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

Case Study Forum: Gait Comparison of Two Prosthetic Knee Units

Jeffrey L. Sutherland, CO
David H. Sutherland, MD
Kenton R. Kaufman, PhD
Michael Teel, CPO

ABSTRACT

Recognizing that outcome assessments are necessary to justify new and costly prosthetic components, the authors have compared the gait characteristics of a single transfemoral amputee using two prosthetic knee units: the Total Knee and the DAW 4Bar pneumatic knee. All other components of the prosthetic systems were identical, and the patient was given sufficient time to adapt to each system. The authors conducted the research and collected kinematic and kinetic data at the Motion Analysis Laboratory of Children's Hospital in San Diego.

Movement measurements varied slightly between the two prosthetic knee units and differed markedly from normal. The gait using the Total Knee unit demonstrated slight movement of the knee into flexion during stance while the gait using the 4Bar pneumatic unit demonstrated full extension throughout stance. During the loading response of the prosthetic limb, a relatively large initial external extension moment occurred with the 4Bar pneumatic knee unit; the Total Knee demonstrated an initial external flexion moment at this same phase. The vertical force and fore/aft shear curves were closer to normal with the Total Knee. Walking velocity and stride length were improved with the Total Knee, but cadence was unchanged.

Overall, the gait parameters were more favorable with the Total Knee unit. A single trial such as this does not provide statistically significant data but can serve as a model for multiple studies to obtain objective information about prosthetic systems.

Key Words: Force Plate; Kinematics; Kinetics; Gait; Knee.

Introduction

Current trends in reimbursement mandate justification of the use of sophisticated and expensive prosthetic components. Much effort and engineering expertise have been invested in the design of various knee units. This labor deserves precise, objective, functional assessment as can be provided by three-dimensional kinematic and kinetic gait analysis. The information obtained from gait analysis also can guide future designs of prosthetic knee components.

This current study focuses on the effect of a mechanism that allows controlled knee flexion during the loading response phase of gait. The Total Knee unit was selected because its mechanism was designed to permit flexion during loading response. Controlled knee flexion will reduce or eliminate excessive elevation of the center of mass (1). Full extension of the knees in stance increases the vertical oscillation of the center of mass (1). The DAW 4Bar pneumatic kneea was selected since it is representative of pneumatic four-bar knee units widely used and available from manufacturers. This study compares these two prosthetic knee units to each other and, more importantly, to normal gait.

Methodology

Materials and Methods

The subject was a 39-year-old male with a right traumatic midshaft transfemoral amputation. He holds an administrative position in law enforcement; he is 178 cm tall and weighs 89.8 kg. The subject had been an amputee for eight years at the time of the study and was a good community ambulator. He wears his prosthesis approximately 16 hours daily. Before the study was conducted, the authors obtained an informed consent.

The authors built two transfemoral prostheses using two different knee units; the fourth author, a certified prosthetist/orthotist, fit the prostheses used in this study. He optimized the alignments for maximal function using standard subjective, trial-and-error techniques. Identical sockets were used. The second socket was made from an alginate model of the first. The sockets were narrow medial lateral ischial containment, semiflexible inner sockets. Both prostheses incorporated the Endolite ankleb and Seattle Lite Footc with identical ankle components. The first study used the Total Knee. The second study used a DAW 4Bar pneumatic knee with swing- and stance-phase control. The alignments were optimized according to each knee unit manufacturer's specifications.

The subject's definitive prosthesis was an Otto Bock Safety Kneed. His second prosthesis was the DAW 4Bar pneumatic knee; the subject used the DAW knee for five years before the present study was undertaken. Trial number one tested the prosthesis with the Total Knee unit. The subject had been wearing the prosthesis with the Total Knee exclusively four weeks prior to the trial. The second trial tested the DAW 4Bar pneumatic knee. The subject had been wearing the prosthesis with the DAW 4Bar pneumatic unit exclusively for 21/2 weeks prior to the trial; however, he was familiar with this unit as it was the same unit used five years before the present study was undertaken. The subject stated he had become equally accustomed to both prostheses and that he liked them both.

The study was performed at the Motion Analysis Laboratory at Children's Hospital, San Diego. Kinematic motion data were collected with a five charge-coupled device (CCD) camera Vicon systeme. Seventeen retroreflective markers were placed on anatomical landmarks of the pelvis and both lower extremities. Reflective targets were illuminated by infrared stroboscopic lights while the subject walked at a self-selected speed and cadence. The three-dimensional position of each target was calculated using AMASSf software and stored in a DEC MVAXg computer. An average of at least 10 walking trials was collected to gather the mean walking temporospatial parameters. The single trial that was assessed to be most representative of the subject's average gait pattern was selected using a Euclidian norm technique for selection. Thus, the data presented clearly represent the subject's walking characteristics when using each prosthetic knee unit. The joint angle curves of 12 different dynamic movements were calculated and plotted for each lower extremity (2). Kinetic (force and moment) data were obtained with the aid of two Kistler piezoelectric force platesh, which provided measurement of vertical force, fore/aft shear, medial/lateral shear, torque and center of pressure. The external moments at the knee and ankle were calculated from the kinematic and force-plate data. These moments provided important information about foot contact forces and the forces of gravity acting across the ankle and knee.

