic limb for the transtibial amputee subjects in the present study, therefore, may well be attributable to the dynamic function of the carbon fiber foot. Prosthetic knee kinematics during swing in transtibial amputee athletes were different from those for the sound limb. Apart from one subject, the knee overextended prematurely during late swing and mean peak flexion and extension angular velocities were lower than those attained on the sound limb. Given that the insertion of the hamstrings and quadriceps muscle groups are unaffected by transtibial amputation, it is unclear why this disrupted pattern was seen. It may have been due to the prosthetic socket restricting knee function or to differences in the inertial properties of the prosthetic shank and foot. In either caseprosthetic modification could ameliorate this kinematic alteration. It is evident, as might be expected, that the transfemoral amputee runner had greater kinematic asymmetry between prosthetic and sound limbs. Because the knee was fully extended early in swing, and remained so until late stance, it seemsthe prosthetic limb was used merely as a rigid support. A fully extended limb meant foot-strike was made with the heel keel. Because the keel is designed to flex on loading this explains why plantarflexion following foot-strike was seen in this subject. Such a limb position would also ensure the resultant instantaneous ground reaction force vector would stabilize the knee during the impact-absorbing phase of stance. Because the prosthesis used by the subject was also worn in everyday use and therefore had to facilitate walking, it is not surprising that such limb function, which is more typical of walking, was observed. Thigh kinematics of the sound limb for all amputees compared well with able-bodied subjects. On the prosthetic side the flexion-extension temporal pattern also matched that of the sound limb. However, maximum flexion angle and angular velocity and maximum extension angle and angular velocity during swing and stance respectively were reduced. It is likely that the prematurely and overextended knee was a contributing factor to this reduction, because an overextended knee would increase the moment of inertia of the leg about the hip. This would affect the swing kinematics and in turn could result in a less than optimum limb position at foot-strike, thus affecting stance kinematics. The results presented here cannot be viewed as generalizable to a wider population. They do, however, provide an insight into how some of the world’s best amputee athletes use their prostheses during sprinting. Using a prosthetic foot that acted much like a spring to respectively control and create dorsiflexion and plantarflexion during stance, simulated ankle function was observed. This simulated ankle function meant transtibial
507
Buckley
amputee subjects were able to achieve an “up-on-the-toes” sprinting gait; such digitgrade gait is typical of able-bodied sprinting. Optimizing the spring characteristics (eg, stiffness, natural frequency) through design modifications or alignment adjustments (eg, stiffness could be increased by anteriorally tilting the shin to increase the foot’s plantarflexed position at foot-strike) could maximize limb function in sprinting. The challenge would be to achieve this while maintaining optimal limb function through the range of speedspreceding maximum sprinting speed. The task of improving transfemoral amputee sprinting is more complex. In this study, the prosthesis was fully extended prematurely in swing. It is likely that this early extension was a function of the swing phase control device being incorrectly adjusted for this mode of gait. Prosthetic modification could ensure the knee flexed and extended at the appropriate rate for running. Modification could also ensure the knee was flexed at foot strike and remained so during early stance; this would encourage limb function more typical of able-bodied sprinting. Again, the challenge would be to develop a limb that is able to function optimally through the range of speeds up to and including maximum sprinting. Having a prosthesis that performs with continually changing stance yielding and swing phase control settings could be accomplished by incorporating microprocessor technology into existing hydraulic knee units. SUMMARY This study examined lower extremity motion in several of the world’s best lower-limb amputee sprinters. Prosthetic limb kinematics for transtibial amputee subjects were similar to sound limb kinematics with subjects achieving an “up-on-thetoes” gait typical of able-bodied sprinting. The kinematics of the prosthetic limb of the