Microprocessor Knee Symposium

Doug Smith, MD; Margrit R. Meier, PhD, CPO; Kim De Roy, MSc, RPT, CPO; Scott B Elliott, BS, BME, CP

Outcome Assessment of a Mr Microprocessor-Controlled Knee

Scott B. Elliott, B.S. BME, C.P.
Ossur North America
Aliso Viejo, California, US

The Rheo Knee^TM by Ossur is a swing and stance control knee system that provides appropriate resistance to flexion and extension through the use of a microprocessor, integrated sensors, and an innovative magnetorheologic fluid actuator. Fine adjustment of swing and stance control parameters may be completed by the prosthetist with the use of a hand held PDA. The main objectives of this operational test were to evaluate the function of production models of the Rheo Knee as well as to verify market acceptance of the new device with users and prosthetists. This paper will focus on discussion of the user questionnaire that was one tool used to evaluate functional performance of the Rheo Knee on the first production units.

13 patients were fit with the Rheo Knee: eight in USA, one in Iceland, two in Germany, one in Holland, and one in France. Subjects were allowed to use an existing socket/frame configuration if alignment was compatible with the Rheo Knee and the existing fit was comfortable for the user. The attending prosthetist was instructed to fit one of the following Flex-Foot systems with the Rheo Knee: Ceterus, Vari-Flex, Talux, or Elation. Table 1 outlines the details of each test user included in the test.

Following fitting of the new prosthesis, the user was instructed to complete a variety of tasks over a minimum of four weeks including: Level ground walking, terrain and cadence transitions, ramp and stair walking, uneven ground walking, seating, and transfers. The user was then asked to evaluate the performance of the Rheo Knee by completing a questionnaire which examined the following: level ground walking, changing walking speed, ramp and stair descent, uneven ground walking, seating, special situations and effort, maintenance, and quality and consistency of behavior.

The user questionnaire has a total of 34 questions. Each question is divided into two questions, a basic question regarding test knee performance and a comparative question relating to the previous or 'existing' prosthesis. In the discussion, the first part will be referred to as subject base and the second part of the question as subject comparison. The user marked his or her ratings on a slide bar (see Figure 1).

At the conclusion of the testing period, subjects were also asked to list the five most beneficial features and five most unfavorable features of the Rheo Knee. These comments are summarized in Tables 4 and 5.

The results are divided into objective and subjective. The objective results summarize the responses on the questionnaire in terms of the average response within a particular performance category. The baseline response is given as well as the comparative response. Maximum and minimum response values are defined as well. The subjective results are a summary of comments taken directly from the users following the test. These responses consist of statements that were made by more than one individual at the conclusion of the testing period.

The results suggest that the Rheo Knee provides beneficial function to the user with respect to all performance categories (see Figure 2). This is due to the fact that the average response within each baseline and comparative performance question was above 50% ('neutral or equal' response level) except for one question – weight of the knee (49%). This indicates that for each baseline and comparative performance category, the user responded that the knee performed in a positive manner, 'equal to' or 'better than' the existing prosthesis.

Highest average baseline ratings are summarized in Table 2 and relate to high quality of swing motion and enhanced stance control functions. High values are also found when addressing the comparative values in these categories indicating that the user greatly prefers the Rheo Knee when compared to the existing prosthesis. The subjective statements summarized in Table 4 tend to back up the objective findings relating to improvements in swing and stance phase control and feel for the user. Users are particularly fond of the natural walking motion and cadence responsive characteristics produced with this particular system. This is most likely due to the low-drag characteristics of the MR actuator coupled with dynamic learning software control for auto-adaptive swing phase resistance. The enhanced knee stability for terrain transitions, stair descent, uneven ground walking, and advanced stumble recovery are also shown to be favorable for users. Force sensors and rapid sampling and command rate (1000Hz) identify when loads are placed on the prosthesis so that instant support may be given to the user during weight bearing. Multiple stance release parameters within the software ensure that stance is activated and deactivated at the appropriate times for the user while walking on a variety of terrain. Less conscious thought and effort is then required of the user to voluntarily maintain knee stability.

