Link to Figures Show Figures
	Article Text Figures References
	Send Link Send HTML Article

Subjective and Objective Analysis of an Energy-Storing Prosthetic Foot

Michael R. Menard, M.D., Ph.D.
D. Duncan Murray, M.D., FRCPC

Introduction

Energy-storing prosthetic feet were designed to enable the amputee to run. They have been fitted for amputee athletes, and are also sought by and recommended for less demanding users, such as geriatric amputees. A variety of energy-storing prosthetic feet are now being marketed, but their relative merits are unclear. Therefore, we wished to determine the clinical situations in which an energy-storing prosthetic foot is superior to a conventional prosthetic foot, and to determine whether any differences exist between different energy-storing prosthetic feet.

We felt that we required an objective criterion to justify the greater expense of some of the newer prosthetic elements. In normal gait analysis, a postulate states that the excursions of the body's center of gravity from a smooth steady progression should be minimized. This postulate led to the identification of six major determinants of gait, and the identification of the role of the toes, foot, ankle, calf, knee, thigh, hip, pelvis, torso, and upper extremities.^3,4 This analysis can-not be applied directly to prosthetic gait^5,6 since below-knee amputees lose the natural foot and ankle, but retain the knee which was "designed" to work with them. The below-knee prosthesis replaces the overall structure of the missing part, but has a simpler mechanical function. In a hierarchy of complexity, the prosthesis provides, for standing, a connection to the ground for static support. During walking, it must absorb shock at heelstrike, then provide a rigid "foot lever" as the body passes over it during stance in order to impede the progression of the shank and to provide an appropriate base by which knee extension can propel the body forward. In addition, it should accommodate to uneven ground. For running, the below-knee prosthesis should provide additional propulsion of the limb by active plantar flexion prior to toe-off.

Prior to the development of the Seattle FootŪ , a compressible heel served for shock absorption and a rigid heel provided support from foot-flat to toe-off (e.g., the SACH foot). However, the amputee couldn't run using this prosthesis.⁷ The energy storing foot is designed to provide an additional active propulsion of the limb at toe-off, which is needed for running, but not for walking. Energy-storing prosthetic feet cannot adapt to the complete range of ambulation on level ground; the user must choose high, medium, or low demand during construction of the prosthesis. They also cannot adapt to situations in which shock absorption without return of energy is required, such as in walking downhill or downstairs.

During our practice, we observed that with energy-storing prosthetic feet, as with other prostheses and orthoses, correct use sometimes did not go hand in hand with user satisfaction. However, the success of a prosthetic fit is ultimately judged by its acceptance by the user and the associated increase in the user's functional capacities. As part of an ongoing analysis of prosthetic gait, we posed the following question: Do those who don't like their prosthesis walk poorly on it, and do those who have poor gaits dislike their prosthesis? To answer this question, we selected a cohort of otherwise healthy, working, traumatic below-knee amputees. These people had been previously fitted successfully, and had recently received a Flex-FootŪ at their usual time of fitting for a replacement prosthesis. We compared their score on an opinion survey of the Flex-FootŪ with their score on gait analysis by trained observers.

In addition, we performed biomechanical measurements of stance duration and ground reaction forces on all subjects. We wished to establish the typical values for these variables in the Flex-FootŪ population, and to characterize the pattern of energy storage and return. We also wished to identify subjects who were markedly different from the mean. We wondered if such "outliers," who had different gait timing and different limb forces, would be observed to walk poorly (low gait score) and/or would dislike their prosthesis (low opinion score). Through such inquiry, we hoped to identify objective measures which would reflect the fundamental determinants of prosthetic gait, and to determine whether quantitative measures could be of use in our day-to-day clinical evaluations of prosthetic gait.

Methods

The subjects chosen for study were all unilateral below-knee amputees considered for fitting with the Flex-FootŪ at the B.C. Workers' Compensation Board Amputee Clinic from the time the Flex-FootŪ first became available to us in June 1985 to six months before the evaluations were carried out. We wanted all subjects to have at least two months experience with a satisfactorily fitted Flex-FootŪ . Thirty-one subjects were eligible. Of these, 26 were fitted with the Flex-FootŪ. Of those fitted, 22 were available for study. Of the remaining four, two refused to participate, one was lost to followup, and one, who had a pre-existing impairment in the non-amputated limb, was excluded.