transfemoral amputee subject was considerably disrupted when compared with the sound limb or able-bodied subjects. The knee was fully extended before and during stance. This meant kinematics for this limb were more typical of those seen in walking and resulted in a sprinting gait that had large kinematic asymmetries between contralateral limbs. Prosthetic developments were discussed. The results presented here cannot be viewed as generalizable to a wider population, but they do provide an insight into how some of the world’s best amputee athletes use their prostheses during sprinting. Because subjects used only a single prosthesis each, the influence of prosthetic componentry could not be assessed.To determine the biomechanical adaptations of the sound and prosthetic limb during this higher mode of gait, future work could adopt a segmental energy flow analysis. Such an approach would also be required to fully assessthe influence of prosthetic design on the dynamics of sprinting. Acknowledgments: This work is part of an ongoing study, in collaborationwith and partly funded by ChasA Blatchford and Sons Ltd., Basingstoke,Hampshire,UK, investigatingthe biomechanicsof running and sprinting in lower-limb amputee subjects.The author thanksAnn DeRidder and Chris Scofieldfor their help during the data collection. References
Mann RA, Hagy JL. Biomechanicsof walking, running, and sprinting.Am J SportsMed 1980;8:345-50. MacIntyreDL, RobertsonDGE.EMG profilesof the kneemuscles during treadmill running. In: JonssonB, editor. Biomechanics X-A. Champaign(IL): Human Kinetics;1987.p. 289-94. Ito A, Saito M, SagawaK, Kato K: Ae M, Kobayashi K. Leg movementanalysisof gold and silver medalistsin men’s 1OOmat the III World Championshipsin athletics.In: Proceedingsof the InternationalSocietyof BiomechanicsXlVth Congress;1993 4-8 July;Paris.Paris:ISB; 1993.p. 624-5. Arch
Phys
Med
Rehabil
Vol 80, May
1999
508
AMPUTEE
SPRINTING,
4. Buckley JG, Zahedi MS. Biomechanics of amputee sprinting. In: Van Coppenolle H, Vanlandewijck Y, Simons J, Van de Vliet P, Neerinckx E, editors. Proceedings of the European Conference on Adapted Physical Activity and Sports: a White Paper on Research and Practice: 1994 Dee 18-20: Leuven, Belgium. Leuven. Belgium: Vitgeverigi Acco; 1995. p. 111-5. ” 5. Lewis JV, Buckley JG, Zahedi MS. An insight into paralympic amputee sprinting. Br J Ther Rehabil 1996;3:440-4. 6. Miller DI. Resultant lower extremity joint moments in below-knee amputees during running stance. J Biomech 1987;20:529-41. I. Enoka RM, Miller DI, Burgess EM. Below-knee amputee running gait. Am J Phys Med 1982;62:66-84. 8. Smith AW. A biomechanical analvsis of amuutee athletic gait. Int J Sport Biomech 1990;6:262-82. d I 9. Czerniecki JM, Gitter A. Insights into amputee rnnning. Am J Phys Med Rehabil1992;71:209-18. 10. Czemiecki JM. Gitter A. Munro C. Joint moments and muscle power output ‘characteristics of below-knee amputees during running: the influence of energy storing feet. J Biomech 1991;24: 63-5. 11. Czerniecki JM, Gitter A, Beck JC. Energy transfer mechanisms as a compensatory strategy in below knee amputee runners. J Biomech 1996;29:717-22. 12. Miller DI, Enoka RM, McCulloch RG, Burgess EM, Frankel VH. Vertical ground reaction force-time histories of lower extremity amputee runners. In: Morecki A, Fidelus K, Kedzior K, Wit A, editors. Biomechanics VII-A. Baltimore (MD): University Park Press; 1981. p. 453-60.
Arch
Phys
Med
Rehabil
Vol 80, May
1999
Buckley
13. Sanderson DJ, Martin PE. Joint kinetics in unilateral below-knee amputee patients during running. Arch Phys Med Rehabil 1996;77: 1279-85. 14. Bartlett RM. A biomechanical analysis program package [masters thesis]. London: City University of London; 1990. 15. Bartlett RM, editor. Biomechanical analysis of movement in sport and exercise. Leeds: British Association of Sport and Exercise; 1997. 16. Bland JM. An introduction to medical statistics. Oxford: Oxford University Press; 1995. 17. Thomas JR, Nelson JK. Research methods in physical activity. Champaign (IL): Human Kinetics; 1990. Suppliers a. Llhas. A. Blatchford & Sons Ltd., Lister Road, Basingstoke, Hampshire RG22 4LU, United Kingdom. b. Flex-Foot, Inc., 27412-A Laguna Hills Drive, Aliso Viejo, CA 22656. C. [ceross; OSSUR UK, Synergy House, Manchester Science Park, Suildhall Close, Manchester Ml5 6SY, United Kingdom. d. MatsushitaElectric Industrial Co., Ltd., Central PO Box 288, Osaka 530-8691 Japan. e. 4com Computers Technical Support Department, Acorn House, Vision Park, Histon, Cambridge CB4 4AE, United Kingdom. f. hlam Instruments, Brunel Institute of Bioengineering, Uxbridge, Middlesex, United Kingdom.