The high-end average baseline ratings tend to show functional performance areas that are quite noticeable to the user. However, it is interesting to examine some of the low-end average baseline values (see Table 3). Notice that the average baseline rating for walking effort on the Rheo Knee was 56%. According to the scale used on this question, this indicates that the user's walking effort is 6% less than 'neutral' or closer to 'less effort.' When looking at the comparative value, one can see that this is one comparative value that actually increases beyond the baseline value to 77%. This indicates that this decrease in walking effort equates to a great increase in the user preferring the Rheo Knee to the existing prosthetic knee. This is most likely due to the fact that TF amputees expend much more energy during normal ambulation when compared to non-amputated individuals. Therefore, any energy saving quality in a product is much appreciated by the user. It is also interesting to note that the users perceived the Rheo Knee to be 'equal' in terms of weight compared to the existing prosthesis. Although the knee is significantly heavier than any of the existing prosthetic knees used by the test group, this equated to only a 1% negative change in the average preference of the user. This may be due to the fact that the weight of the knee is centered around the actuator and battery which are positioned in the most proximal portion of the knee. Therefore inertial effects of weight during swing phase are minimized in the current design. One may then conclude that weight itself is not the determining factor of preference or perception of the prosthetic knee user but the combination of weight distribution and functional performance output of the device may be the more influential factors.

According to the subjective comments of the users, the most negative attributes of the Rheo Knee centered on daily maintenance (i.e. charging) and compatibility with activities such as kneeling (see Table 5). Other negative comments were directed at cosmesis (lack of soft outer cover) and incompatibility with clothing.

Performance On An Obstacle Course: Otto Bock C-Leg VS. Otto Bock 3R60 VS. Catech SNS

1 Margrit R. Meier, PhD, CPO, Andrew H. Hansen, PhD 1,2, Steven A. Gard, PhD, 1,2
1 Northwestern University Prosthetics Research Laboratory and Rehabilitation Engineering
Research Program; 2 VA Chicago Health Care System
Chicago, Illinois

Introduction: The C-leg® is considered by many prosthetists and manufacturers to be the leading microprocessor-regulated knee mechanism currently available on the market. What sets it apart from others is its hydraulic knee unit with microprocessor-controlled stance and swing phase damping characteristics that provide monitoring and intervention capabilities during the entire walking cycle.

Previous investigations of microprocessor-controlled knee joints have reported mixed results, ranging from data showing clear benefits for amputees to those that suggest there is no difference at all when compared to conventional knee mechanisms [1-4]. The reported results are not sufficient to objectively determine the benefits of the C-leg® as most of the reported studies are related to the Intelligent Prosthesis (IP) by Blatchford. Objective evidence is needed to determine if there is significant benefit when prescribing expensive microprocessor-controlled knee mechanisms over high-performance passive knee units that cost significantly less. In this short paper the following specific objectives of the study are presented: (1) to determine the participants' walking performance while walking over an obstacle course and (2) to examine the influence of mental loading while walking over the obstacle course with three different knee units.

Methodology: General: In a crossover study design, each participant wore each prosthetic knee joint-Otto Bock C-leg, Otto Bock 3R60 and Mauch SNS-for a period of four weeks. Test prostheses were fabricated using a duplication of the participant's current prosthetic socket, and each participant was fitted with a Dynamic Plus foot. Participants: Persons with unilateral transfemoral amputation, aged between 40 and 60 years, with a body-weight less than 125 kg, were included in the study if they presented with no serious complications that interfered with their walking ability; had six or more months of experience with a definitive prosthesis; were able to walk unassisted at a comfortable speed without undue fatigue and without health risk; and were able to climb stairs. Protocol: The obstacle course was set up in the VA Chicago Motion Analysis Research Laboratory (VACMARL). It consisted of foam section (3m long, 1m wide, 20 durometer on a shore A scale), narrow slaloms around three chairs, a vacuumized bean-bag section (3m long, 1m wide) simulating sand, a rock section (3m long, 1m wide), a short downward sloping ramp (1.5m long, 1.4m wide), a 90-degree left turn, and a final stair step (height: 12cm) (Figure 1). The mental loading test consisted of a mathematical calculation task where the participant had to count vocally backwards in 3-step increments (first visit), in 7-step increments (second visit) and in 3-step increments (third visit). Participants completed the obstacle course twice, once without mental loading, and once with mental loading. No familiarization run was allowed. They were videotaped allowing time to be measured. Statistical Analysis: Due to the non-parametric data distribution Friedman Test was used to assess the overall performance of the three knee joints. If a variable reached significant level, Wilcoxon Signed Rank Test was used to test between differences of each knee joint. A Bonferroni correction was applied to account for multiple testing, lowering the significance level to 0.016.

Results: Data from 11 participants, two females and nine males, were analyzed. Their mean age was 45.8 ±9.5 years, mean height was 175 ± 9 cm, and mean weight was 81.8 ± 14.1 kg. They were all established walkers with their amputation having occurred 20.1 ± 14.2 years ago. Seven participants had their amputation due to a traumatic incidence, one due to Peripheral Vascular Disease (PVD), two due to infection and one due to a congenital deficiency. Three out of the 11 participants had a knee-disarticulation amputation.

The median time taken to complete the obstacle course with the 3R60 knee joint was 34.9seconds (s), the minimum time (min.) was 23.9s and the maximum time (max.) was 84s. Adding the mental task altered the time only minimally: 34.2s (min. 22.9s, max. 82s). For the Cleg, the total time was slightly lower when compared to the 3R60 knee joint: median time 32.1s (min. 22.1s, max. 73.1s). By adding the mental task the median time for the C-leg increased to 33.9s (min. 18.1s, max. 69.8s). The difference between the 3R60 knee joint and the C-leg was non-significant for both conditions (without mental task: p=0.169; with mental task: p=0.045). Participants performed best on the obstacle course when fitted with the SNS unit. Their total median time without mental task was 30.9s (min. 26s, max. 75.2s). Adding the mental task increased the median time slightly to 32s (min. 23.8s, max. 75.2s). The difference between the SNS and the 3R60 knee joint was significant for both conditions (without mental task: p=0.011; with mental task: p=0.005). However, the difference between the C-leg and the SNS knee joint was non significant (without mental task: p=0.674; with mental task p=0.678) (Figure 2).

Discussion: The participants completed the obstacle course in the shortest time when fitted with the SNS knee joint, followed by the C-leg, and they were slowest with the 3R60 knee joint regardless if a mental task was administered or not. Roughly summarized: the more complex knee joint (3R60) and the more sophisticated knee joint (C-leg) performed less favorable in the given context. It could be that the more complex and sophisticated knee joints require more time and training in order for the user to be able to take full advantage of their characteristics. Thus the given 4-week accommodation period may not have been enough. However, it could also mean that for soft or uneven walking surfaces, a simpler knee joint-represented by the SNS knee joint-simply performs better, as participants have a quicker and direct impact on its behavior.

The mental task had its biggest impact on the C-leg: participants performing with the C-leg reduced their performance speed by 6% compared to their non-mental task performance. The SNS knee joint induced only a 4% speed reduction compared to the non-mental task performance. This may indicate that the microprocessor-driven knee joint did not reduce mental loading during the obstacle performance as anticipated. Participants slowed down more with the C-leg to perform the two tasks simultaneously-walking safely and calculating correctly-than with the SNS unit. In contrast to the two single-axis knee joints, performance with the multi-axis knee joint (3R60) and the mental task enhanced participants' speed slightly by 2%, possibly indicating the influence of the stability provided by the positioning of the knee's instantaneous center of rotation [5]. In-depth detailed analysis may be possible to perform once the aimed sample size of 15 participants has been reached.

Limitations: The obstacle course was set up within the gait laboratory and thus represented a controlled environment that may not be representative of outdoor conditions. However, the different walking surfaces and narrow curved pathways demanded higher ambulation skills than walking on the level laboratory surface, thus challenging participants' performances.

Acknowledgements: This work was supported by the Department of Veterans Affairs, Rehabilitation Research and Development Service, and is administered through the Jesse Brown VA Medical Center, Chicago, IL.

References

Smith DG, Willingham LL, Allyn KJ, Buell NC and Hafner BJ. Functional evaluation of the transition from a non-micropocessor controlled prosthesis to a micro-processor controlled prosthesis for transfemoral amputees: Early results of a clinical trial. International Society for Prosthetics and Orthotics (ISPO), 11th World Congress, Hong Kong, August 1-6, 2004.
Stinus H. Biomechanik und Beurteilung des mikroprozessorgesteuerten Exoprothesenkniegelenkes C-leg. (Abstract in English). Zeitschrift für Orthopädie und ihre Grenzgebiete 2000; 138: 278-282.
Buckley JG, Spence WD and Solomonidis SE. Energy cost of walking: comparison of "Intelligent Prosthesis" with conventional mechanism. Archives of Physical Medicine and Rehabilitation 1997; 78: 330-333.
Datta D and Howitt J. Conventional versus microchip controlled pneumatic swing phase control for transfemoral amputees: user's verdict. Prosthetics and Orthotics International 1998; 22: 129-135.
Gard SA, Childress DS and Uellendahl JE. The influence of four-bar linkage knees on prosthetic swing-phase floor clearance. Journal of Prosthetics and Orthotics 1996; 8(2): 34- 40.

Figure 1
Overview of obstacle course set-up within the VACMARL laboratory: Foam Section (3m long), Slalom Section around three chairs, vacuumized Bean Bags to mimic sand (3m long), Rock Section (3m long), Ramp (1.5m long), Corner (90° degree left turn) and a Step (12cm high). Two video cameras were set in such a way that the entire obstacle course could be filmed, allowing time measurements for each section.

Figure 2
Total time taken (in sec) to complete the obstacle course for each prosthetic knee joint.
w/o MT: without Mental Task
* SNS-3R60: p=0.011
w MT: with Mental Task
** SNS-3R60: p=0.005

Powered Prosthetics: Clinical Outcome of Unprecedented Function

Kim De Roy, MSc. R.P.T. - CPO
Össur hf
Reykjavik, Iceland

The Power Knee¹ is a prosthetic knee solution intended to provide mechanical power to replace knee joint kinematics and compensate for lost muscle strength in above knee amputees. A motorized actuator module generates power according to the users' need to adequately endure different portions of locomotion. Portions requiring specific power management are level ground walking, stair- and incline a- or descending, sitting down and standing up. The motorized kneeunit substitutes the ex-centric and concentric muscle-work required during this type of actions, allowing natural effortless performance. The system aims to reduce effort in level ground and incline walking and to offer unprecedented function such as reciprocal stair-climbing as shown in Figure 1.

The actions of the Power Knee are partially controlled by sensory data collected at the sound side. A combination of predefined strategies related to the positioning and loading of the sound and the prosthetic side will result in predefined actions of the leg. The user controls the prosthesis according to the portion of locomotion he or she is confronted with.

The artificial intelligence of the motorized prosthesis operates on high and low-level layers in order to continuously observe the whole state of the respective human-system interface. The high level is responsible for the management of biomechanical events and of the amputeesprosthesis interaction. The low level artificial intelligence manages the artificial knowledge, the systems response and the interaction between units. Finally, a user interface (software) is available for individual fine-tuning of the function of the prosthesis during the different portions of locomotion. Adjustments can be made based upon inter-individual anthropometrical differences as well as on personal physical conditions and level of rehabilitation.

The purpose of the study was to investigate the clinical outcome and elaborate on the clinical value of this type of innovation. At the same time the opportunity was used to evaluate the required process of adaptation and rehabilitation. Giving the fact that it concerns an unprecedented prosthetic solution based on new technology the experimental protocol was intended to evaluate the robustness, repeatability, usability, and stability of the Power Knee. For that purpose both controlled and noncontrolled environment activities were investigated. Part of the protocol was intended to verify the prosthesis behaviors in the subjects� normal environment.

Ten subjects participated in clinical trials evaluating different aspects in the functionality of the Power Knee. All subjects were unilateral above knee amputees in good physical condition, aged between 20 and 65 years. The subjects were equipped with a complete prosthesis kit including the prosthetic module itself, a charger, APM2-module and IPO3. A complete documentation was provided including a user manual and logbook.

The results for the different test protocols are summarized per protocol. A first protocol described testing in an uncontrolled home environment and lead to the following conclusions. Through several experimental trials, the subjects grew in their ability and confidence. According to their comments, they needed a familiarization period at first to adapt their movements with the active prosthesis to their usual environment. After the period of adaptation, they were able to complete a wide range of activities and appreciated the multiple possibilities offered by the prosthesis. The most frequently mentioned advantages were:

Gait comfort for longer distances
Gait quality
Cadence adaptation
Support for incline walking
Increased walking speed
Ability for reciprocal stairs ascend
Stability of the system while walking on irregular surfaces (i.e. mud, grass, gravel)

In a second experimental protocol the subjects' capacity to trigger all appropriate portions of locomotion supported by the prosthesis system was evaluated including: standing, walking on a leveled floor, walking up and down inclined planes, going up and going down staircases, as well as sitting down on a chair and standing up from a chair. Within one session different individuals reach different ability levels. Certain subjects are able to initiate the detection of all portions of locomotion with different levels of ability. Other subjects require more time in controlling some tasks, but still had a good comprehension of what was required to correctly initiate all portions within two sessions of 2-3 hours. Specific tasks can be distinguished as more difficult to learn. As expected, the portions which required the most training were stairs related. This can partially be explained through the fact that reciprocal stairs ascend is a task which most patients haven't performed since their amputation and requires a specific training. Gradual increase of the power provided by the prosthesis led to early adaptation and steady progression.

A third protocol investigated the time required for the user to obtain full control over the prosthesis to perform and fully complete all different portions of locomotion. On average subjects require 3-4 sessions of or a total time of 7-9 hours to become experienced in the completion of the different portions of locomotion using the Power Knee. The stairs trials were most complex portions to train. When initiating going upstairs and down-stairs, the subjects need to initially control the required movements. In order to ensure user-safety, training should be done is a controlled environment. For stair ascent, it is recommended to perform thorough training on the triggering of this portion of locomotion prior to proceeding to cyclic upstairs walking. For stair descent, particular emphasis should be made on proper foot clearance after sound limb toe strike as well as proper gait symmetry. During the early stages of rehabilitation and for users with a lower ability level it is recommended to limit stair ascent and descent using modes that provide a stiffleg which fully supports the user and ensures users� safety.

In conclusion to the results of the completed experimental protocols it is clear to state that the Power Knee enables the above knee amputee to regain a broad, if not complete, spectrum of daily lifeactivities. All subjects successfully adapted to the powered prosthesis and were able to safely perform unprecedented as well as traditional activities. Based upon the subjective comments recorded from the different subjects throughout the trials, advantages related to improved energy efficiency are likely, yet hypothetical at this stage and will require further investigation. The usability of the system is expected to of great benefit to a broad range of above knee amputees. Although no user will have the same ability in using the prosthesis, limited time will be required to complete a full training on the prosthetic system. The complexity of certain tasks causes no limitation to the overall performance on the Power Knee. Learning capabilities of subject are dependant on motivation, functional level, physical condition and cognitive abilities. The implementation of the rehabilitation function in the user-interface has obvious benefits to the early adaptation and the clinical outcome of the subject using the prosthesis.

Figure 1: Reciprocal stairs ascend using the Power Knee