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

Gait Analysis and Energy Cost of Below- Knee Amputees Wearing Six Different Prosthetic Feet

Daryl G. Barth, BS, CPO
Laura Schumacher, BS, CP
Susan Sienko Thomas, B.Sc.

Introduction

New prosthetic materials and designs have broadened the range of prosthetic feet available. As a result, it is becoming more difficult for prosthetists and prescribing physicians to choose which foot is best for individual amputees. Much of the research evaluating the dynamics of prosthetic feet is subjective (1-3). Quantitative research on below-knee amputee gait has been performed in the following areas: dynamic evaluation of the foot through motion analysis, evaluation of the forces created by and acting on the body when wearing a prosthetic foot, and amputees' energy cost when using various foot designs (2,4-9).

The purpose of this study is to scientifically measure the dynamic gait characteristics of the amputee by using motion analysis and to measure the energy cost of the below-knee amputee when wearing the SACH, S.A.F.E. II, Seattle Lightfoot, Quantum, Carbon Copy II and Flex-Walk prosthetic feet (10). The analysis will separate the subjects into the Vascular Group and the Traumatic Group for more specific evaluation of activity level by type of amputation. Research results will provide quantitative insight to facilitate proper prosthetic foot selection.

Dynamic evaluation of the foot through motion analysis has determined that the gaits of below-knee amputees differ from those of "normal" individuals. Specifically, linear measurements, range of motion and foot-floor contact forces differ. During normal gait, the motion about the joints and the magnitude of the foot-floor contact forces between the right and left lower extremities are relatively symmetrical (11). However, during amputee gait, the foot-floor contact forces and range-of-motion about the ankle, knee and hip reveal asymmetry between the sound-side and the prosthetic-side (2,5,8,12)

Prosthetic component design, selection and alignment of the amputee's prosthesis are all directed toward obtaining optimal gait. Prosthetists are taught to achieve optimal sound-side and prosthetic-side symmetry; and investigators have established that the use of symmetry is a good analytic method to evaluate pathologic gait (12). Presently, only limited information describes how the newer prosthetic feet perform dynamically in achieving this optimal, symmetrical gait. This study evaluates the effect of six different prosthetic feet on prosthetic-limb and sound-limb symmetry. Ankle joint angles and foot-floor contact forces between the prosthetic and sound limbs are quantitatively measured and analyzed.

Some foot manufacturers have suggested that their prosthetic feet reduce the amputee's energy cost. This energy cost savings can be evaluated by comparing different prosthetic feet. Researchers have found that each amputee walks at a self-selected optimal, energy-efficient speed (3,6,13,14,15). By evaluating the amputee at self-selected velocity for each foot, the energy cost for that foot can be determined. Previous studies have shown that below-knee amputees have a higher energy cost than normal individuals (3,13,14,16,17). Nielsen demonstrated minimal differences between the SACH and Flex-Foot prosthetic feet when walking at slow velocities; however, ambulation at higher velocities conserved energy when wearing the Flex-Foot (6). Furthermore, investigators have detected energy cost differences among vascular and traumatic below-knee amputees (13,14). This study evaluates the differences in energy cost among prosthetic feet and the differences in energy cost between the Vascular and Traumatic groups.

Method

Six male unilateral right below-knee amputees were selected as subjects for this study (see Table 1 ). Each subject had an existing below-knee definitive prosthesis with a soft, removable liner (insert). All had been wearing a variant of a patellar-tendon-bearing definitive prosthesis for at least two years. None of the subjects had residual limb pain, swelling or pressure sores, and none exhibited major gait deviations. Residual limb length for all subjects was at least 6.0 percent of total body height, but did not exceed 8.5 percent. These criteria reflect the importance of maintaining a consistent limb length among subjects when evaluating energy cost (13,18).

The six subjects tested were divided into two groups: the Vascular Group and the Traumatic Group. The Vascular Group was comprised of three subjects diagnosed with diabetic or non-diabetic peripheral vascular disease. They had a mean age of 64 years, a mean time since amputation of 5 years and a mean residual limb length of 13.6 cm, which averaged 7.6 percent of total body height. The Vascular Group was considered to have a lower activity level. The Traumatic Group was comprised of three subjects who did not suffer from peripheral vascular disease. They had a mean age of 39.3 years, a mean time since amputation of 21.6 years and a mean residual limb length of 13.1 cm, equaling 7.3 percent of total body height. The Traumatic Group was considered to have a higher activity level.