The opinion survey was similar to that used in our previous study.⁸ It included demographic data, general health, medication use, nature of the injury resulting in the amputation, whether revision of the amputation was necessary, pain (contemporary and in comparison to previous prosthesis), and ability to perform activities (contemporary and in comparison to previous prosthesis). A mail-in survey was followed up by a personal interview for clarification at the time of biomechanical testing. Four participants, who returned the mailed survey, were unable to attend for gait observation and biomechanical testing.

Video recordings of each subject were made to test for reliability and inter-observer variation. In order to assign a score to the gait of each subject, several commonly observed abnormalities were selected for static analysis and for each phase of gait.⁹ Static analysis involved six observations of the residual limb and socket, and four observations of the standing alignment. Analysis while walking consisted of four observations at heelstrike, five during early stance, six during midstance, five during late stance, and five during swing. Finally, four were observed walking up and down stairs and a ramp. Each of the 39 items received a score of "1" if normal and "0" if abnormal. Subscores and total scores were simple, unweighted sums of the item scores.

Biomechanical measurements were made in an established gait laboratory with a Kistler 9261A multi-component force platform and a Data General Micro-Nova MP2OO minicomputer. Each subject first executed "training" runs and was coached until the subject established a consistent, comfortable free walking speed, and consistently landed on the forceplate without targeting. Then data was collected with three runs of consistent speed being saved for the natural limb and three for the prosthetic limb. For the purposes of comparison, the data was normalized to 100 equal time intervals of stance for each subject. Data was transferred to a Macintosh computer for graphic depiction and numerical analysis in Excel.

Results

Opinions

With respect to overall response to the Flex-FootŪ , 15 (68%) felt their gait was improved, six (27%) felt there was no change, and one felt it was worse. Fourteen out of 20 users felt their recreational activity increased with the Flex-FootŪ while six observed no change. Twenty users could feel the dynamic action of the prosthesis and two did not respond to this question. Of the 14 users who stated a preference, six felt that the amount of dynamic action was appropriate, five preferred more, and three preferred less.

Nine out of 20 respondents experienced a decrease in limb pain, eight experienced no change, and three experienced increased pain. Five out of 18 respondents had a decrease in skin problems, 11 had no change, and two had more skin problems. Sixteen expressed no difficulties with the normal limb, while six stated some problems.

The subjects' impression of the effect on his capability to engage in various activities is summarized in Figure 1 . Overall, walking, jogging, and dancing showed the most improvement. Walking upstairs was improved, while walking downstairs was quite difficult for a number of users, but less troublesome for others. Similarly, walking on uneven ground was difficult for a number of users, but not for others. No one thought their gait was less smooth than before the Flex-FootŪ and 18 thought it was smooth. Twelve out of 22 felt their balance was improved, and 14 out of 22 felt their endurance was improved. Nineteen out of 22 had no mechanical problems. Two experienced breakage of the Flex-FootŪ assembly.

A numerical opinion score probably is less appropriate as a summary measure of user satisfaction than the direct question itself. Nevertheless, when we assigned + 1 to each activity which was improved with the Flex-FootŪ , 0 to each which was not changed, and -1 to each which was worse. The net score represents improvement in some items and worsening in others. No one felt that their overall ability to ambulate was worse with the Flex-FootŪ , and two felt that it was unchanged. The remainder of subjects felt that their ability to ambulate was better. Similarly, a numerical pain opinion scale, with negative values indicating a reduction in pain with the Flex-FootŪ , ranged from -2 to +2 (out of a maximum range of -9 to + 9), with a mean of -0.7. With one exception, the alterations in pain never showed improvement in one area or worsening in another.

Gait Observation

The distribution of scores from gait observation is shown in Figure 2 . The overall impression of the observers was that the members of this cohort walked very well. The average total score was 28.9 (range 22-34) out of a maximum of 39. The greatest variation in scores occurred at heelstrike and in early stance. A consistent observation was the occurrence of some degree of medial heel whip in all but one subject.

Quantitative Measurements

The forceplate measurements are summarized in Figure 3 ,Figure 4 ,and Figure 5 . The forces are expressed as the fraction of the body weight for each subject (1.0 equals 100% of the body weight). The wide solid line is the mean value for the natural limb, and the wide dashed line is the mean value for the prosthetic limb. The narrow solid and dashed lines indicate a standard deviation of + 1 and -1 from the mean value. The difference between the ground reaction force on the prosthetic and the natural side at corresponding times during stance were calculated for each subject, then averaged. The difference is the amount by which the force on the natural side exceeds the force on the prosthetic side.

For the vertical component (Figure 3) , positive values of the difference curve indicate a greater vertical force on the prosthetic side. The differences are large at heelstrike and late in stance. At heelstrike, the prosthetic side rises more gradually to peak load. The measurement technique cannot distinguish between greater energy of the natural side at impact and greater energy absorbing capacity of the prosthetic heel, but it clearly shows assymmetry at this stage of gait. Late in stance, the positive peak of the difference curve indicates that the natural side is pushing into the ground more at this stage when the heel has risen, and the later negative peak indicates that the prosthetic side is pushing more into the ground just before toe-off.

The mediolateral force (Figure 4) has been adjusted so that a positive force is one which pushes from the stance side toward the center of the body. This was done to facilitate comparison of the prosthetic and natural sides, since in some subjects, the force was in opposite, absolute directions for both limbs, and since not all amputations were on the same side. For the medio-lateral component, the forces are pushing in a medial direction except in the first 10% of stance. Therefore, except for the first part of stance, positive values indicate a greater force on the natural side pushing in a medial direction, and negative values indicate a greater force on the prosthetic side pushing in a medial direction. During the first part of stance, just after heelstrike, the forces are pushing in a lateral direction, which indicates the advancing limb has a slight lateral to medial component to its movement just before heelstrike. (This is probably due to the pelvic rotation that occurs as the limb advances; the acetabula are furthest from the line of advancement during midstance and closest to this line at heelstrike.³) Therefore, during the first 10% of stance, the negative value for the difference in the mediolateral components indicates that the prosthetic limb either has less of a lateral to medial component, or else absorbs it better at heelstrike (Figure 4) . The next two positive peaks in the difference curve indicate that a smaller lateral to medial push occurs on the prosthetic side, but the late negative peak indicates that the prosthetic side experiences a large push of this sort just before toe-off.

The anteroposterior force (Figure 5) is positive for a ground reaction force which decelerates the subject and negative for one which accelerates him. Again, the difference curve is more difficult to interpret because the meaning of the sign of the curve is different during the deceleration phase and the acceleration phase. During the deceleration phase, the decelerating force is smaller on the prosthetic side, but the proportion of stance over which deceleration occurs is longer. During the acceleration phase, the prosthetic side again delivers less force until just before toe-off.

The mean duration of stance on the prosthetic side was 0.670 seconds and on the natural wide was 0.683 seconds. This difference is statistically significant (p < 0.025 by the t statistic for paired data, two-tailed test). the ratio of the stance duration on the prosthetic side to that on the natural side ranged from 0.92 to 1.04 (mean 0.98). the mean walking speed was 1.40 m/sec (range 0.94 - .62 m/ sec).

Comparison of Evaluation Techniques

The opinion scores and the observation scores were compared to see if there was a concordance of low opinion and poor gait. This was difficult to determine because overall, the subjects liked their prosthetic fit and walked well. the scores cannot be compared numerically, because they are nonparametric, i.e., a score of "4" is not necessarily twice as good as a score of "2". therefore, we identified outliers in each category. In two subjects, the low opinion score was associated with a low gait observation score. Interestingly, these two subjects had the lowest subscores on the quality of the residual limb and socket. However, the subject with the next lowest residual limb subscore felt his activity level had improved and his pain had decreased, and he was observed to walk well. Overall, no pattern of correlation was found between the opinion scores and the observation scores. The quantitative measurements were analyzed for outliers as well. The deviation of the measured forces from the mean for the whole group was calculated (root mean square deviation) and the scores of two subjects were more than two standard deviations from the mean. Both subjects had low observation scores (22 out of 39); one had a low residual limb subscore (three out of six), but both felt their activity was greater and their pain was reduced. overall, there was no consistent pattern between opinion score, observation score, and forceplate score.

