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

Energy-Efficient Knee-Ankle Foot Orthosis: A Case Study

Kenton R. Kaufman, PHD
S.E. Irby, MS
J.W Mathewson, MD
R.W. Wirta
D.H. Sutherland, MD

ABSTRACT

The energy required to walk using a newly designed knee-ankle-foot orthosis (KAFO) has been evaluated. The new KAFO locks the knee during stance and allows free-knee motion during the swing phase of gait.

The energy required for gait on level ground and on a slope with a 5-percent incline was evaluated in a post-polio subject. Comparisons were made between the standard locked-knee KAFO and the free-knee (unlocked) configuration. The oxygen consumption rate (ml/kg/ mm) and energy cost (ml/kg/m) were significantly lower during free-knee gait. The results of the study show that a KAFO design that allows free-knee motion during swing is effective in lowering the energy required for walking.

Introduction

Walking is a complex process in which the body segments can move in many ways. To function efficiently, the lower extremity should have the ability to: 1) support body weight during the stance phase of a locomotion cycle, 2) rotate and coordinate the joints to achieve forward progression, 3) adjust limb length by flexing the knee during the swing phase of gait, and 4) further smooth the trajectory of the center of gravity by slightly flexing the knee in midstance (1).

Normal walking requires muscular strength, joint mobility and coordination of the central nervous system. The absence of any of these capabilities can challenge a person's ability to walk. Yet impaired walking can be mitigated by conventional orthotic intervention, which varies with the level and type of lower-extremity dysfunction.

Some orthoses are more effective than others for specific types of dysfunction. Patients who use a locked knee-ankle-foot orthosis (KAFO) clear the foot during swing phase by adopting compensatory movements at other joints. Examples of compensations for a stiff knee gait are ipsilateral circumduction, hiking of the pelvis or contralateral al vaulting. Such compensations result in increased muscular effort and increased vertical displacement of the body's center of mass.

People with partial or complete paralysis of the lower extremity may require orthotic intervention for stability during stance. These individuals often are prescribed KAFOs, which can compensate for severe weakness of the lower-limb muscles.

Two types of KAFOs generally are prescribed: eccentric (free-knee) joints or locked- (fixed-) knee joints. Eccentric knee orthoses are stable in extension as long as the ground reaction-force vector passes anterior to the hinge axes. The eccentric hinge orthosis design provides limited stance stability and allows flexion/extension at all times. However, the patient must maintain the force vector anterior to the knee axis hinge during stance for stability. The locked-knee KAFO achieves maximum stability through the use of locks. However, this design does not allow any swing-phase knee motion.

A KAFO should provide complete stability in stance phase and unrestricted knee movements in swing phase (2). The authors designed a digital logic-controlled electromechanical KAFO with such capabilities (3-5). The purpose of this study was to compare the energy expenditure of walking with restricted and unrestricted knee motion when using this type of KAFO. The authors hypothesized the elimination of stiff knee gait would result in decreased energy consumption.

Materials and Methods

Logic-Controlled Electromechanical Free-Knee Brace

A small, lightweight, electronically controlled knee lock that can be installed on a conventional KAFO has been developed (3-5). The system is composed of mechanical hardware and an electronic control system (see Figure 1 ).

The mechanical hardware portion consists of a polypropylene orthotic design, a mechanical clutch and a clutch-release actuator solenoid. To adapt the electromechanical components to a standard orthosis, the medial-side knee hinge struts are left intact, and the lateral hinge is removed. Specifically fabricated stainless steel brackets connect the clutch mechanism to the lateral thigh and shank struts. The knee-hinge clutch mechanism is a wrap-spring clutch (a special class of overrunning clutches; an overrunning clutch allows torque to be transmitted from one shaft to another in only one direction of rotation). The assembly is, in effect, a band break connecting two cylinders placed end-to-end and rotating on a common axis (see Figure 2 ). The spring is fastened to the left-hand arbor and slips on the right-hand arbor. The backstopping, or locking-up, occurs when a torque is applied to the clutch, which tends to wrap the spring tightly onto the shaft while locking the shafts together. Conversely, when torque is applied to the opposite side, the spring unwraps from the shaft, allowing the shaft to slip easily in the opposite direction. A solenoid is used to control the clutch.

