Secondary Logo

Journal Logo


Assessment of Gait Symmetry in Transfemoral Amputees Using C-Leg Compared With 3R60 Prosthetic Knees

Petersen, Andreas Overbeck MSc; Comins, Jonathan RPT; Alkjær, Tine PhD

Author Information
JPO Journal of Prosthetics and Orthotics: April 2010 - Volume 22 - Issue 2 - p 106-112
doi: 10.1097/JPO.0b013e3181ccc986
  • Free

The microprocessor-controlled knee prosthesis; C-Leg (Otto Bock, Duderstadt, Germany), has been on the market for a decade. Several studies have investigated whether the gait of transfemoral (TF) amputees differs when subjects use C-Leg compared with conventional knee prosthesis. The gait pattern of TF amputees is characterized by 1) temporospatial gait asymmetry,1 2) increased stride width,1,2 and 3) reduced gait speed as compared with normal subjects.3 According to the manufacturer, C-Leg should give the user a more symmetric gait and the ability to selectively increase and decrease gait speed. From a subjective and visual perspective, the authors have previously noticed that subjects' movement patterns with regards to step length and kinematics seem to be more symmetric when they are equipped with a C-Leg. This article proposes to test this claim by quantifying the gait symmetry in TF amputees walking with two different prostheses: the C-Leg and the conventional 3R60.

Asymmetry in the gait pattern of TF amputees has been assessed quantitatively in several studies. Asymmetry in amputee gait is normally revealed by a prolonged duration of stance phase on the sound side, whereas the duration of the stance phase on the prosthetic side is similar to normal subjects.1 However, symmetry is influenced by gait speed. Thus, symmetry in TF amputee gait approaches normal values with increased gait speed.1 Some TF amputees exhibit a vaulting pattern, which is also an indicator of asymmetric gait. The ratio of the impulse of the vertical ground reaction force (GRF) between the sound and the prosthetic limb has been used as another objective measure of asymmetry and studies of this have reported that TF amputees exhibit greater asymmetry than normal subjects.4

A study of 50 TF amputees showed significantly improved kinematic gait symmetry when using C-Leg compared with conventional prostheses.5 In particular, knee joint angle was addressed and was shown to be improved in the physiological range. Furthermore, the study showed that the use of C-Leg resulted in better symmetry in load-duration and reduced loading on the prosthetic limb. Another study described walking with C-Leg as more natural than walking with a mechanical prosthetic knee because the C-Leg improved the knee flexion during loading response significantly.6 In contrast to this, Segal et al.7 observed a lack of knee flexion during loading response when subjects used C-Leg. They demonstrated that knee flexion for the prosthetic knees at loading and at opposite heelstrike was similar between walking with the C-Leg and the hydraulic prosthetic knee Mauch SNS.

The step length has been reported to be more identical between the sound- and the prosthetic side when subjects ambulate with C-Leg than with a hydraulic knee prosthesis.7 In contrast, one study reported a tendency toward decreased symmetry when TF amputees used C-Leg, but these results were not statistically significant.8

Johansson et al. compared several gait parameters including step length, single support, and double support. There was no significant improvement in investigated parameters when C-Leg was compared with the hydraulic Mauch SNS prosthesis.9 However, gait symmetry was not addressed in this study because no comparisons between the prosthetic and the sound limb were performed.

Factors other than type of prosthetic knee influence gait symmetry of TF amputees. Two studies have investigated the use of energy-storing feet and their effect on gait symmetry. Goujon et al.10 concluded that gait symmetry is not determined by type of prosthetic foot but rather by a subject's gait pattern. However, other parameters such as step length and propulsive action were improved when subjects used energy-storing feet. Graham et al.11 concluded that energy storing feet can lead to improved symmetry with regard to step length, but that they do not significantly improve stance phase duration or loading of the prosthetic limb when compared with conventional feet.

A previous study revealed that in 20 of 21 C-Leg users, the prosthesis was incorrectly aligned such that the sagittal knee axis in the standing position was posterior to the vertical projection of the GRF vector.12 Correct alignment, with the knee axis anterior to the weight-bearing line, is a prerequisite for proper function of the C-Leg, yet many C-Leg studies do not address this issue. It is particularly important to ensure accurate alignment when gait symmetry is investigated.

The purpose of this study was to compare the effect of using a C-Leg knee prosthesis and using a conventional knee prosthesis (3R60) on gait symmetry in TF amputees. The hypothesis of this study was that the microprocessor-controlled knee prosthesis C-Leg leads to an improved gait symmetry compared with gait with 3R60. 3R60 is a polycentric hydraulic knee prosthesis, which allows knee flexion during loading response. The two prosthetic knees were compared because both incorporate a hydraulic damping mechanism but their damping mechanisms are different: the 3R60 is a passive mechanical knee whereas the C-Leg uses a microprocessor to variably control the damping of the knee.

In this study, we introduce the concept of the butterfly plot (also called a Pedotti diagram) in the area of amputee gait. As far as we know, the use of butterfly plots as a qualitative or a quantitative clinical endpoint has not been addressed in the study of amputee gait. Vertical and horizontal GRFs combined with information of centre of pressure are shown in such a diagram. It is a plot of the direction and magnitude of the GRF vector emanating from the instantaneous centre of pressure during the entire stance phase. A butterfly plot can be useful to investigate a person's gait strategy,13 e.g., such a diagram can reveal the degree to which a person brakes the forward progression at heelstrike, or what the forward propulsive forces look like at toe-off. Furthermore, it is possible to compare the plot from the left and right leg to assess gait symmetry and gait pattern. In this article, we will also introduce a novel indicator for gait symmetry developed from the butterfly plot, and investigate if this parameter can be used as an objective and simple measure of asymmetry in the TF amputees.

In this study, four measures of symmetry were investigated: 1) spatial symmetry, expressed as an absolute symmetry index (ASI), 2) temporal symmetry, also expressed as an ASI, 3) duration of single support on the sound side and the prosthetic side, and 4) comparison of butterfly plots between the sound side and the prosthetic side, and across the two conditions.



Inclusion criteria included unilateral TF amputees between 25 and 65 years of age, amputation due to trauma or cancer, habitual C-Leg use, Functional Classification Level 3 or 4, use of the prosthesis at least 8 hours a day, and ability to walk without aids. Exclusion criteria included underlying musculoskeletal disease and neurological disease.

Nine unilateral TF amputees (eight men, one woman) were recruited to participate in the study, and five of these subjects (four men, one woman) completed the protocol. Two subjects withdrew before the study due to problems with their sockets, one subject completed the C-Leg test but due to difficulties while walking with the 3R60 he had to withdraw from the study, and finally one subject was excluded from the study because he could not carry out the tests correctly. The participants were 26 to 57 years old, 167 to 190 cm in height, and weighed 67 to 104 kg. Detailed information about the subjects is given in Table 1. