Biomechanical comparison of the energy-storing capabilities of SACH and Carbon Copy II prosthetic feet during the stance phase of gait in a person with below-knee amputation.

Because of its simplicity of design and durability, the solid-ankle cushion heel (SACH) foot is the most frequently prescribed prosthetic foot in the United States.[1] Because people with disabilities have increased participation in a variety of athletics, however, there is a demand for prosthetic feet that will endure and improve athletic performance. As a result of this demand, a new class of solid-ankle prosthetic components known as "energy-storing" feet has been developed. Among these newer designs are the Seattle Foot [TM],(*) the Flex-Foot [TM],[dagger] and the Carbon Copy II (CC II) foot.[double dagger] [2-4]

The keel of the CC Il foot consists of a carbon-composite flexible primary deflection plate extending to the proximal interphalangeal region of the forefoot and an auxiliary deflection plate extending to the midfoot region.[3-5] This keel is designed to deform from heel-strike to late stance, storing energy like a compressed spring, and then to release this stored energy just before toe-off to initiate swing. Furthermore, the lightweight materials used in the CC II foot presumably contribute to a smooth pendulum action during swing. The manufacturer of the CC 11 foot claims that this energy-storing mechanism enables people with amputations to improve their performance in vigorous activities such as running and jumping and in nonvigorous activities such as level walking.[5]

There is little evidence available to demonstrate that the CC II foot's keel performs as an energy-storing spring. This evidence is mostly limited to clinical comparisons and personal anecdotes.[3] Torburn et al,[6] who extensively studied five subjects with amputations using the CC II foot, reported no effect of the energy-storing design on stride characteristics, angular kinematics, joint kinetics, or oxygen consumption during level walking.

During the design and manufacturing process, the testing of prosthetic feet involves methods in which the feet are attached to instruments that measure their response to externally applied loads.[7-9] Such methods provide valuable information about the material and mechanical properties of a particular prosthetic component, but they have limited application to estimates of the foot's performance under the loading conditions of gait.

In the modern biomechanics laboratory, the kinematic and kinetic gait characteristics of subjects with amputations can be measured three dimensionally. This capability provides the opportunity to test different prosthetic feet under the conditions in which they ultimately will be used. Instead of attaching different prosthetic feet to inanimate instruments, investigators can attach them to a human "instrument," who will provide the ideal dynamic force input for analyzing the response of different feet to walking.

One variable of interest in the study of energy-storing prosthetic feet is joint muscle power, which is equal to the product of moment of force and joint angular velocity.[10] In the prosthetic limb, if the moment and angular velocity have the same polarity (ie, are both positive or are both negative in sign), the resulting positive power represents the rate of return of energy by the prosthesis. If the moment and angular velocity are opposite in polarity (ie, one is positive, and the other is negative), however, the resulting negative power represents the rate of storage of energy by the prosthesis. For a given gait cycle, the difference between the energy stored and the energy returned by the prosthesis represents dissipation of energy.

Robertson and Winter[11] showed that the ankle plantar flexors control the forward rotation of the shank over the supporting foot early in stance and that they are responsible for over 80% of the mechanical power generated at push-off during the normal gait cycle. People with below-knee (BK) amputations, therefore, are likely to compensate for the loss of the posterior shank musculature, either at other joints or through substitution by the components of the prosthesis, or through some combination of these two strategies.

Winter and Sienko[12] investigated the biomechanics of gait in 8 subjects with BK amputations using SACH (5 subjects), Uniaxial (2 subjects), and Greissinger [R] [subsection] (1 subject) prosthetic feet. Of particular interest in their observations were the differences in moment and power at the prosthetic ankle joint. They found that with all three prosthetic feet, the dorsiflexor moment was prolonged in early stance as compared with normal gait. All three feet stored energy (ie, had negative power) during mid-stance, but only the Uniaxial and Greissinger [TM] feet returned energy in late stance with 20% and 30% efficiency, respectively. These investigators have demonstrated the use of a method of measuring the energy storage and return capacity of prosthetic feet, which are designed to compensate for the normal function of the ankle plantar flexors.

