Angelika Zissimopoulos, MS; Stefania Fatone, Ph.D., BPO (Hons); Steven Gard, Ph.D.
Northwestern University Prosthetics and Orthotics Research Laboratory and Rehabilitation Engineering Research Program; Physical Medicine and Rehabilitative Engineering, Feinberg School of Medicine, Northwestern University
Knee-ankle-foot orthoses (KAFOs) are used when knee control is desired in addition to ankle or foot control or when the severity of a pathology requires the support of a KAFO rather than a knee orthosis (1, 2). Depending on the desired level of stability, the knee joints of KAFOs are typically either locked or unlocked. In locked knee gait, compensatory mechanisms such as hip hiking, vaulting, and circumduction are often employed to create toe clearance during swing phase. However, research has indicated that these gait compensations increase energy expenditure during ambulation (3-6). Therefore, new orthotic knee joints have been developed to provide stance-phase knee stability with knee flexion during swing. While not all KAFO users have adequate musculature to use these stance-control KAFOs, there are many patients with adequate strength and voluntary control who would benefit from a device that could lock for support during stance-phase and unlock to allow knee flexion and sufficient toe clearance during swing phase. Currently, research evaluating not only KAFOs in general (7) but also new stance-control devices is limited. Therefore, the purpose of this study was to investigate the biomechanical and energetic effects, during gait, of using a KAFO incorporating an orthotic stance-control knee joint.
Nine able-bodied subjects (five females and four males) with no known musculoskeletal or neuromuscular problems participated in this study. The study was approved by the Northwestern University Institutional Review Board, and all subjects provided informed consent prior to participation. Subject specific data were collected including age, height, mass (while wearing the orthosis), and foot length (while wearing shoes). Gait and energy expenditure data were collected with the Stance-Control Orthotic Knee Joint (SCOKJ)®1 operating in three modes: locked, unlocked, and auto (knee flexion blocked in stance but allowed in swing). The SCOKJ®, which operates using a cam mechanism, was chosen as a representative stance-control device. Motion of a cable connecting the ankle and knee engages and disengages the cam mechanism and cable motion can be controlled by weight bearing or by ankle motion. For this study, the SCOKJ® was set up to be activated by ankle motion; plantarflexion pushed the cable up, engaged the joint, and prevented knee flexion while dorsiflexion pulled the cable down, disengaged the joint, and allowed knee flexion.
A qualified orthotist unilaterally cast each subject (left or right side determined by random assignment) and the stance-control KAFO was custom fabricated from reinforced polypropylene and aluminum uprights. Double action ankle joints2 were used to articulate the ankle. Upon successful fabrication and adjustment of the KAFO, subjects were trained to walk with the device on level ground and on a Cosmed model T170 treadmill3 for 20-30 minutes every day up to 10 days, or until they were able to reliably operate the knee joint. Following the training period, subjects underwent a gait evaluation (over-ground walking) and an energy expenditure evaluation (treadmill walking). All data were collected at the VA Chicago Motion Analysis Research Laboratory (VACMARL) of the Jesse Brown VA Medical Center (Chicago, IL).
During the gait evaluation, which consisted of five testing conditions with at least three walking trials collected for each condition, a Helen Hayes marker set was used to collect bilateral kinematic data from eight Eagle Digital RealTime cameras4 sampling at 120 Hz. For the first three conditions, subjects walked at their freely-selected speed with the orthosis in each of the randomized three modes. To create a speed-matched data set, the mode which yielded the slowest walking speed was noted and subjects then walked in the other two modes at this same speed. The speed was monitored during these final two walking conditions to ensure that a speed-matched data set was created. For the energy expenditure evaluation, data were collected using the Cosmed K4b2 portable spirometer3 and the same treadmill. Subjects fasted for two hours before the start of the energy expenditure evaluation to eliminate the effect of food metabolism on energy expenditure measurements. The same testing order was used for both the gait evaluation and the energy expenditure evaluation. For each testing condition during the energy expenditure test, the treadmill was set to match the average speed calculated for each condition during the gait evaluation. The energy expenditure test consisted of three stages: 1) five minutes of pre-exercise resting energy expenditure, which started once the subject was fitted with the Cosmed equipment and calibration was finished; 2) seven minutes of walking energy expenditure on the treadmill for each of the five walking conditions; and 3) at least five minutes of post-exercise resting energy expenditure in a comfortable seated position or until heart rate returned to pre-exercise levels. The laboratory was kept quiet and uninterrupted during the energy expenditure session. For all statistical analyses, a repeated measures ANOVA with Bonferroni correction was used with a significance level of p = 0.05.
Subjects had a mean age, height, and mass (and standard deviation) of 25 ± 2 yrs, 173 ± 8 cms, and 69 ± 10 kgs, respectively. Compared to the locked mode, subjects walked significantly faster and with a higher cadence in the unlocked mode (p = 0.011 and p = 0.005, respectively). However, speed and cadence were not significantly different between the unlocked and auto modes (p = 0.087 and p = 0.217, respectively) or between the locked and auto modes (p > 0.999).