Discussion

The overall goal of our research is to understand the essential factors which determine correct prosthetic fit and use. The best method by which to analyze gait for clinical purposes has not yet been determined. previous force plate studies of prosthetic gait have examined few subjects and have not made quantitative comparisons.^2,10,11

For amputees, the usual method of analysis is visual observation by the prosthetist, with the overall aim of making the gait as smooth and symmetrical as possible. Both static and dynamic alignments are performed according to accepted rules,⁹ but the skill and experience of the prosthetist determine the outcome. An important part of the art of prosthetics is the therapeutic relationship the prosthetist establishes with his patient. Clinical relationships are known to have a powerful influence on the compliance and satisfaction of the patient. ² Therefore, it has the potential to be a powerful confounding factor in clinical gait analysis.

Some workers argue that visual observation is inadequate as a diagnostic method for prosthetic gait because few variables can be evaluated. ¹³ However, most normal gaits look similar in terms of the usual quantitative measurements of cadence, joint angles, and ground reaction forces. ¹⁴ This is because these "external" variables reflect the net effects of many different factors: the patient's endowment of bone and muscle; habits and motivation; any pathological processes which are present; and compensatory activities. In order to discover the biomechanical mechanism of a particular gait, the forces and moments at the joints must be estimat ed. This requires the use of a link-segment model to calculate all forces and moments from simultaneous recordings of limb position and ground reaction forces. In amputees, the stump-socket articulation would have to be incorporated into the model.

Most practicing prosthetists probably would feel that such a sophisticated analysis is unnecessary. Overall our subjects walked very well and were very satisfied with their prosthesis and fit. According to this reasoning, the clinical models of normal and pathological gait actually used by the prosthetist⁹ do not require the amount, type, or precision of data which quantitative analysis is capable of providing. A question arises, however, concerning problems in the true value of the current crop of "high-tech" prosthetic components: What benefits can energy-storing prosthetic feet provide, and is one appreciably different from another?

In order to address this problem, we followed Rose¹⁶ in formulating a clinical model of the gait under study, and performed only those quantitative measurements which were needed to test it. The measurements were only to be as accurate and complete as was necessary to obtain a useful answer.

Concerning the analysis of the Flex-FootŪ, the clinical problem is that the amputees appear to adopt a medial heel whip, and to have some difficulty with uneven ground and with descent of stairs.

Our initial hypothesis was that there was too much spring in the Flex-FootŪ for the ordinary user. However, only three of the subjects thought the springiness of their prosthesis was excessive, and they did not consistently have poor gait scores or markedly asymmetrical gaits as measured by stance times. Therefore, we went to a more sophisticated measurement technique, the force platform, after realizing that the ground reaction forces would not be determined solely by the prosthesis, but by the entire locomotor apparatus on both the prosthetic and natural sides. The subtle difference we observed between the Flex-FootŪ side and the natural side was the delivery of an accelerating force very late in stance on the FlexFootŪ side, at a stage of stance where the natural side was delivering very little force (Figure 3) . In retrospect, it seemed obvious that an energy-storing prosthetic foot should behave in this manner, as a spring with its force proportional to the amount it is bent. This impulse at toe-off makes the prosthesis feel "alive". However, the calf and foot are not normally used to propel the limb with an active push-off by the toes during walking; during walking the posterior calf muscles are electrically silent during the last approximately 10% of stance, and the anterior calf muscles are dorsiflexing the foot to clear during swing.^3,17,18,19 Then, the medial heel whip would serve as a compensatory movement by the rest of the limb to dissipate this unnecessary, propulsive force.

However, during running, the posterior calf muscles are used for propulsion of the limb, and a strong propulsive force during late stance is appropriate and useful. Indeed, provision of such a force was the goal of the original energy-storing foot design, the Seattle Foot ?²⁰

Conclusion

Energy-storing prosthetic feet provide a propulsive force very late in stance during walking, while the natural limb does not. This makes the limb feel "lively" and is appreciated by almost all users, but, it appears that it must be dissipated by the adoption of a medial heel whip. It is not known whether these relatively subtle events have any deleterious effect on the stump, the joints of the amputated and intact limbs, or the total energy requirement for gait. Our work so far has revealed associations, not causes. We hypothesize that a very small amount of energy storage and return would be just as pleasing to the user during walking, but would not induce an appreciable amount of medial heel whip. We also hypothesize that the forces and moments in the limbs might be harmful if the amount of stored energy which must be dissipated is excessive. If these hypotheses can be verified, they would provide the basis for rationale for an energystoring prosthetic foot prescription based on the habitual type and intensity of activity of the user, and on the range and pattern of energy storage and return available in a particular prosthesis.