In our study, kinematic and kinetic variables of the lower extremities were analyzed during level, nonvigorous gait of a person with a unilateral BK amputation using SACH and CC II prosthetic feet. The purpose of this study was to compare the variables during the stance phase of gait that were directly related to the energy-storing capabilities of the different feet. These variables were the progression of the center of pressure, prosthetic ankle joint angle, net prosthetic moment about the ankle, and prosthetic ankle joint power. Other kinematic gait variables (ie, step length, single-limb support time, swing period, double-limb support time, cadence, average velocity, and hip and knee flexion-extension) were analyzed to illustrate the consistency of the person's gait pattern while using the two different prosthetic feet. This consistency provided justification for the use of a person with a BK amputation as a component of the instrumentation rather than as a subject of the study.

Method

Instrumentation

Data were collected at The Biomechanics Laboratory, Department of Rehabilitation Medicine, National Institutes of Health (NIH), Bethesda, Md. A video-based (50-Hz), three-dimensional, kinematic data-acquisition system (VICON [R])(~~) and two AMTI(#) model OR6-3A (200-Hz) force plates were used for data collection. A data-capture volume measuring 0.606 (X), 1.813 (Y), and 1.200 (Z) m in the orthogonal laboratory coordinate system was defined by 15 reference targets placed midway along a 10-m walkway, such that the positive Z axis defined the vertical upward dimension and the negative Y axis defined the direction of ambulation (Fig. 1). The position and orientation of five cameras were determined (external calibration) before trial data acquisition by imaging the reference targets. A camera nonlinearity correction process (internal calibration) was performed before the external calibration process and was applied to all kinematic data.

The average external calibration error, a least-squares residual difference between the actual and individual camera-estimated calibration target locations, was 2.06 mm for the five cameras, with a range of 0.94 to 3.12 mm. The estimated resolution of this camera configuration relative to the field of view is 1 part in 3,000.[13] This means that, for a 3-m field, objects can be resolved to within 1 mm. Experiments conducted to determine the accuracy and precision of measurements made in this laboratory demonstrated root mean square errors no greater than 0.9 degree when angular displacements of a uniaxial pendulum were sampled.[14] When repeated measurements were made of ankle-subtalar flexion-extension of five subjects during gait, the mean width of one standard deviation about the mean angular displacement curve for each subject was no greater than 0.94 degree.[14] The magnitude of this within-subjects variability is small relative to the magnitude of the angular displacements about the joints of the lower extremity during gait and indicates that the video-based technique possesses the resolution needed to sample repeatable angular kinematic data from human subjects during gait to within 1 degree.

The force plates were positioned ill the middle of the data-collection volume such that the position and magnitude of vertical and anterior-posterior ground reaction forces generated by subsequent contralateral stance phases could be measured.[15] The time of onset and cessation of vertical ground reaction forces and a kinematic-based technique were used to estimate the times at which temporal and spatial gait events, such as heel-strike, mid-stance, and toe-off, occurred.[16]

The person who served as the component of the instrumentation in this study was an active 33-year-old woman with a unilateral BK amputation. She stood 1.73 m tall with either prosthetic foot and weighed 52 kg with the SACH foot and 53 kg with the CC II foot. She received a BK amputation secondary to synovial cell carcinoma of the right foot 51 months prior to participation in this study. She had since walked without any assistive device using a patellar tendon-bearing prosthesis with supra-condylar cuff and SACH foot. She gave her informed consent, then underwent a thorough musculoskeletal screening examination before completing this study in accordance with the NIH requirements for human experimentation. She had full and pain-free range of motion (ROM) of all joints and "normal strength" (defined as the ability to resist the examiner in a manual "break" test) of all muscle groups of the intact limb. On the side of the amputated limb, hip ROM and strength were equal to measurements obtained for the intact side. Knee ROM was also full and pain-free on the side of the amputation. The strength of both the knee extensors and flexors on the side of the amputated limb could not be overcome in a break test using the shortened tibial segment (11 cm from the lateral knee joint line to the distal end of the residual tibia) as the lever arm.

Girth measurements taken with a measuring tape at 10-cm intervals from the level of the greater trochanter to the inferior border of the patella of the intact and residual limbs differed by an average of 4.64 cm (3.00-7.00 cm) and were greater on the intact side. Limb length, measured with a measuring tape from the greater trochanter to the floor adjacent to the calcaneocuboid joint, was 85.0 cm on the intact side, 85.5 cm on the prosthetic side with the SACH foot, and 86.0 cm on the prosthetic side with the CC II foot.

For the purpose of this study, a certified prosthetist constructed and fitted an endoskeletal patellar tendon-bearing prosthesis with medial wedge suspension and a total-contact soft liner. The shank pylon was left exposed to allow for the exchange of the foot component. The person with the BK amputation donned the prosthesis only once during the entire data-collection period. She wore the same pair of soft-soled, low-heeled shoes with both prosthetic feet. Prosthetic foot type was randomly selected by a drawing. The prosthetist affixed and aligned the SACH foot first, followed by the CC II foot. The person with the BK amputation was allowed to accommodate to the foot by freely ambulating for approximately 10 minutes.

Retroreflective spherical targets 25.4 mm in diameter were affixed to the skin superficial to the superior aspect of the greater trochanters, the left lateral femoral condyle, and the lateral aspect of the left lateral malleolus. Targets were also affixed to the prosthesis lateral to the right femoral condyle, to the distal lateral aspect of the shank at the approximate location of the lateral malleolus, and to the foot at the approximate location of the fifth metatarsal head.

Data Collection

The person with the BK amputation walked at a self-determined "free" velocity through the data-collection volume. A data-collection trial commenced approximately 2 seconds before the person entered the data-collection volume and terminated approximately 2 seconds after she left the data-collection volume. Ten acceptable trials were gathered for each prosthetic foot. An acceptable trial was one in which the person with the BK amputation achieved complete consecutive steps occurring on sequential force plates and in which the kinematic data were continuous. The data-collection phase lasted for 55 minutes. A total of 46 trials of data were collected in order to obtain the 20 trials required for analysis. The person with the BK amputation did not report pain or fatigue, nor was tissue breakdown prevalent on inspection after data collection. There was a 30-minute rest period when the foot was exchanged, but the prosthesis was not removed at that time.

Data Reduction

The ADTECH Motion Analysis Software System (AMASS)(**) was used to create three-dimensional trajectories of the targets from the raw camera data. The NIH Automated Gait Evaluation Software (AGES) was used to compute the temporal, spatial, and kinetic variables from the target position time histories and force-plate data.

The variables of step length, single-limb support time, swing period, double-limb support time, cadence, and average velocity during a stride were defined in accordance with Winter.[17] Angular kinematics were defined as follows: hip joint angle was the anterior measure of the angle projected into the laboratory YZ plane formed by the intersection of the laboratory Z axis and the thigh segment; knee joint angle was the posterior measure of the angle projected into the laboratory YZ plane formed by the intersection of the thigh and shank segments; and ankle joint angle was the anterior measure of the angle projected into the laboratory YZ plane formed by the intersection of the shank and foot segments. Prosthetic ankle joint angle reflected foot deformation, although no prosthetic ankle joint motion was expected to occur with the solid-ankle design. Center of pressure was measured relative to its location at initial foot contact with respect to the laboratory Y axis. Net muscular and prosthetic moments about the hip, knee, and ankle in the sagittal plane were calculated as the inverse of the cross product of the ground reaction force vector during stance and the displacement vector from the joint target to the center of pressure.[18] Prosthetic ankle joint power was calculated as the product of the ankle joint angular velocity and the net prosthetic moment about the ankle.[10]

Data Analysis

Because each of the two prosthetic feet was tested 10 times using the person with the BK amputation as a biped system (and not as a subject) in which only the prosthetic limb condition was modified, two-way analyses of variance (ANOVAS) for repeated measures (with foot and side as factors) were used to evaluate step length, single-limb support time, and swing period. Duration of doublelimb support, cadence, and average velocity for the two prosthetic feet were compared by Student's t tests. The alpha level of significance for the ANOVAS and t tests was adjusted to .017 according to the Bonferroni method for multiple comparisons. This method guards against Type I errors in the case of multiple comparisons by dividing the chosen level of significance ([alpha] =.05) by the number of similar comparisons (three each for the ANOVA and Student's t test), thereby making each individual test more stringent.19 Although this method is considered conservative and may result in low statistical power, when only a few comparisons are made with 9 degrees of freedom, the adjusted significance levels are not dramatically different from those obtained when more exact methods for making multiple comparisons are used.[20] Ensemble averages were generated for hip joint angle, knee joint angle, and ankle joint angle kinematics throughout a stride and for center of pressure, net muscular moment about the hip, net muscular moment about the knee, net prosthetic moment about the ankle, vertical ground reaction force, and anterior-posterior ground reaction force during stance.