The electronic control system is composed of digital logic-integrated circuits. A combination logic network monitors input data and produces electrical output commands based on the input states. The inputs to the control circuitry are signals generated by strategically located foot contact sensors. Based on the input, the controller algorithm generates an actuation signal that is sent to the solenoid for release of the clutch during the swing phase of gait.

Subject

A 40-year-old male subject with poliomyelitis was studied. The subject incurred poliomyelitis at age 2 that affected the lower left extremity with primary weakness of the hamstrings (grade two), quadriceps (grade one) and calf muscles (grade two). The ankle motion was limited due to a triple arthrodesis at age 12. The subject exhibited normal range of motion of the hips and knees on both sides. The subject was selected for participation in this study because he uses orthoses in his professional life and could provide valuable feedback regarding the KAFO design and function.

Knee Motion

Kinematic parameters were acquired using a computerized video motion analysis system (VICON ) with five infrared cameras. The spatial distribution of the cameras was optimized to yield reliable motion data bilaterally at the hip, knee and ankle. Reflective markers were placed on the subject's lower limbs to identify the relative rotations of the limb segments. Markers were placed on a sacral stick and bilaterally on the anterior/superior iliac spines, greater trochanters, lateral femoral condyles, lateral malleoli, calcanei and fifth metatarsal heads as well as on a 12-cm wand taped to the lateral aspect of the tibias. The motion analysis system was calibrated prior to each gait analysis. Video motion data were acquired over a 3-in length of a 9-in walkway to yield one complete gait cycle. One set of data corresponding to the standing position (static data) was recorded to calculate joint centers.

After a brief orientation session, the subject was asked to walk along the walkway. At least five trials were conducted. The time and distance parameters were calculated and averaged, and the walking cycle that most closely typified the patient's gait was selected for further analysis. The selection was made by calculating the Euclidian norm for the deviation of velocity, cadence and step length from the average velocity, cadence and step length. The cycle with the smallest Euclidian norm was considered the representative cycle. The two-dimensional coordinates recorded by each of the cameras were reduced to a set of three-dimensional marker coordinates using the analytical software package (AMASS b) provided within the VICON system.

Once the marker positions were computed, the rotation of the limb segments was calculated using software developed in the Motion Analysis Laboratory (6,7). The software was configured to calculate and display joint angles and deviations from normal gait at the hip, knee and ankle as a function of the gait cycle (8). Angular rotations were calculated about three axes of the hip, knee and ankle. The authors used Fourier analysis to quantitate wave forms and smooth the data. The gait cycles were manipulated (extended and compressed) to yield a normalized gait cycle. All gait events were expressed as a percentage of the gait cycle, irrespective of the actual time for a stride.

Energy Expenditure Measurement

Energy expenditure testing was conducted on an electronically controlled treadmill. During testing, the subject wore his customary shoes and was required to maintain an erect walking posture without walking aids.

The subject was studied using the orthosis in both locked and unlocked configurations. Data were collected at treadmill grades of 0- and 5-percent incline. Walking velocities ranging from 15 to 80 in/mm (the functional range of walking speeds in adults) were tested. The average slow and fast walking speeds in adults ages 20-59 years range from 43 to 106 in/mm (9).

The subject breathed through a suspended mouthpiece that allowed vertical, lateral and forward/backward movement with changes in head position. Gas samples were analyzed for oxygen content. A mass spectrometer (Perkin Elmer) and a volume turbine (Sensor Medics) were interfaced with commercially available software (Firstbreath) to calculate energy consumption variables on a breath-by-breath basis. The mass spectrometer was interfaced to the mouthpiece via a known length of capillary tube, precalibrated for lag time. Prior to the test, the mass spectrometer was calibrated with precision gases.

Once physiological steady state was reached, data collection commenced. (Steady state was determined by observing the slope of the oxygen consumption rate and was reached within two minutes for all conditions.) All gas volumes were corrected to standard values of temperature, saturation and pressure (STPD). Standard protocols for energy expenditure measurement were followed (10,11).

Energy expenditure during walking was expressed by three parameters (12). The rate of oxygen consumption (1702) is the amount of oxygen consumed per minute (ml/kg/min). The oxygen consumption rate indicates the intensity of sustained exercise and is related to the length of time exercise can be performed. The energy cost per meter (ml/kg/m) describes the amount of oxygen needed to walk a unit distance and indicates physiologic work. The energy cost equals the oxygen consumption rate divided by the speed of walking.

A comparison of the energy cost per meter to an averaged value for normal walking enabled the authors to determine gait efficiency. Normal values for energy expenditure were taken from the literature. Energy expenditure data for normal gait on a level surface were available from Waters (9) and Bobbert. The relationship between energy consumption and speed during normal walking on a slope was taken from Bobbert (13). Gait efficiency was defined as the ratio of the energy cost for an able-bodied individual divided by the energy cost for the patient. Since the energy cost for the patient is nearly always greater than normal, the gait efficiency is less than 100 percent depending on the degree of disability. Thus, gait efficiency is a measure of the patient's rate versus an able-bodied individual's rate of energy expenditure at comparable speeds (9).

For accurate measurements, the oxygen consumption rate (ml/kg/mm) must be distinguished from energy cost (ml/kg/m). The oxygen consumption rate indicates the intensity of physical effort during exercise and is a time-dependent parameter. Energy cost is not time-dependent. A high oxygen consumption rate indicates a high intensity of exercise. On the other hand, a high energy cost indicates a high degree of gait disability.

Data Analysis

A repeated measures design was used. The subject was tested during level, over-ground and inclined-slope walking. The KAFO was tested in locked-and unlocked-knee configurations. Regression analysis was used to determine the relationship between oxygen consumption rate or energy cost and velocity. The data were partitioned into separate groups for each KAFO configuration, and a linear regression analysis determined the relationship of oxygen consumption rate to walking speed. Waters and Lunsford (14) have confirmed oxygen consumption rate increases in a positive linear fashion with walking velocity. The authors used regression analysis to test whether the two straight line regression equations (locked versus unlocked KAFOs) were coincidental (15). The shape of the energy-cost curve was considered parabolic (16). A paired t-test was used to determine statistical differences in gait efficiency.

Results

Knee Motion Pattern

Dynamic gait analysis clearly has shown improvements in the knee-motion pattern while using the new KAFO (see Figure 3 ). When the orthosis was tested in the locked configuration, the knee position was set at 25 degrees throughout the gait cycle to simulate a standard, locked KAFO. When the clutch-control algorithm was actuated, the knee maintained a stable, locked configuration during stance of 25 degrees but obtained 65 degrees of knee flexion during the swing phase of the gait cycle. The knee swing-phase motion pattern approached the motion of an able-bodied individual when the free-knee algorithm was used to control the knee function of the orthosis.

Oxygen Consumption Rate (1702)

At each treadmill speed, oxygen consumption rate increased in a linear manner for both locked and unlocked configurations (see Figure 4 ). The increase in oxygen rate was significant in both the locked (r²=0.96, p=0.00l) and unlocked positions(r²=0.96, p=0.00l) on level ground (0-percent slope). A similar pattern was true for the 5-percent incline with the increase in oxygen rate again significant when using the locked (r²=0.99,p=0.027) and unlocked(r²=0.98, p=0.099) configurations.

For each slope condition, oxygen consumption rate was greater with the use of the locked configuration. Comparison of the regression lines at 0percent slope revealed the intercepts were not significantly different (p<0.05) yet the slopes of the two lines were not the same (p<0.025). Comparison of the regression lines for 5-percent incline showed the slopes were parallel (p<0.05), but the lines were not coincidental (p=0.07). Thus, the unlocked configuration reduced metabolic energy requirements for ambulation. Nevertheless, the energy requirements for the subject were higher than those for able-bodied individuals (see Figure 4 ).

Energy Cost

The energy cost per meter walked varied with the speed of walking (see Figure 5 ). The relationship was defined by a second-order relationship. The change in energy cost with speed was significant when the orthosis was locked (r²=0.97, p=0.005) and unlocked (r²=0.94,p=0.016). Similar to the oxygen consumption rate, the energy cost was higher for the subject tested than for able-bodied individuals (see Figure 5 ).