Acknowledgments

Subjects were examined at the Amputee Service of the Workers' Compensation Board by Ian Dukes, under the supervision of the Director A.B. Kennard, M.D., and at the Biomechanics Laboratory, University of British Columbia, and the Department of Physical Education, under the supervision of David Sanderson, M.D.

Michael R. Menard, M.D., Ph.D. and D. Duncan Murray, M.D., FRCPC, practice at the University of British Columbia, Canada.

Murray is also the Clinical Director of Prosthetic and Orthotic Services at University Hospital, Shaughnessy Site, Room A170, 4500 Oak Street, Vancouver, B.C., Canada V6H 3N1.

References:

1. Michael, J., "Energy Storing Feet: A Clinical Comparison," Clin. Prosthet. Orthot., 11, 1987, pp.154-168.
2. Wagner, J., S. Sienko, T. Supan and D. Barth, "Motion Analysis of SACH vs. Flex-FootŪ in Moderately Active Below-Knee Amputees," Clin. Prosth. Orthot., 11,1987, pp.55-62.
3. Inman, V.T., H.J. Ralston and F. Todd, "Human Walking," 1981, Williams & Wilkins, Baltimore, Maryland.
4. Tesio, L., P. Civaschi and L. Tessari, "Mo tion of the Center of Gravity of the Body in Clinical Evaluation of Gait," Am. J. Phys. Med., 64, 1985, pp.57-70.
5. Breakey, J. "Gait of Unilateral Below-Knee Amputees," Orthotics and Prosthetics, 30, 1976, pp.17-24.
6. Skinner, H.B. and D.J. Effeney, "Gait Analysis in Amputees," Am. J. Phys. Med., 64, 1985, pp.82-89.
7. Kegel, B., M.L. Carpenter and E.M. Burgess, Functional Capabilities of Lower Extremity Amputees," Arch. Phys. Med. Rehab., 59, 1978, pp.109-120.
8. Murray, D.D., W.J. Hartvikson, H. Anton, L. Hommonay and N. Russell, "With a Spring in One's Step," Clin. Proth. Orthot., 12, 1988, pp. 128-135.
9. Mensch, G. and P.M. Ellis, "Physical Therapy Management of Lower Extremity Amputations," 1986, Aspen Publishers, Rockville, Maryland.
10. Suzuki, K., "Force Plate Study of the Artificial Limb Gait," J. Jap. Orthop. Ass., 46, 1972, pp.503-516.
11. Bach, T.M. and D.P. Wooley, "The Flex- Foot: Initial Evaluations at Lincoln Institute," Aust. Orthot. Prosthet. Mag., 1, 1986, pp.5-8.
12. Haynes, R.B., "A Critical Review of the Determinants of Patient Compliance with Therapeutic Regimens," Compliance with Therapeutic Regimens, 1976, D.L. Sackett and R.B. Haynes, Johns Hopkins University Press, Baltimore, Maryland.
13. Saleh, M. and G. Murdoch, "In Defense of Gait Analysis," J. Bone Joint Surg., 67B, 1985, pp.237-241.
14. Murray, M.P., A.B. Drought and R.C. Kory, "Walking Patterns of Normal Men," J. Bone Joint Sarg., 46A, 1964, pp.335-360.
15. Winter, D.A., "Biomechanics of Human Movement," John Wiley & Sons, Toronto, Canada, 1979.
16. Rose, G.K., "Clinical Gait Assessment: A Personal View," J. Med. Eng. Tech., 7, 1983, pp. 273-279.
17. Simon, S.R., R.A. Mann, J.L. Hagy and L.J. Larsen, "Role of the Posterior Calf Muscles in Normal Gait," J. Bone Joint Surg., 60A, 1978, pp.465-476.
18. Mann, R.A. and J.L. Hagy, "The Function of the Toes in Walking, Jogging, and Running," Clin. Orthop. Rel. Res.,142, 1979, pp.24-29.
19. Sutherland, D.H., L. Cooper and D. Daniel, "The Role of the Ankle Plantar Flexors in Normal Walking," J. Bone Joint Surg., 62A, 1980, pp.354-363.
20. Hittenberger, D., "The Seattle Foot," Orthotics and Prosthetics, 40, 1986, pp.17-23.