Results

Stride Characteristics

The means and standard deviations for step length, single-limb support time, and swing period are presented in Table 1. Step length was 26% longer for the prosthetic limb than for the intact limb, single-limb support time was 28% longer for the intact limb than for the prosthetic limb, and swing period was 34% longer for the prosthetic limb than for the intact limb.

[TABULAR DATA OMITTED]

The results of the two-way ANOVAS (see Tab. 2, for example) indicated that the differences between the intact and prosthetic sides for these variables were significant (P<.017). The exchange of prosthetic feet, however, did not affect any of these variables on either the intact or the prosthetic side; no significant differences (P>.017) were found to exist across feet and in the footxside interactions.

Table 2. Two-Way analysis of Variance for Repeated Measures of Step Length by
Foot and Side
Source      df    SS        MS         F         P
Foot        1     0.00096   0.00096     3.08    .11(a)
Error       9     0.00281   0.00031
Side        1     0.25249   0.25249    132.35   .0001(b)
Error       9     0.01717   0.00191
Footxside   1     0.00064   0.00064      1.15   .31(a)
Error       9     0.00502   0.00056
(a) Not significant at [alpha] =.017.
(b) Significant at [alpha] =.017. Level of significance
[alpha] =.05) adjusted for multiple comparisons by Bonferroni
method of correction.[19]

Means and standard deviations for double-limb support time, cadence, and average velocity are presented in Table 3. The results of the t tests indicate that there were no significant differences (P>.017) between prosthetic feet for any of these variables.

[TABULAR DATA OMITTED]

Hip and Knee Kinematics

Because the angular kinematic ensemble averages of the intact limb for both prosthetic feet were virtually identical on visual inspection when plotted, the values for the intact limb are presented as a single ensemble average in Figures 2 through 6.

Hip Joint Angles

The hip joint angle demonstrated a flexion-extension pattern over the course of a single stride (Fig. 2a). At heel-strike, the hip joint angle showed greater hip flexion on the side of amputation than on the intact side for both prosthetic feet. From 20% to 40% of the gait cycle, the hip joint angles were nearly identical for both sides and feet. After 40% of the gait cycle, the hip joint angle continued to be more flexed on the side of amputation than on the intact side for both prosthetic feet.

Knee Joint Angles

The ensemble averages for knee joint kinematics of the intact and prosthetic limbs are depicted in Figure 2b. The knee joint angle for the intact limb demonstrated a flexion-extension-flexion pattern during stance that is typical of norrnal gait. Conversely, the ensemble averages for the prosthetic limb for each type of foot clearly lacked this typical triphasic pattem. Instead, the knee was maintained predominantly in a fully extended position. During swing, the prosthetic side knee joint was flexed approximately 10 degrees more than the intact side knee joint.

Vertical Ground

Reaction Forces

Figure 3a shows the ensemble averages for the vertical component of the ground reaction force. In general, the vertical ground reaction force curve generated during a single stance phase can be described as a composite of two "peaks" that occur chronologically: an initial maximum peak (F1), followed by an intermediate trough (F2), and a second maximum peak F3).21 Although it was not clear in this study which condition had the greatest duration of loading (expressed as time lapsed to reach F1) during the early phases (0%-5%) of stance, it appears that the intact limb duration exceeded that of the prosthetic limb while using the CC II foot, which in turn exceeded that of the prosthetic limb while using the SACH foot. The vertical ground reaction force reached F1 slightly earlier in stance in the intact limb than in the prosthetic limb while using either type of foot. The F1 peak was greater for the SACH foot than for the CC II foot, but it occurred later in the stance phase. The F2 troughs for the prosthetic limb while using either type of foot were slightly higher in magnitude and occurred slightly later than for the intact limb. The F3 peaks were lower in magnitude than the Fl peaks for both sides and both prosthetic feet. The magnitude of the F3 peak was greatest for the intact limb; these magnitudes were virtually identical for both prosthetic limb conditions. The duration of unloading (expressed as the time lapsed from F3 to 100% of stance) appeared to be greatest for the prosthetic limb with the SACH foot, followed by the prosthetic limb with the CC II foot. The duration of unloading of the intact limb was only slightly shorter than that of the prosthetic limb with the CC II foot.