Gait Efficiency

The gait efficiency when walking with locked as well as unlocked knee joints was compared to normal gait (13). Linear regression analysis determined the gait efficiency did not vary with the speed of walking on level ground for either the locked (p=0.09) or unlocked (p=0.l4) configurations. The mean gait efficiency was 58 (+/- 3) percent (mean +/- standard error) for the locked and 62 (+/- 3) percent for the unlocked condition (see Figure 6a ). This difference in gait efficiency was statistically significant (p=0.032).

Similarly, linear regression analysis confirmed the gait efficiency did not vary with the speed of walking up a 5percent incline in either the locked (p=0.256) or unlocked (p=0.545) state. The gait efficiency was greater on the incline: 73 (+/- 2) percent in the locked configuration and 82 (+/-2) percent in the unlocked state (see Figure 6b ). This difference in gait efficiency on the incline for the differing KAFO configurations (locked versus unlocked) also was statistically significant (p=0.006).

Discussion

Surveys show rejection rates for KAFOs are high, ranging from 60 percent to nearly 100 percent (17-20). The primary reasons for discontinuing the use of an orthosis were changes in the needs of the user, difficulty in obtaining the orthosis from the supplier, unacceptable performance of the orthosis and difficulty using the orthosis. In addition, orthotic intervention was abandoned in favor of wheelchairs because patients found that, even with the aid of an orthosis, walking still required too much energy.

Patients who require KAFOs typically accept orthoses for a very short period following injury or disease but soon choose to use wheelchairs, presumably because walking with locked knees is energy inefficient. Cerny et al. (21) have shown that walking with KAFOs is more inefficient than wheelchair propulsion for individuals with paraplegia who depend on KAFOs to walk, even for patients who customarily use KAFOs for locomotion. Walking with KAFOs is much less energy efficient than normal walking, whereas values for wheelchair propulsion approximate values for normal walking (21). These data suggest wheelchair propulsion is selected as the primary mode of locomotion because walking with two KAFOs is more physiologically taxing.

Most of the research and development efforts aimed at improving impaired gait have been directed at prosthetic systems. Design engineers face fewer technical problems in developing prosthetic limb replacements than in developing orthotic systems. For example, an orthosis adds weight and volume to the lower extremity, which limits the size and weight available for accommodating the orthosis. Other than the application of modern plastics to orthotic designs, there have been no real changes in the function of conventional KAFOs for decades (22).

The effects of restricted knee motion on the metabolic cost of walking have been studied by other investigators. Using plaster casts, Ralston (23) immobilized the knee at flexion angles of 0, 15, 30 and 45 degrees. Comparisons were made at a walking velocity of 74 in/mm. The changes in energy expenditure varied with the amount of knee flexion. The smallest change was at 15degree knee flexion where the gait efficiency was 78 percent of that of the unrestricted knee. The largest change was at a knee flexion of 45 degrees where the gait efficiency was 64 percent of that obtained with the knee unrestricted. Waters et al. (24) performed a similar study placing plaster casting around the fully extended knee without interfering with ankle or hip motion. At a walking speed of 64 in/mm, the oxygen rate consumption was 12.7 ml/kg/min, and the energy cost was 0.20 ml/kg/m. The gait efficiency was 76 percent of that of an able-bodied individual. Mattsson et al. (25) also performed a study with the knee immobilized and the ankle free. The subjects walked at a self-selected walking speed. The walking speed was 61 in/mm with the immobilized knee. At this speed, the energy cost was 0.160 ml/kg/m without immobilization and 0.196 ml/kg/m with immobilization. Thus, the gait efficiency for knee immobilization was 82 percent.

In the present study, the reported energy cost is higher and the reported gait efficiency is lower than the measures reported in these other studies. The difference probably is due to two factors. First, the subjects of the other studies were healthy adults with no muscular weakness. Thus, even though each subject's knee was immobilized, he or she could use muscular substitutions to compensate for the immobilization. Second, the other studies immobilized the knee but allowed the ankle complete freedom. This would not necessarily be possible when wearing a KAFO.

