The Dawn of Powered Lower-Limb Prostheses
March 2013 Issue
During the last decade, the O&P industry has enjoyed a technological renaissance. As technologies improve and become more reliable, their adoption becomes more widespread. For example, the microprocessor regulation of prosthetic knees, once a novel and emerging technology, has largely been embraced as the standard of care in transfemoral prosthetics. More recently, the microprocessor regulation of prosthetic foot and ankle mechanisms has been introduced and is now being refined, with growing application into prosthetic care. As the use of "intelligent" prostheses that are capable of altering their responses to real-time situational needs becomes more common, the next logical step in this progression is the introduction of externally powered prostheses that go beyond the variable dampening characteristics of today's microprocessor control by generating their own powered movements. This article reviews the most commonly described applications of such powered prostheses in current literature.
Powered Ankle: Kinematics, Power, and Energy Consumption
A number of strategies have been explored to provide powered ankle motion, including pneumatics, springs, and motors.1-4 Of these, the only system that has been developed to the point of commercial release is the motor-driven iWalk BiOM, which has been the subject of several recent trials. In a 2011 study by Mancinelli et al., five unilateral transtibial amputees were fitted with a powered BiOM foot and ankle and an Össur Ceterus foot, a traditional passive-elastic prosthesis, prior to testing sessions that examined gait kinematics, kinetics, and oxygen consumption.5 (Editor's note: The Ceterus foot was discontinued in February 2011. It was replaced with the Re-Flex Rotate foot at the end of 2010.) The BiOM is designed to produce a powered "push-off" event, so it is unsurprising to observe that subjects experienced a mean maximum plantarflexion angle of 10.5 degrees at push-off when using it. By contrast, the mean maximum plantarflexion angle in the Ceterus foot was 2 degrees of dorsiflexion. The value in normative controls was reported at approximately 17 degrees of plantarflexion.
In addition to the improvements in ankle kinematics using the powered foot and ankle prosthesis, study subjects experienced additional benefits in ankle kinetics. The energy storage and return properties of the Ceterus are such that despite its passive design, it supplied an appreciable mean peak power generation of 1.87 watts per kilogram. The BiOM, however, facilitated a larger mean peak power generation of 2.89 watts per kilogram, closely approximating the normative value of 2.92 watts per kilogram.5
Given the increases in propulsive push-off during terminal stance in the powered prosthesis condition, improved oxygen consumption values could also be reasonably anticipated. In fact, the effect of the BiOM on energy consumption compared to those observed with the Ceterus were extremely variable across subjects, ranging from 0.58 percent improvement when using the Ceterus to 17 percent improvement when using the BiOM. Across the small cohort, the average improvement when using the BiOM compared to the Ceterus was 8.4 percent.5 Taken together, these preliminary observations define some of the benefits that may be obtainable with an externally powered prosthetic ankle mechanism, including ankle kinematics that better approximate normative values, improved push-off at the ankle, and reduced oxygen consumption values.5
Observations across Variable Speeds
These findings were followed by a 2012 clinical trial in which comparisons were drawn from the observations of seven unilateral transtibial subjects with their legacy prosthetic feet and with the BiOM, as well as healthy, matched controls.6 In contrast to the 2011 trial in which subjects were only observed at a single, self-selected walking speed, in this trial subjects walked on a treadmill at five predefined velocities. The researchers recorded mean energy consumption values for the healthy, matched controls and unilateral transtibial subjects wearing the BiOM foot across the various gait velocities, along with the relative improvements recorded with the BiOM compared to their legacy prosthetic feet (Table 1).6
In addition to the improvements in energy consumption at all but the slowest preselected velocities, subjects tended to prefer faster walking speeds in the BiOM foot than they did in their legacy prosthetic feet. The mean preferred velocity in the latter condition, 1.16 meters per second, increased 23 percent to 1.42 meters per second with the powered prosthetic foot and ankle.6
The previous two trials were characterized by acclimation periods of only a few hours and some level of involvement by employees of iWalk, Bedford, Massachusetts, the BiOM's manufacturer. A third trial was independently conducted by Ferris et al. at the U.S. Army's Center for the Intrepid, Fort Sam Houston, Texas, on subjects who had a minimum three-week acclimation period to the powered foot and ankle prior to data collection. In addition to similar biomechanical analyses as those used in the previous trials, these subjects completed the Prosthetic Evaluation Questionnaire (PEQ), a standardized user satisfaction assessment.7
Observations from this trial confirm many of the findings of the previous two. While the plantarflexion angle experienced in the BiOM during initial swing was less than that observed in matched controls, it was substantially greater than that recorded with the subjects' legacy dynamic-response feet.7 Similarly, while the power generation from the ankle during preswing in the legacy prosthetic feet was 40 percent less than that observed in able-bodied controls, the power obtained from the BiOM foot during this phase of gait was 35 percent greater than that seen in the controls and 125 percent greater than that seen with the legacy prosthetic feet.7 Also, consistent with the findings of the previous trials, the subjects' self-selected walking speeds increased an average of 6 percent in the powered foot and ankle condition.7