Anterior-Posterior Ground

Reaction Forces

All anterior-posterior ground reaction force ensemble averages depicted a braking, or negative-acceleration, phase followed by a propulsive, or positive-acceleration, phase (Fig. 3b). The peak force for the braking phase of the intact limb occurred at approximately 15% of stance and was of much greater magnitude than for either prosthetic limb condition. Neither prosthetic limb condition demonstrated a well-distinguished peak force during the braking phase. The durations of the braking phase were nearly identical for the prosthetic limb conditions, and they were approximately 5% longer than for the intact limb.

With the exception of the last 10% of stance, the intact limb demonstrated a higher force throughout the propulsive phase and a much higher peak propulsive force than did the prosthetic limb while using either type of foot. The peak propulsive forces for the two prosthetic limb conditions were approximately equal, but this force occurred slightly later in stance for the CC II foot than for the SACH foot.

Net Muscular Moments

About the Hip

The ensemble averages for the net muscular moment about the hip varied substantially between intact and prosthetic limbs (Fig. 4a). There were slight differences in the rate, peak moment, and time at which the peak moment occurred between prosthetic limb conditions. In general, the prosthetic side demonstrated a large extensor moment, about twice the magnitude of that of the intact limb, which peaked at approximately 15% of stance and continued until approximately 50% of stance. For the remainder of stance, the net muscular moments about the hip for the prosthetic side remained close to zero. The net muscular moment about the hip on the intact side began with a slight flexor moment and progressed rapidly to an extensor moment that peaked at approximately 10% of stance. It then dropped to nearly zero at 15% of stance, where it remained until late stance, at which time a flexor moment was produced.

Net Muscular Moments

About the Knee

At the knee as well as the hip, the net muscular moments were quite similar for the two prosthetic limb conditions and quite different from the intact limb (Fig. 4b). The two prosthetic limb conditions demonstrated a flexor moment throughout the entire stance phase. The knee of the intact limb demonstrated an initial flexor moment, followed by a larger extensor moment that peaked at approximately 20% of stance and continued until approximately 35% of stance. The remainder of stance on the intact limb demonstrated a flexor moment of greater magnitude than on the prosthetic side and reached a peak at approximately 70% of stance.

Center of Pressure

Regardless of side or condition, the center of pressure under the foot of the supporting limb moved anteriorly a distance of approximately 210 mm from heel-strike to toe-off (Fig. 5). The center of pressure under the intact foot was more anterior than that of either prosthetic foot for the duration of stance. Until 15% of stance, the center-of-pressure values for the SACH and CC Il feet were similar. Between 15% and 80% of stance, the center of pressure for the SACH foot was more anterior than that of the CC II foot. At 80% of stance, the centers of pressure for the SACH and CC II feet united.

Prosthetic Foot Function

In the sagittal plane during stance, the CC II foot ankle joint underwent a greater excursion (4.8 [degrees]) than did the SACH foot ankle joint (2.3 [degrees]) (Fig. 6a). Ankle joint angular velocities were also greater in both the negative and positive directions for the CC II foot (Fig. 6b). The curves showing the net prosthetic moment about the ankle were similar for the two prosthetic feet (Fig. 6c), although plantar-flexor moments increased more rapidly in the SACH foot.

The power profiles for the two prosthetic feet were virtually identical for the first 35% of stance (Fig. 6d). From 35% to almost 80% of stance, the power values for the CC II foot were more negative than those of the SACH foot, indicating a greater rate of energy storage. From 80% to 100% of stance, the power values for the CC II foot were more positive than those of the SACH foot, indicating a greater rate of energy return. During the latter 20% of stance, the peak power for the CC 11 foot was 19 W as compared with 5 W for the SACH foot.