A test prosthesis was fabricated for each subject. In making this prosthesis, the soft liner was removed and the subject's existing definitive laminated socket was duplicated using Otto Bock duplicating foam. A polypropylene socket was vacuum-formed over the model. The existing definitive soft liner was worn in the test prosthesis, allowing subjects to use their comfortable, well-fitting soft liner, the same thickness of prosthetic socks and current method of suspension in the test prosthesis. Otto Bock titanium endoskeleton components were attached to the polypropylene socket and were used to make alignment adjustments.

The criteria used in selecting the six feet to be tested were that they are currently used in clinical practice, attach with an ankle bolt and adapt to the footplate (Otto Bock 2R31) of the endoskeletal test prosthesis. This study analyzes the prosthetic foot only; therefore, dynamic ankle and shin components were not evaluated.

Individual feet were chosen following each manufacturer's suggested guidelines for amputee weight and activity level (10). Each subject tested the feet in random order. The foot to be tested was attached to the test prosthesis at the ankle and optimally aligned by the prosthetist. Since it has been determined that prosthetic alignment has an effect on gait symmetry, each foot was worn on the test prosthesis for three weeks with opportunity for alignment changes if deemed necessary (8). At the end of three weeks, the foot was quantitatively evaluated using motion analysis. Energy cost was measured using a treadmill on the same day. Following the entire testing procedure for each foot, the foot was removed at the ankle bolt and replaced with the next foot, optimally aligned and the process repeated until each subject had tested all six feet.

The objective evaluation of the amputee's gait using each of the feet was performed in the Motion Analysis Laboratory at Southern Illinois University School of Medicine. Passive retro-reflective markers were attached to the patient at the shoulder, elbow, wrist, ASIS, lateral epicondyle of the knee, lateral malleolus, 5th metatarsal and calcaneous, bilaterally. Stick markers were applied anteriorly at the femur and laterally in line with the tibia, bilaterally, with an additional marker at the L5-S1 level. Surface electrodes were placed at the rectus femoris, vastus lateralis, medial hamstring and lateral hamstring, bilaterally. Using a five-camera VICON motion analysis system, the subjects walked at their self-selected comfortable walking speed for three trials. Immediately following the analysis, velocity, cadence, stride length and single-limb stance times were calculated to determine the average speed to be used during the oxygen consumption test. In the motion analysis, joint motion was evaluated to allow for a dynamic comparison of the range-of-motion. AMTI force plates measured the dynamic forces, allowing a comparison of the sound and prosthetic limbs. Electromyographic data was collected to determine the differences in muscle activity patterns when different prosthetic feet were worn.

Energy cost was evaluated in the Memorial Medical Center Pulmonary Function Laboratory. Equipment included a heavy-duty treadmill (0.8 to 10.0 mph), metabolic measurement cart, valve with rubber mouthpiece and nose clip, and ECG electrodes. The metabolic measure cart allows noninvasive measurement of basic physiologic responses to exercise, including oxygen uptake, carbon dioxide production, ventilation and heart rate.

Subjects were tested by registered respiratory therapists using standard procedures for exercise testing. Resting measurements were recorded for a minimum of 10 minutes. ECG electrodes were positioned, and the subject stepped onto the treadmill. Treadmill speed started at 0.8 mph and increased in 0.1 mph increments to the subject's self-selected walking speed determined in motion analysis. Measurements were recorded for 10 minutes at steady-state exercise intensity as determined by visual monitors plotting oxygen uptake against time. Heart rate was monitored continuously for safety and to support steady-state occurrence. Energy cost was calculated by using data from a three-minute interval during steady-state exercise (15,16,19). It was expressed in milliliters of oxygen uptake per kilogram of body weight per meter traveled: Energy Cost ml O₂/ Kg-m (6,9,13,20). This allowed a comparison of energy cost between the Vascular and Traumatic groups, as well as the energy cost of the amputee when wearing different prosthetic feet. It is outside the scope of this article to discuss the controversy of how to study gait efficiency and energy cost. The majority of the literature reviewed suggests the best method to evaluate energy expenditure is the energy cost per distance traveled and, consequently, is the method used in this study (6,13,14,20).