In addition to confirming many of the basic improvements observed with the powered prosthetic foot and ankle, the authors of this third trial examined a number of inter-limb asymmetries at the knee and hip with both prosthetic foot options and found that many of these asymmetries persisted with the powered prosthetic foot and ankle, suggesting that the restoration of ankle power may not fully restore normal gait kinematics and kinetics.7
Finally, with respect to user satisfaction, of the ten subjects who completed the PEQ, seven ultimately preferred the powered prosthesis to their legacy prosthetic foot. Modest increases were seen across the PEQ evaluations of "ambulation," "frustration," "utility," and "well-being" with the powered prosthetic foot and ankle. The most striking evaluation in favor of the powered prosthesis was "perceived response," in which a 10 percent improvement was observed. The only evaluation point that favored the legacy prosthetic feet was "sounds" where subjects reported a 21 percent decrease in satisfaction with the powered foot and ankle.7
Powered Ankle/Knee: Kinematics, Cadence, and Energetics
The most thoroughly documented and described powered prosthetic solution for transfemoral amputations in the literature is the Vanderbilt University powered knee and ankle system.8-11 While the knee is not yet commercially available, several articles have been published to describe its functionality across a range of tasks and environments. For example, a 2008 study of a single subject using a prototype version of the powered ankle/knee prosthesis revealed that during level ground ambulation, the mechanism facilitated more normalized ankle and knee kinematics in both weight acceptance and push-off than the subject's conventional prosthesis, which included a dynamic-response foot and microprocessor knee.8
Across three different walking velocities-"slow," "normal," and "fast"-the subject's kinematics with his conventional prosthesis were characterized by limited stance flexion of the knee during loading response and limited ankle motion throughout the gait cycle, with approximately 10 degrees of motion in dorsiflexion and limited plantarflexion. The powered ankle/knee prosthesis allowed a more physiologic amount of ankle plantarflexion and knee flexion during weight acceptance and a propulsive push-off at the ankle during pre-swing.8
Similar to the effects of the powered ankle on transtibial amputees, the transfemoral subject's self-selected walking speed increased with the powered ankle/knee mechanism; however, it did so by a much more appreciable 24 percent.8 When the subject walked on a treadmill at a sustained velocity of 0.89 meters per second, his oxygen uptake in his conventional prosthesis was 23 percent greater than that recorded with the powered prosthesis.8 These improvements are tempered by their occurrence in a single subject only but may indicate that the provision of power is even more beneficial following transfemoral amputation.
In addition to the benefits during level ground walking, the developers of the powered knee/ankle system have described its ability to adapt to variable inclines and declines while standing as well as upslope walking.9-10 In the case of the former, the authors describe a condition in which the subject was asked to stand on a series of wedges, in 5-degree increments from 15 degrees of incline to 15 degrees of decline, while wearing both the conventional and powered prosthesis.9
While wearing the conventional prosthesis, the subject relied heavily on his sound-side limb on both inclined and declined surfaces. In the case of declined surfaces, the relative stiffness of the conventional foot/ankle mechanism forced the prosthesis into excessive knee flexion, which required the patient to offload the prosthesis to maintain his stability. In the case of inclined surfaces, the same stiffness of the conventional foot/ankle mechanism was such that weight could only be borne through the toe of the prosthesis and required the patient to hike the prosthetic-side hip to accommodate the functional leg-length discrepancy.9
The Vanderbilt team, on the other hand, describes a state controller in the powered ankle/knee prosthesis that allows it to determine the angle of inclination within 1 degree and adapt the sagittal ankle angle accordingly. As a result, the subject was able to maintain a more symmetrical load distribution in the powered prosthesis across the range of wedges.9
With respect to upslope walking, the developers of the Vanderbilt knee observed kinematics in the same subject's conventional knee prosthesis and powered knee/ankle prosthesis across level ground and surfaces inclined approximately 5 degrees and 10 degrees.10 The kinematics in the conventional knee were similar across the three surfaces. Stance flexion of the knee was limited, and to the extent that ankle motion occurred, it was into dorsiflexion during mid to late stance. The inability of the conventional prosthesis to adapt to the upslope inclinations required the subject to perform a compensatory vault of the sound-side ankle to obtain swing phase clearance.
The powered knee/ankle was able to adapt to the variable slopes. As the degree of incline increased, the powered prosthesis experienced greater knee flexion during loading response, adopted a sagittal ankle alignment of progressive dorsiflexion, and generated progressively increasing stance phase knee extension moments.10 In addition, the magnitude of the pushoff during terminal stance also increased with more inclined surfaces, reflecting the increased energy requirements associated with upslope walking.10
In a 2012 article, Lawson et al. describe the differences observed in the same case subject during stair ascent and descent in his powered knee/ankle prosthesis and his conventional prosthesis.11 Beginning with stair ascent, like most users of transfemoral prostheses, the subject reported that he preferred a "step-to" stair ascent in which he stepped up with the sound limb and then brought his prosthesis up to that step before initiating the next. However, because he ascended stairs step over step in the powered prosthesis, he attempted a reciprocal stair ascent in his conventional prosthesis as well.