The work performed by the different prosthetic feet can be determined by integrating the power profiles with respect to time (Tab. 4).[10] During the negative (energy-storing) portion of the power profile, the SACH foot performed only 26% of the work performed by the CC II foot. During the positive (energy-returning) portion of the power profile, the SACH foot performed only 50% of the work performed by the CC II foot. The CC II foot returned 57% of its stored energy as compared with a 30% return by the SACH foot. The net work performed by the CC Il foot was 19% more negative than that of the SACH foot.

Discussion

Many of the results, specifically stride characteristics and angular kinematics for both the intact and prosthetic limbs, are typical for subjects with BK amputations.[22] The results of the kinetic analysis of the intact limb are fairly typical of normal ambulation.[23]

Prosthetic foot type had no effect on the stride characteristics or, except for prosthetic ankle joint angle, on the kinematic gait variables of the person with BK amputation in this study. Because the angular accelerations of her anatomical joints and the timing of her gait pattern were consistent across feet and because her mass remained relatively constant under both conditions, we can reasonably assume that she was imparting similar loads to the two prosthetic feet during the stance phase. Therefore, the variables that directly reflected prosthetic foot performance, namely center of pressure, ankle joint angle, net prosthetic moment about the ankle, and ankle joint power, probably reveal differences between the feet related to their energy-storing and energy-returning characteristics.

As previously noted, the SACH and CC II feet center-of-pressure values were very similar for the first 15% of stance, although the rate of the vertical ground reaction force increase in early stance was faster for the CC II foot and the F1 peak was higher for the SACH foot. It is during this phase that the cushioned heel of the SACH foot and the cushioned heel and energy-storing keel of the CC II foot compress. Differences in deformation characteristics between the two feet would directly affect these measures. Beyond 15% of stance, the center of pressure of the SACH foot was more anterior than that of the CC II foot. This may result from the weight transference onto the deformable components of the CC II foot as compared with the rigid keel of the SACH foot. Associated with the more anterior center of pressure under the SACH foot was a greater knee flexor net muscular moment through mid-stance and a greater hip extensor net muscular moment from 20% to 30% of stance onward. The prosthetic knee, however, was in a fully extended position. Thus, the net muscular flexor moments recorded throughout stance may not have a muscular origin, but rather may be derived from resistance of the joint itself.

Both feet were shown to store and return energy during stance. Both energy storage and energy return, however, were greater in the CC II foot than in the SACH foot. The CC II foot also dissipated a greater amount of energy as compared with the SACH foot, although the CC II foot's overall energy-returning efficiency was greater.

The significance to the person with a BK amputation of the energy-storage and energy-return capabilities of these prosthetic feet is of primary importance in determining their clinical usefulness. One way to make such an assessment relies on the concepts of gravitational potential energy (PE) and kinetic energy (KE). Recall that

(1)             PE = mgh
(2)             [Delta] PE = W
(3)             KE = 0.5 [mv.sup.2]
(4)             [Delta] KE = W

where m = body mass, g = gravitational constant, h = height of the body's center of mass, W = positive work (ie, energy returned), and v = average velocity of gait.[10] That is, the energy returned by the prosthetic foot can have two possible effects on the body: that of raising the center of mass (ie, increasing [Delta] PE) and that of increasing average velocity of gait (ie, increasing [Delta] KE). Substitution into equation (2) of experimentally obtained values for m and W, and assuming that [h.sub.initial] is equal to 1 m, shows that under the conditions of level, nonvigorous gait, the energy returned by the SACH foot was sufficient to raise the body's center of mass 0.67 mm, whereas that returned by the CC II foot was sufficient to raise the center of mass 2.57 mm. Substitution into equation (4) of experimental values for m, [V.sub.initial], and W shows that the energy returned by the SACH foot was sufficient to increase average velocity by 115 mm/s, whereas that returned by the CC II foot was sufficient to increase average velocity by 224 mm/s. As the average external calibration error for the kinematic data-acquisition system was 2.06 mm, it is likely that such small effects were barely appreciated experimentally. Furthermore, these estimated effects on [Delta] PE and [Delta] KE are probably not clinically significant, even in the case of the CC II foot.