The energetics of walking with a KAFO have been tested by Cerny et al. (21) in adults with low-level spinal cord injuries. These individuals walked at a velocity of 32.4 in/mm; their oxygen consumption rate was 20.4 ml/kg/min; and their energy cost was 0.99 ml/kg/in. These measures are higher than those reported in the present study. However, 80 percent of the subjects in the Cerny study used a wheelchair as their primary mode of locomotion. Thus, these subjects were more disabled than the subject in the present study.

While the present study only includes one subject, the energy consumption for this subject falls within parameters defined by able-bodied individuals with knee immobilization and subjects with a high degree of disability. No attempt is made to extrapolate beyond these limits.

The difference in the subject's oxygen consumption rate when using the locked and unlocked knee configurations was about 1 ml/kg/mm. In an able-bodied individual, this difference would translate into a change in walking velocity of about 8 m/min (9). Most adults would prefer to walk at speeds between 74 and 82 m/min (9,26). Thus, this difference in energy expenditure would equal approximately a 10-percent difference in walking velocity for an able-bodied adult subject. During normal walking, body segments can move in many ways. Vertical displacements of the body are against gravity and, as a rule, require more energy than do horizontal displacements (27). Six key motions have been described that minimize vertical displacement of the body during gait. These motions are termed the "determinants of gait" and include pelvic rotation, pelvic tilt (pelvic obliquity), knee flexion at heel strike, ankle-foot interaction, knee motion and lateral pelvic motion (28). Interacting with gravity, the lower limbs propel the body forward efficiently (29). Use of KAFOs may improve static posture but fail to improve function because free motion of the knee and ankle are not allowed. The vertical hip oscillation of a patient using a conventional KAFO was increased by 65 percent over that when no orthosis was used during gait (30).

Conclusion

The improved KAFO described in this study provides an articulated knee-joint system (3-5) that reduces the metabolic energy requirements during gait. This orthosis provides knee stability during stance while allowing free-knee motion during the swing phase of gait. The ability to freely move the leg during the swing phase of gait results in more energy-efficient ambulation. The principle presented in this study should be applied to all future KAFO designs. The results of this study are applicable to any patient who suffers from partial or complete paralysis of the lower extremity and requires a KAFO for ambulation.

Acknowledgements

This work was supported by NIH Grant 1ROl HD30150.

KENTON R. KAUFMAN, PhD is director of orthopaedic research at the Motion Analysis Laboratory at Children's Hospital in San Diego, 3020 Children's Way, San Diego, CA 92123-4282; (619) 576-1700. He also is adjunct associate professor at the University of California-San Diego.

S.F. IRBY MS, works at the Motion Analysis Laboratory at Children's Hospital in San Diego.

J.W. MATHEWSON, MD, directs the Cardiovascular Stress Laboratory at Children 's Hospital in San Diego. R.W. WIRTA works at the Motion Analysis Laboratory at Children's Hospital in San Diego.

D.H. SUTHERLAND, MD, is the medical director at the Motion Analysis Laboratory at Children's Hospital in San Diego and a professor at the University of California-San Diego.

References:

1. Perry J. The mechanics of walking: a clinical interpretation. In: Perry J, Hislop HJ, eds. Principles of the lower-extremity bracing. New York: Amer Phys Ther Assn, 1967:9-32.
2. Sutherland DH, Olshen RA, Cooper L, Wyatt M, Leech J, Mubarak 5, Schultz P. The pathomechanics of gait in Duchenne muscular dystrophy. Dev Med and Child Neur 1981; 23:3-22.
3. Irby SE. A digital logic-controlled electromechanical free-knee brace, MS1 Thesis, San Diego State University, San Diego, Calif., 1994.
4. Kaufman KR, Irby SE, Wirta RW, Us sell DW, Mathewson JW, Sutherland DH. Knee-ankle-foot orthosis for free-knee gait. Second world congress of biomechanics. Amsterdam, The Netherlands, July 10-5, 1994:280.
5. Malcolm LL, Sutherland DH, Cooper L, Wyatt M. A digital logic-controlled electromechanical orthosis for free-knee gait in muscular dystrophic children. Orthop Transactions 1980;5:90.
6. Sutherland DH, Olshen RA, Biden EN, Wyatt MP The development of mature walking. London: MacKeith Press, 1988.
7. Kaufman KR, Moitoza JR, Sutherland DH. Relation between external markers and tibial rotation measurements. The international symposium on 3-D analysis of human movement. Montreal, Canada: 1991 ;52-4.
8. Kaufman KR, Wyatt M, Sutherland DH. Implementation of prediction regions for motion data. Dev Med and Child Neur 1991;33(A):Supp64:29-30.
9. Waters RL, Lunsford BR, Perry J, Byrd R. Energy-speed relationship of walking: standard tables. J of Orth Res 1988;6:215-22.
10. Jones NL. Clinical exercise testing, 3rd ed. Philadelphia: W.B. Saunders Co., 1988.
11. Wasserman K, Hansen JE, Sue DY, Shipp BJ. Principles of exercise testing and interpretation. Malvern: Lea and Febiger, 1987.
12. Waters RL. Energy expenditure. In: Perry J, ed. Gait analysis: normal and pathological function. Thorofare, N.J.: Slack Inc., 1992:443-89.
13. Bobbert AC. Energy expenditure in level and grade walking. J of App Phys 1960;15:6:1015-21.
14. Waters RL, Lunsford BR. Energy cost of paraplegic locomotion. JBJS 1985;67A:1245-50.
15. Kleinbaum DG, Jupper LL. Applied regression analysis and other multivariable methods. Belmont: Wadsworth Publishing Co., 1978.
16. Ralston HJ. Energy-speed relation and optimal speed during level walking. Internationale Zeitschrift fur Angew Physioleinschl Arbeitsphysiol 1958;17:277-83.
17. Kaplan LK, Grynbaum BB, Rusk HA, Anastasia T, Gassler S. A reappraisal of braces and other mechanical aids in patients with spinal cord dysfunction: results of a follow-up study. Arch Phys Med and Rehab 1966;47:393-405.
18. Rosman M, Spira E. Paraplegic use of locking braces: a survey. Arch Phys Med and Rehab 1974;55:310-4
19. Phillips B, Zhao H. Predictors of assistive technology abandonment. Assis Tech 1993;5:36-45.
20. Coghlin JK, Robinson CE, Newmarch B, Jackson G. Lower-extremity bracing in paraplegia: a follow-up study. J of Paraplegia 1980;18:25-32.
21. Cerny K, Waters R, Hislop H, Perry J. Walking and wheelchair energetics in persons with paraplegia. Phys Ther 1980; 60:9:1133-9.
22. Lehneis HR. Orthotics: the state of the art. J of Rehab Res and Dev 1993; 30:4:vii-viii.
23. Ralston HJ. Effects of immobilization of various body segments on the energy cost of human locomotion. Ergonomics 1965; Suppl 53.
24. Waters RL, Campbell J, Thomas L, Hugos L, Davis P Energy cost of walking in lower-extremity plaster casts. JBJS 1982: 64:896-9.
25. Mattsson E, Brostrom L-A. The increase in energy cost of walking with an immobilized knee or an unstable ankle. Scandinavian J Rehab Med 1990;22:51-3.
26. Finley FR, Cody KA. Locomotive characteristics of urban pedestrians. Arch of Phys Med and Rehab 1970;51:423-6.
27. Inman V, Ralston HJ, Todd E Human walking. Baltimore: Williams & Wilkins, 1981.
28. Saunders JB, Inman V, Eberhart H. The major determinants in normal and pathologic gait. JBJS 1953;35:3:543-58.
29. Inman V. Conservation of energy and ambulation. Bull of Pros and Res 1968; 10:26.
30. Allard PL, Duhaime M, Thiry PS, Drown G. Use of gait stimulation in the evaluation of a spring-loaded knee joint orthosis for Duchenne muscular dystrophy patients. Med and Bio Eng and Comp 1981; 19:165-70.

Website built by oandp.com