Results

In motion analysis, significant differences were found in linear measurements (velocity, cadence, stride length and single-limb stance), late-stance ankle dorsifiexion, the amount of ankle dorsiflexion change from early to late stance, weight acceptance foot-floor reaction forces, and right and left step lengths. Early-stance initial plantarfiexion of the ankle, total ankle range-of-motion, pushoff foot-floor reaction forces and electromyographic data were also collected and analyzed; however, no significant differences were found.

Statistical analyses consisting of split plot analysis of variance were used to evaluate linear measurements and energy cost. Paired t-tests helped evaluate ankle range-of-motion and foot-floor reaction forces. The level of statistical significance was p<.l.

Linear Measurements

The results of motion analysis showed statistically significant differences at p<.05 between the Vascular and Traumatic groups in the linear measurements of velocity, cadence, stride length and single-limb stance time as shown in Table 2 . The Vascular Group walked at 45.0 meters/minute, 82.4 steps/minute and had a stride length of 1.1 meters. They had a right single-limb stance (prosthetic side) 32.4 percent of the gait cycle and left single-limb stance (sound side) 33.8 percent of the gait cycle. The Traumatic Group walked at 64.4 meters/minute, 94.7 steps/minute and had a stride length of 1.4 meters. They had a right single-limb stance (prosthetic side) 36.4 percent of the gait cycle and left single-limb stance (sound side) 39.1 percent of the gait cycle. Normal males walk at 79-90 meters/minute, 113-116 steps/ minute and have a stride length of about 1.5 meters. Normal single-limb stance is 38 percent of the gait cycle for each extremity (2,19,20). These data indicate Traumatic Group ambulation differs from that of the Vascular Group, and both vary from normal. This concurs with the findings of other investigators and affects prosthetic foot prescription criteria (3,13,14).

Within one complete stride, the distance from heel contact to opposite heel contact is the step length. During normal gait, equal right and left step lengths are taken. In this study, no significant difference between right and left step lengths was found when comparing all six subjects. However, some significant differences were found in the Traumatic Group (n=3). Figure 1 shows the step length results in the Traumatic Group and compares the prosthetic-limb step length to the sound-limb step length for each prosthetic foot. In Figure 1 , the sound limb is represented by the zero horizontal axis, and the bar graphs above and below this axis indicate a longer or shorter prosthetic-side step length relative to the sound limb. The Traumatic Group, when wearing the Flex-Walk and the S.A.F.E. II, had a significantly shorter sound-limb step length. However, when wearing the SACH, they had a significantly longer sound-limb step length (p<.O9). This suggests that the prosthetic foot dynamics of the more active traumatic amputee affects step-length symmetry. (Although the Quantum appears to have a significantly shorter sound-limb step length in Figure 1 , this was not statistically significant due to a large standard deviation.)

Ankle Range-of-Motion

Ankle joint motion was also analyzed from motion analysis. Some significant differences were found in late-stance ankle dorsiflexion occurring at opposite heel contact, ankle initial plantarfiexion to dorsiflexion change occurring from early to late stance and weight acceptance forces occurring shortly after heel contact. Symmetry between the sound limb and prosthetic limb was used for comparison.

Late-stance ankle dorsiflexion which occurs at opposite heel contact is an important factor in amputee gait. This dorsiflexion affects the late-stance stability necessary for optimal late-stance balance and the advancement of the opposite extremity. Figure 2 illustrates dorsiflexion results by comparing the prosthetic limb to the sound limb (represented by the zero horizontal axis) to evaluate symmetry for each of the prosthetic feet. The prosthetic limb, when wearing the SACH (n=6), showed significantly less dorsiflexion (p<.05) than the sound limb. When wearing the Flex-Walk (n=6), the prosthetic limb had significantly greater dorsifiexion (p<.05) than the sound limb. The S.A.F.E. II, Seattle Lightfoot, Quantum and Carbon Copy Ii (n=6) did not exhibit any statistically significant differences in late-stance dorsifiexion. These results are important clinically, as they verify that the SACH provides good late-stance stability through limited dorsiflexion. This late-stance stability, for example, is required by amputees with anterior-knee instability on the prosthetic side resulting from a moderate knee flexion contracture, weak prosthetic-side knee extensors or poor late-stance balance. The Flex-Walk, on the other hand, provides less late-stance stability and more late-stance dorsiflexion ankle motion, which is desirable for higher activity levels or for activities requiring more ankle dorsiflexion.