Results

The time-distance parameters for both studies are shown in Table A . Cadence was 102 steps per minute and was identical in both trials. Walking velocity was 117 cm/sec with the Total Knee unit and 99 cm/sec with the DAW 4Bar unit; this represents an 18-percent increase of velocity with the Total Knee unit. Stride length was 137 cm with the Total Knee unit and 117 cm with the 4Bar unit; this represents a 17-percent increase of stride length with the Total Knee unit. Right step length was 74 cm with the Total Knee unit and 62 cm using the 4Bar unit; this represents a 19-percent increase of right step length with the Total Knee unit.

The 12-joint angle rotation measurements from each trial yielded many identical motion patterns. Only the movements that showed differences are presented. The Total Knee unit trial showed greater right hip extension in terminal stance phase compared with the 4Bar unit and normal (2,3) (see Figure 1 and Table A ). Swing-phase hip flexion compared more favorably to normal with the Total Knee unit than with the 4Bar unit. The difference was decreasing flexion in terminal swing in preparation for load acceptance. Internal rotation around the vertical axis was increased over normal in both the first and second studies but was exaggerated in the 4Bar unit (see Figure 2) . The 4Bar unit showed no knee flexion during the entire period of stance while the Total Knee unit showed a very small movement of the knee into flexion during single-limb stance (see Figure 3) . This movement into flexion was tentative and did not duplicate the normal initial knee flexion wave (3,4).

The ankle motions were identical in the first and second studies, but they differed markedly from the normal ankle motion curve (3,4) (see Figure 4) . The force-plate measurements differed from normal. The right side vertical force curve in the Total Knee unit showed symmetry and slopes that were closer to normal than in the 4Bar unit (see Figure 5) . The fore/aft shear measurements in both studies showed increased fore/shear and decreased aft/shear as compared to normal (see Figure 6) . However, the right side 4Bar unit showed increased fore/shear during loading and decreased aft/shear during terminal stance as compared with the Total Knee. Medial/lateral shear forces showed very little difference in the first and second studies (see Figure 7) . It should be noted that both deviated from normali with an absence of initial medial shear.

The ankle external moments, not shown, were identical with the Total Knee and the 4Bar unit. The sagittal plane knee external moments (see Figure 8) of both studies showed abnormal patterns. The sagittal moments were characterized by an abnormal increase in extension in both studies. There was an abnormally high early extension moment in the 4Bar unit but none with the Total Knee unit. In both studies the initial flexion moment was much smaller than normali.

Discussion

The patient stated he became equally accustomed to both prostheses and had no preference. The authors observed minor but discernible differences when using the Total Knee. These differences are more apparent in the temporospatial parameters and force curves than in the joint angle measurements.

With respect to knee motion there was a slight movement into flexion during single-limb stance when using the Total Knee unit (see Figure 3) . This movement into flexion began at about 10 percent of the gait cycle. This is later than the normal initial knee flexion wave (3,4). Cadence was unchanged with both studies; however, stride length, step length and velocity were increased with the Total Knee (see Table A) .

Vertical force curves showed more normal timing when using the Total Knee (see Figure 6) . The second peak slope was more normal when using the Total Knee, indicating more normal loading during terminal stance. The midstance minimum between the two vertical force peaks represented relative unweighting of the limb as the body center of mass reached its highest point in the gait cycle. With normal gait, this midstance minimum and the slopes that define it are relatively symmetricali. When using the 4Bar knee unit, the midstance minimum and the slopes were asymmetrical, with increased loading and more rapid unweighting in the interval directly after loading response. The second peak of the vertical force had a similar magnitude. The general appearance of the vertical force curves suggested a more efficient loading response with the Total Knee unit. However, the 50-percent difference in walking velocity could have affected the curves since force curves are known to be velocity sensitive (5).

The increase in fore shear (see Figure 6) when using the pneumatic 4Bar knee is correlated with the increase in the first peak vertical force (see Figure 5) . The deficient aft shear is correlated with the decreased height and slope of the second peak in the vertical force curve. The force curves when using the Total Knee suggest a smoother transfer to and from the prosthetic limb than with the 4Bar knee.