Before we can conclude, however, that the CC II foot is not a clinically effective energy-saving prosthetic foot, several points merit consideration. First, the person who participated in this study received no formal instruction or training in the use of the CC II foot. This discrepancy in experience with the two feet was intentional for the purpose of this study, because the research design relied on consistent stance phase loading of the feet by the person in her role as a part of the laboratory instrumentation. We purposely avoided introducing greater variability into the instrumentation by minimizing the person's use of the CC II foot prior to her participation in this study. With training and experience using the CC II foot, however, the person would be expected to utilize the energy-storing capabilities of the CC II foot more advantageously. Second, because ankle joint angle was determined as a projection into the laboratory YZ plane, any rotations of either the foot or shank segments out of that plane would have resulted in an underestimate of ankle joint angular displacement and hence of its derivatives angular velocity and acceleration. Thus, net prosthetic moment about the ankle and ankle joint power would subsequently have been underestimated. Although our use of a single person with demonstrated consistency of gait kinematics supports conclusions based on the relative effects of the two prosthetic devices, the calculations estimating the absolute changes in PE and KE may be lower than the actual values obtained for both the SACH and the CC 11 feet. Finally, under more strenuous loading conditions, such as vigorous walking, running, or jumping, the greater efficiency of energy return by the CC II foot may well improve functional performance. Indeed, it was for such activities that the concept of the energy-storing foot was conceived.

Conclusions

The person who participated in this study showed no discernable change in overall gait characteristics, as measured by stride characteristics and angular kinematics of the hip and knee of both limbs while using either of the two prosthetic feet. She was therefore an ideal instrument to use in the mechanical testing of the different prosthetic foot designs. Differences evoked from the use of either the SACH or the CC II foot were discernable when measurements were clearly specific to keel design and appear to be associated with the energy-storage and energy-return characteristics of the feet. The CC II foot was shown to return energy at 57% efficiency as compared with only 30% efficiency for the SACH foot.

Despite the superior efficiency of the CC II foot as an energy-storing and energy-returning device, estimates of the change in PE and KE caused by the energy returned during level, nonvigorous gait are clinically unremarkable. Future studies are needed to investigate similar effects during more strenuous activities.

[TABULAR DATA OMITTED]

(*) Model and Instrument Development, 861 Poplar Pl S. Seattle, WA 98144. (dagger) Flex-foot Inc, 14 Hughes. B-201, Irvine, CA 92714. (double dagger) Ohio Willow Wood Co, 15441 Scioto Darbv Rd, PO Box 192, Mount Sterling. OH 43134. (~~) Oxford Metrics Inc, 14206 Carlson Cir, Tampa, FL 33625. (#) AMTI, 151 California St, Newton, MA 02158. (**) ADTECH, 2002 Ruatan St, Adelphi, MD 20783.