Figure 3 shows both late-stance ankle dorsiflexion results and step-length results in the Traumatic Group (n=3) by comparing the prosthetic side to the sound side (represented by the zero horizontal axis) to evaluate symmetry for each prosthetic foot. In the Traumatic Group, the Flex-Walk had significantly more late-stance ankle dorsiflexion (p<.l) compared to the sound limb, resulting in a significantly shorter sound-limb step length (p<.05).

The opposite appeared to occur when wearing the SACH. The SACH showed less late-stance ankle dorsiflexion compared to the sound limb and exhibited a significantly longer sound-limb step length (p<.07). Comparison of the step lengths reported in the linear measurements to the late-stance ankle dorsiflexion reported here for the Traumatic Group suggests the amount of late-stance dorsiflexion of the prosthetic foot affects sound-limb step length. As more late-stance dorsiflexion occurs with the Flex-Walk, a shorter sound-limb step length is required. Less late-stance dorsiflexion with the SACH results in a longer sound-limb step length. Therefore, step-length symmetry appears to be affected by the prosthetic foot in the more active amputee.

Ankle joint range-of-motion occurring from initial plantarflexion to late-stance dorsiflexion is considered to be the change in dorsiflexion during stance phase of gait. Figure 4 shows the change in dorsiflexion results by comparing the prosthetic limb to the sound limb (represented by the zero horizontal axis) to evaluate symmetry for each prosthetic foot. The prosthetic limb, when wearing the Flex-Walk (n=6), had significantly more change in dorsiflexion (p<.05). The S.A.F.E. II (n=6) also exhibited more change in dorsiflexion (p<.07). This change in dorsifiexion is important because lower activity level amputees with a low velocity and a short step length can use a prosthetic foot with less change in dorsifiexion. Higher activity level amputees will need the prosthetic foot to provide a greater change in dorsifiexion. Activities affected by change in dorsiflexion, such as running or walking on inclines, also deserve special attention in prosthetic foot selection.

Forces

As seen in Figure 5 , normal vertical forces increase in weight acceptance at heel contact, decrease through midstance and increase again to the same magnitude at pushoff. Figure 6 depicts normal fore-shear forces resulting from the braking motion at heel contact and symmetrical aft-shear forces resulting from pushoff (11). In this study, weight acceptance and pushoff floor reaction forces were evaluated from motion analysis. Peak vertical and fore-shear forces occurring in early stance phase were arithmetically combined for total weight acceptance force analysis. Peak vertical and fore-shear forces occurring in late stance were arithmetically combined for total pushoff force analysis. The prosthetic limb was compared to the sound limb to evaluate symmetry. Significant differences were found in weight acceptance; no significant differences were found in pushoff.

Figure 7 shows the weight acceptance force results by comparing the prosthetic limb to the sound limb (represented by the zero horizontal axis) to evaluate symmetry for each prosthetic foot. When wearing the Carbon Copy II (p<.05) and the Quantum (p<.07), the six subjects showed significantly greater sound limb weight acceptance forces. Reducing weight acceptance forces is important in protecting the sound limb from further stress and trauma. The amputee with diabetes and/or peripheral vascular disease is more susceptible to stress created by higher sound-limb weight-acceptance force, which should be considered in prosthetic foot selection.

Energy Cost

Figure 8 shows the results of the energy cost analysis in ml O₂/Kg-m. Energy cost was evaluated by comparing the Vascular Group to the Traumatic Group and the different prosthetic feet to one another. The Vascular Group had a significantly greater energy cost than the Traumatic Group during self-selected comfortable walking (p<.05). No significant differences were found in energy cost among the prosthetic feet tested, and no significant changes in self-selected walking velocity occurred.