Orthotic side stance-phase knee flexion was significantly smaller for the locked mode compared to the unlocked and auto modes (p = 0.009 and p < 0.001, respectively), and was also significantly smaller for the auto mode compared to the unlocked mode (p < 0.001). Orthotic side swing phase knee flexion was not significantly different between the auto and unlocked modes (p = 0.09), but was significantly smaller in the locked mode compared to the auto and unlocked modes (p < 0.001) (Figure 1). All subjects exhibited hip hiking, (Figure 2) and three utilized circumduction in the locked mode to provide orthotic side toe clearance during swing. These compensations were not observed in the auto mode. Finally, oxygen cost was significantly lower in the unlocked mode compared to the locked and auto modes (p = 0.007 and p = 0.001, respectively), but was not significantly different between the locked and auto modes (p > 0.999).
In this study, data were collected from young, healthy, able-bodied adults to analyze the effects of the SCOKJ® on gait of a homogenous population, rather than the heterogeneous population typical of KAFO users (8). In general, subjects walked with the most gait deviations in the locked mode. Subjects compensated for the loss of locked mode swing phase knee flexion by hip hiking and to a lesser extent, circumducting. In the auto mode, these gait compensations were reduced.
The auto mode freely-selected walking speed was expected to be significantly faster than the locked mode speed. However, walking speed was not significantly different for these two modes. It is possible that the training period was not long enough for subjects to adapt to the operation of the stance-control KAFO. Perhaps a different training strategy would allow subjects to become more comfortable with operating the device. Another explanation is that the stancecontrol KAFO itself limited walking speed. Ankle plantarflexion and dorsiflexion create push rod motion, which in turn engages and disengages the knee joints. The length of the push rod was adjusted at the beginning of the study until the knee joints properly engaged and disengaged. This adjustment depends on step length and therefore walking speed. Subjects may have been constrained to their initial walking speed and step length, which was possibly slower simply due to a lack of familiarity with the device. In the future, it may be beneficial to readjust the push rod length after some training to allow for an improved walking speed.
As expected subjects achieved similar amounts of swing phase knee flexion in the auto and unlocked modes, while stance-phase knee flexion was significantly lower in the auto mode compared to the unlocked mode since stance-phase knee flexion was blocked in the auto mode (Figure 1). Energy expenditure results, however, were not consistent across subjects. Some subjects had the highest oxygen cost in the locked mode while for others the auto mode was the highest. Such inconsistency has been reported by previous investigators (9, 10). In this study, perhaps subjects needed a longer training period to adjust to operating the SCOKJ®. A longitudinal study may help illustrate changes in energy expenditure over time. Further studies should also explore effects of the type of training (i.e. treadmill, level over-ground walking, or community walking).
Finally, it should be noted that results from this study may not be consistent with results that would be obtained from a pathological subject population. Subjects in this study had greater muscle function, strength, and joint range of motion than a typical KAFO user would likely have. This may have influenced various aspects of this study including control of the ankle joint, walking speed, and variability during walking. For example, pathological subjects may have a more fixed walking speed and step length, which would improve the consistency and reliability of engaging and disengaging the SCOKJ®.
In conclusion, allowing swing phase knee flexion alleviated some of the compensatory actions associated with locked knee gait, but did not consistently improve energy expenditure. Perhaps the benefits of improved walking kinematics and improved gait aesthetics are more important than reductions in energy expenditure. Further studies are required to establish the impact of stance-control orthoses on persons with pathology.
Condie DN, Lamb J. Knee-ankle-foot-orthoses, in Biomechanical Basis of Orthotic Management, P. Bowker, et al., Editors. Oxford: Butterworth Heinemann Ltd.,1993: 146-167.
Goldberg B, Hsu JD. Atlas of Orthoses and Assistive Devices. 3rd ed. St. Louis: Mosby-Year Book, 1997.
Mattsson E, Brostrom LA. The increase in energy cost of walking with an immobilized knee or an unstable ankle. Scandinavian Journal of Rehabilitation Medicine 1990; 22:1: 51-3.
Ralston HJ. Effects of immobilization of various body segments on the energy cost of human locomotion. Ergonomics 1965: 53-60.
Waters RL, Campbell J, Thomas L, Hugos L, Davis P. Energy costs of walking in lower-extremity plaster casts. The Journal of Bone and Joint Surgery 1982; 64:6: 896-9.
Hanada E, Kerrigan DC. Energy consumption during level walking with arm and knee immobilized. Archives of Physical Medicine and Rehabilitation 2001; 82:9: 1251-4.
Fatone S. A review of the literature pertaining to KAFOs and HKAFOs for ambulation. Journal of Prosthetics and Orthotics 2006; 18:3S: 137-168.
Irby SE, Bernhardt KA, Kaufman KR. Gait of stance control orthosis users: The Dynamic Knee Brace System. Prosthetics and Orthotics International 2005; 29:3: 269-282.
Hebert JS, Liggins AB. Gait evaluation of an automatic stance-control knee orthosis in a patient with postpoliomyelitis. Archives of Physical Medicine and Rehabilitation 2005; 86: 1676-1680.
McMillan AG, Kendrick K, Michael JW, Aronson J, Horton GW. Preliminary evidence for effectiveness of a stance control orthosis. Journal of Prosthetics and Orthotics 2004; 16:1: 6.