References

[1] Goh JCH, Solomonidis SE, Spence WD, Paul JP. Biomechanical evaluation of SACH and uniaxial feet. Prosthet Orthot Int. 1984;8:147-154, [2] Burgess EM, Hittenberger DA, Forsgren SM, Lindh D. The Seattle prosthetic foot: a design for active sports. Orthotics and Prosthetics 1983;37:25-32. [3] Michael J. Energy storing feet: a clinical comparison. Clinical Prosthetics and Orthotics 1987;11:154-168. [4] Edelstein JE. Prosthetic feet: state of the art. Phys Ther. 1988:68:1874-1881. [5] Carbon Copy II Product Literature. Mt Sterling, Ohio: Ohio Willow Wood Co. [6] Torburn L, Perry J, Ayyappa E, Shanfield SL. Below-knee amputee gait with dynamic elastic response prosthetic feet: a pilot study. J Rehabil Res Dev. 1990;27:369-384. [7] Daher RL. Physical response of SACH feet under laboratory, testing. Bulletin of Prosthetic Research. 1975;10-23:4-50. [8] Kabra SG. Narayanan P. Ankle-foot prosthesis with articulated human bone endoskeleton: force-deflection and fatigue study. J Rehabil Res Dev. 1991;28:13-22 [9] Kabra SG, Narayanan R. Equipment and methods for laboratory testing of ankle-foot prostheses as exemplified by the Jaipur foot. J. Rehabil Res Dev. 1991;28:23-24. [10] Marion JB. Classical Dynamics of Particles and Systems. 2nd ed. New York, NY: Academic Press Inc; 1970:45-91. [11] Robertson DGE, Winter DA. Mechanical energy generation, absorption and transfer amongst segments during walking. J Biomech. 1980;13:845-854. [12] Winter DA. Sienko SE. Biomechanics of below-knee amputee gait. J Biomech. 1988; 21:361-367. [13] Gundersen LA, Valle DR, Barr AE, et al. Bilateral analysis of the knee and ankle during gait: an examination of the relationship between lateral dominance and symmetry. Phys Ther, 1989;69:640-650. [14] Kepple TM, Stanhope SJ, Lohman KN, Roman NL. A video-based technique for measuring ankle-subtalar motion during stance. J Biomed Eng. 1990;12:273-280. [15] Stanhope SJ, Jarrett MO. A position adjustable force plate mounting system. Proc Annu Int Conf IEEE Eng Med Biol Soc. 1988;10 (part 2):655. [16] Stanhope SJ, Kepple TM, McGuire DA, Roman NL. A kinematic-based technique for event time determination during gait. Med Biol Eng Comput. 1990;28:355-360 [17] Winter DA. The Biomechanics and Motor Control of Human Gait. Waterloo, Ontario, Canada: University of Waterloo Press; 1987. [18] Sutherland DH, Cooper L, Daniel D. The role of the ankle plantarflexors in normal walking. J Bone Joint Surg [Am]. 1980;62: 354-363. [19] Godfrey K. Statistics in practice: comparing the means of several groups. N Engl J Med. 1985;313:1450-1456. [20] Miller RG. Simultaneous Statistical Inference. 2nd ed. New York, NY: Springer-Verlag New York Inc; 1981:15-16. [21] Chao EY, Laughman RK, Schneider E, Stauffer RN. Normative data of knee joint motion and ground reaction forces in adult level walking. J Biomech. 1983;16:219-233. [22] Breakey J. Gait of unilateral below-knee amputees, Orthotics and Prosthetics. 1976;30:17-24. [23] Bresler B, Frankel JP. The forces and moments in the leg during level walking. Trans Am Soc Mech Eng. 1950;72:27-36.

AE Barr, PT, was Staff Physical Therapist, Department of Rehabilitation Medicine, Warren G Magnuson Clinical Center, National Institutes of Health, Bethesda, MD 20892, when this study was completed. She is now Senior Research Physical Therapist, Motion Analysis Laboratory and Department of Rehabilitation Services, The Hospital for Special Surgery, 535 E 70th St, New York, NY 10021 (USA). Address correspondence to Ms Barr. K Lohmann Siegel, PT, was Senior Staff Physical Therapist, Department of Rehabilitation Medicine, Warren G Magnuson Clinical Center, National Institutes of Health, when this study was completed. She is now Research Physical Therapist, The Biomechanics Laboratory, Department of Rehabilitation Medicine, Warren G Magnuson Clinical Center, National Institutes of Health. JV Danoff, PhD, PT, is Associate Professor, Department of Physical Therapy, Howard University, Washington, DC, and Research Consultant, Department of Rehabilitation Medicine, Warren G Magnuson Clinical Center, National Institutes of Health. CL McGarvey III, PT, is Director of Physical Therapy, Department of Rehabilitation Medicine, Warren G Magnuson Clinical Cenler, National Institutes of Health. A Tomasko, PT, was Physical Therapy Student Intern, Department of Rehabilitation Medicine, Warren G Magnuson Clinical Center, National Institutes of Health, when this study was completed. He is now Staff Physical Therapist, Avon Physical Therapy Associates, 54 W Avon Rd, Avon, CT 06001. I Sable, CPO, is in private practice at Capitol Orthopedics, 8200 Wisconsin Ave, Bethesda, MD 20814. SJ Stanhope, PhD, is Chief of the Biomechanics Section and Director of The Biomechanics Laboratory, Department of Rehabilitation Medicine, Warren G Magnuson Clinical Center, National Institutes of Health.