Discussion

The significant results in this study can be summarized for each prosthetic foot tested to provide helpful criteria in prosthetic foot selection for amputees. The SACH foot should be used when amputees require maximum late-stance stability by limited dorsiflexion. This is necessary, for example, in amputees with poor late-stance stability resulting from weak knee extensors, knee-flexion contracture or poor mid- to late-stance balance. The SACH foot is also appropriate for lower-activity-level amputees requiring less dorsiflexion.

The S.A.F.E. II foot should be used when increased early- to late-stance change in dorsiflexion is desired. This is necessary for amputees walking on inclines or uneven terrain, or with higher activity levels requiring more ankle motion. The S.A.F.E. II's increased ankle motion better accommodates various walking surfaces and velocities.

The Seattle Lightfoot exhibited no significant differences between the sound limb and the prosthetic limb. Therefore, it provides sound-limb and prosthetic-limb symmetry for amputees, making the Seattle Lightfoot a good choice for average-activity-level amputees with no specific gait abnormalities or considerations.

The Quantum and Carbon Copy II both showed greater sound-limb weight acceptance forces, which should be considered in fitting diabetic and/or dysvascular amputees where protection of the sound limb is important.

The Flex-Walk should be used when increased early- to late-stance change in dorsiflexion and maximum late-stance dorsiflexion is desired. The Flex-Walk better accommodates increased ankle range-of-motion requirements, such as variations in walking velocity and walking on inclines. The Flex-Walk is appropriate for the higher-activity-level amputees.

The differences reported in this study using motion analysis show each prosthetic foot is activity- and gait-specific for the individual amputee. At the conclusion of this study, each subject was fitted with a new prosthesis for his participation. The subjects were allowed to select the prosthetic foot of their choice. All of the subjects selected prosthetic feet similar in function and dynamics to the prosthetic foot on their existing definitive prostheses. This emphasizes that, over a period of time, the amputee becomes accustomed to a specific "feeling" and "function" from the prosthetic foot. It is this feeling and function that affects the amputee's subjective evaluation when comparing different prosthetic feet.

Conclusion

Using the quantitative measures of motion analysis, some significant differences were found in linear measurements, ankle range-of-motion and foot-floor reaction forces of the various feet tested. These data provide information about the dynamic performance of the various feet which can be helpful in prescribing the optimal prosthetic foot for individual amputees. The Vascular Group had a significantly greater energy cost than the Traumatic Group, but no significant difference in energy cost occurred among the different prosthetic feet.

This study concludes:

• Comparison of sound-limb to prosthetic-limb symmetry is the best method to analyze and evaluate different prosthetic feet.
• Some vascular amputee gait characteristics significantly differ from those of the traumatic amputee. These differences in gait characteristics affect analysis and should be considered in prosthetic foot selection.
• Walking velocity, gait pattern and type of activity all contribute to prosthetic foot selection.
• A larger subject population is necessary to find further significant differences among prosthetic feet. In this study, the small subject population tested by motion analysis yielded large variability that may still hide some important differences in these feet.

Acknowledgments

The authors thank the Southern Illinois University Mimi Covert Memorial Motion Analysis Laboratory, Memorial Medical Center Pulmonary Function Laboratory, SIU Orthotic and Prosthetic Services, and Steve Verhulst, PhD, in the SIU Department of Statistics and Measurements for their help and contribution to this study.

This research was funded by grants from the SIU Central Research Committee, Flex-Foot Inc. and M+IND Inc.

All prosthetic feet were contributed by their manufacturers.

Daryl Barth, BS, CPO, is an assistant professor in the division of orthopaedics and assistant director of Orthotic and Prosthetic Services of Southern Illinois University School of Medicine in Springfield, Ill. Mr. Barth also serves as orthotic/prosthetic residency program director for SIU.

Laura Schumacher, BS, CP, is a graduate of the prosthetics/orthotics program at the University of Washington and is currently employed by Orthomedics, Western Division of OSI. This study was conducted by Ms. Schumacher in conjunction with her prosthetic residency at SIU Orthotic and Prosthetic Services.

Susan Sienko Thomas, B.Sc., is manager/kinesiologist of the gait analysis laboratory at Children's Memorial Hospital in Chicago, Ill. She is also an instructor of clinical surgery in orthopaedics at Northwestern Medical School. Previously, she was the kinesiologist/administrative director of the motion analysis laboratory at SIU's School of Medicine.