The four-bar linkage knee, as represented by the 4Bar DAW unit used in this study, provides knee stability (resistance to flexion) when the instant center of rotation of the prosthetic knee is posterior to a straight line drawn from the femoral head to the heel (6). Extra energy, through increased hip flexion effort, is required to initiate swing phase with four-bar units (6). The authors have no references regarding the energy required to initiate swing phase with the Total Knee. As shown by Radcliffe (7), while there are many versions of four-bar knee units, they all rely on the floor-reaction force vector being anterior to the instant center of the knee unit for stability in stance. This study did not include hip moments or power; thus, the authors did not address the issue of energy expended in initiating swing. The characteristic of four-bar knee units to increase stability with increased knee extension is counterproductive to normal walking in another way. An unyielding knee extension during loading response and throughout stance could produce an excessive elevation of the common center of mass.

This study compared the subject's gait pattern to the pattern a normal individual would have. Normal gait is not attainable for the amputee. Is it reasonable to compare the gait of a patient with a transfemoral amputation to that of normal subjects? Should normal gait be the proper standard for the amputee? No artificial limb yet devised can replace the function of the normal limb. The muscles that control the knee, ankle and foot joints are absent. The proprioception and sensation that would be supplied by the missing limb are lacking. The patient must learn coping mechanisms to adjust to walking without feedback from the missing limb and without substantial muscle motors.

Given these facts, we can take the view that the gait of a transfemoral amputee is predestined to vary substantially from normal walking and accept the four-bar prosthetic knee unit as a time-honored and satisfactory component. Perhaps amputees do not need knee flexion in stance phase; however, the authors do not accept this notion. Normal gait is associated with smooth passage of the common center of mass. Knee flexion, produced by eccentric action of the quadriceps muscles in the normal subject, minimizes the elevation of the common center of mass as it passes forward over the weight-bearing lower limb (1,8). Controlled knee flexion in the stance phase can and must be included in prosthetic knee systems if the gait of the transfemoral amputee is to be brought closer to that of normal subjects.

Differences in stride length, time of toe-off and percent of single stance appeared to favor the unit that allowed some knee flexion during stance. The toe-off time was premature with the 4Bar unit (see Table A) . The percent for single stance, an important indicator of stability, was abnormally low with the 4Bar unit. The magnitude of knee flexion at toe-off was very much less than in normal gait with both units. This is important as preparation for swing is an important component in normal gait. The sequence of flexion for swing phase in normal subjects begins with knee flexion, followed by hip flexion and, finally, ankle dorsiflexion. Neither design permitted normal flexion in preparation for swing phase. This implies further research and additional design characteristics are needed if the gait of transfemoral amputees is to be brought closer to that of normal subjects.

Conclusion

Intuitively, a knee unit that would allow motion similar to normal motion would be ideal. The 4Bar knee allowed no flexion during stance. This is consistent with its four-bar design characteristics (6,7). The Total Knee allowed a small amount of flexion during stance. This flexion, which was less than occurs in normal gait, appeared later in the stance phase; thus, the degree of knee flexion during loading with the Total Knee did not appear to be optimal. The authors cautiously attribute the discernible improvements in time-distance parameters, kinematics and kinetics to the knee unit that permitted some flexion in stance phase without obvious compromise of stability. A delay in knee flexion during preswing was noted in the trials with both prosthetic knee units, implying ease of release may not be optimal for either mechanical design.

This study was of a single subject only; the authors realize no statistical statements can be made about a single-subject study. It also should be noted that other manufacturers currently have prosthetic knee units that allow knee flexion during stance. It will be necessary to include them in future studies. The authors hope and expect many well-planned prospective studies of multiple subjects will follow. This will allow objective, functional assessment of prosthetic systems. 5

References:

1. Saunders JB, Inman V, Eberhardt H. The major determinants in normal and pathological gait. JBJS 1953;35:A:543-58.
2. Sutherland DH, Kaufman KR, Moitoza JR. Kinematics of normal human walking. In: Rose J, Bamble JG (eds). Human walking, 2nd ed. Baltimore: Williams & Wilkins, 1984:24.
3. Sutherland DH. Gait disorders in childhood and adolescence. Baltimore: Williams & Wilkins, 1984:24.
4. Sutherland DH, Olshen RA, Biden EN, Wyatt MP. The development of mature walking. Philadelphia: MacKeith Press, 1988:65-153.
5. Paul JP. The effect of walking speed on the force actions transmitted at the hip and knee joints. In: Proceedings of the Royal Society of Medicine, February 1970;63:2.
6. de Vries J. Conventional 4-bar linkage knee mechanisms: a strength/weakness analysis. J Rehab Res Devel 1955;32:1:36-42.
7. Radcliffe CW. Prosthetics. In: Rose J, Gamble JG (eds). Human walking, 2nd ed. Baltimore: Williams & Wilkins, 1994:165-99.
8. Skinner SR, Antonelli D, Perry J, Lester DK. Functional demands on the stance limb in walking. Orthopedics 1985;8:355-6.