In doing so, he adopted the common compensational strategies of minimal hip flexion combined with circumduction to advance the prosthetic-side limb to the next step. Then, using an aggressive sound-side ankle push-off and heavy reliance on the hand rail, he was able to elevate his body over the relatively stiff prosthesis.11 The powered prosthesis, on the other hand, facilitated an almost physiologically normal stair ascent pattern in which the prosthetic-side step was initiated at approximately 70 degrees of knee flexion and a powered knee extension brought the knee to approximately 10 degrees of flexion as the subject ascended to the next step.11 The ability of the Vanderbilt knee to facilitate powered knee ascent made national news in November 2012 when another transfemoral amputee ascended 103 floors (2,109 steps) of Chicago, Illinois' Willis Tower in the annual stair-climbing charity event, SkyRise Chicago.12
The differences in performance between the two prostheses were more subtle during stair descent when the prosthetic knee functions as a damper. Using his conventional prosthesis, the subject used the common strategy of placing the prosthetic foot roughly halfway over the edge of the stair immediately beneath him and allowing the knee to flex through its dampened state as it lowered him to the next step.11 While the knee kinematics observed with the powered prosthesis were similar, the descent strategies differed at the ankle. Compared to the conventional strategy in which there was no appreciable motion at the prosthetic ankle, in the powered prosthesis, the descending step began at almost 20 degrees of plantarflexion. Stair descent began with resisted dorsiflexion of the foot, followed by resisted knee flexion.11 Thus, the greatest difference between the two strategies is the powered knee/ankle prosthesis' ability to remain flat on the step throughout the descent.
Promises of the Future
While the application of external power in lower-limb prosthetics has advanced on many fronts, this article has focused on the two applications that have the greatest amount of supportive academic literature-the BiOM in transtibial applications, and the Vanderbilt prosthesis in transfemoral applications. The two approaches provide a preliminary understanding of where external power may take the O&P profession with respect to such variables as gait kinematics, kinetics, and energetics across both level ground walking and navigating environmental obstacles.
Phil Stevens, MEd, CPO, FAAOP, is in clinical practice with Hanger Clinic, Salt Lake City, Utah. He can be reached at
- Klute, G. K., J. M. Czerniecki, and B. Hannaford. 2002. Artificial muscles: Actuators for biorobotic systems. International Journal of Robotics Research 21:295-309.
- Versluys, R., G. Lenaerts, M. Van Damme, I. Jonkers, A. Desomer, B. Vanderborght, L. Peeraer, G. Van der Perre, and D. Lefeber. 2009. Successful preliminary walking experiments on a transtibial amputee fitted with a powered prosthesis. Prosthetics and Orthotics International 30:356-63.
- Hitt, J. K., T. G. Sugar, M. Holgate, and R. Bellman. 2010. An active foot-ankle prosthesis with biomechanical energy regeneration. Journal of Medical Devices 4(1):011003-11.
- Au, S., M. Berniker, and H. Herr. 2008. Powered anklefoot prosthesis to assist level-ground and stair-descent gait. Neural Networks 21(4):654-66.
- Mancinelli, C., B. L. Patritti, P. Tropea, R. M. Greenwald, R. Casler, H. Herr, and P. Bonato. 2011. Comparing a passive-elastic and a powered prosthesis in transtibial amputees. Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society 2011:8255-8.
- Herr, H. M., and A. M. Grabowski. Bionic ankle-foot prosthesis normalizes walking gait for persons with leg amputation. 2012. Proceedings B of the Royal Society 279:457-64.
- Ferris, A. E., J. M. Aldridge, C. A. Rábago, and J. M. Wilken. 2012. Evaluation of a powered ankle-foot prosthetic system during walking. Archives of Physical Medicine and Rehabilitation 93(11):1911-8.
- Sup, F., H. A. Varol, J. Mitchell, T. J. Withrow, and M. Goldfarb. 2009. Preliminary evaluations of a self-contained anthropomorphic transfemoral prosthesis. IEEE ASME Transactions on Mechatronics 14:667-76.
- Lawson, B. E., H. Varol, and M. Goldfarb. 2011. 2011. Standing stability enhancement with an intelligent powered transfemoral prosthesis. IEEE Transactions on Biomedical Engineering 58:2617-24.
- Sup, F., H. A. Varol, and M. Goldfarb. 2011. Upslope walking with a powered knee and ankle prosthesis: Initial results with an amputee subject. IEEE Transactions on Neural Systems and Rehabilitation Engineering 19:71-8.
- Lawson, B. E., H. Varol, A. Huff, E. Erdemir, and M. Goldfarb. 2012, Control of stair ascent and descent with a powered transfemoral prosthesis. IEEE Transactions on Neural Systems and Rehabilitation Engineering published online ahead of print.
- New York Daily News. 2012. Zac Vawter Climbs 103 Floors of Chicago's Iconic Willis Tower with Bionic Leg. November 5. www.nydailynews.com/news/national/man-bionic-leg-climbs-103-floors-article-1.1196727