References:

1. Michael J. Energy-storing feet: a clinical comparison. Clinical Prosthetics and Orthotics 1987 11:3 : 154-168.
2. Skinner, HB, Effeney DJ. Gait analysis in amputees a special review. American Journal of Physical Medicine 1985 ;64:2:82-89.
3. Fisher SV, Gullickson G Jr. Energy cost of ambulation in health and disability: a literature review. Archives of Physical Medicine and Rehabilitation 1978;59: 124-133.
4. Gage JR, Hicks R. Gait analysis in prosthetics. Clinical Prosthetics and Orthotics 1985 ;9:3: 1723.
5. Winter DA, Sienko SE. Biomechanics of below-knee amputee gait. Journal of Biomechanics 1988;21:5:361-367.
6. Nielsen DH, Shurr DG, Golden JC, Meier K. Comparison of energy cost and gait efficiency during ambulation in below-knee amputees using different prosthetic feet a preliminary report. Journal of Prosthetics and Orthotics 1988;1:1:2431.
7. Wagner J, Sienko 5, Supan T, Barth D. Motion analysis of SACH vs. Flex-Foot in moderately active below-knee amputees. Clinical Prosthetics and Orthotics 1987;11:1:55-62.
8. Hannah RE, Morrison JB, Chapman AE. Prostheses alignment: effect on gait of persons with below-knee amputations. Archives of Physical Medicine and Rehabilitation 1984;65:159-162.
9. Perry J. Integrated function of the lower extremity, including gait analysis. Adult Orthopaedics 1984;2:1161-1207.
10. Carbon Copy II produced by Ohio Willow Wood, P.O. Box 192, 15441 Scioto Darby Road, Mount Sterling, OH 43143. Flex-Walk by Flex-Foot Inc., 27071 Cabot Road, Suite 106, Laguna Hills, CA 92653. Quantum by Hosmer Dorrance Corp., P.O. Box 37, 561 Division St., Campbell, CA 95009-0037. SACH by Kingsley Manufacturing Co., P.O. Box 5010, 1984 Placentia Ave., Costa Mesa, CA 92628-5010. S.A.F.E II by Campbell-Childs Inc., 400 Industrial Circle, White City, OR 97503. Seattle Lightfoot by M+IND (Model and Instrument Development), 861 Poplar Place S., Seattle, WA 98144. 11. Winter DA. The Biomechanics and Motor Control of Human Gait. Univ. Waterloo Press, 1988: 1-55.
11. Hannah RE, Morrison JB. Kinematic symmetry of the lower limbs. Archives of Physical Medicine and Rehabilitation 1984;65:155-158.
12. Waters RL, Perry J, Antonelli D, Hislop HJ. Energy cost of walking of amputees: the influence of level of amputation. Journal of Bone and Joint Surgery 1976;58-A:1:42-46.
13. Pagliarulo MA, Waters RL, Hislop HJ. Energy cost of walking of below-knee amputees having no vascular disease. Physical Therapy 1979; 59:5:538-543.
14. Bard G, Ralston HJ. Measurement of energy expenditure during ambulation, with special reference to evaluation of assistive devices. Archives of Physical Medicine and Rehabilitation 1959;40:415-420.
15. Ganguli S, Datta SR, Chatterjee BB, Roy BN. Metabolic cost of walking at different speeds with patellar-tendon-bearing prosthesis. Journal of Applied Physiology 1974;36:4:440-443.
16. Molen NH. Energy/speed relation of below-knee amputees walking on a motor-driven treadmill. Internationale Zietsch rift Fur A ngewandte Physiologie Einschliesslich Arbeitsphysiologie 1973 ;31: 173-185.
17. Gonzales EG, Corcoran PJ, Reyes RL. Energy expenditure in below-knee amputees: correlation with stump-length. Archives of Physical Medicine and Rehabilitation 1974;55: 111-119.
18. Blessey RL, Hislop HJ, Waters RL, Antonelli D. Metabolic energy cost of unrestrained walking. Physical Therapy 1976;56:9: 1019-1024.
19. Lunsford BR, Perry J, Byrd R. Energy-speed relationship of walking: standard tables. Journal of Orthopedic Research 1988;6:2:215222.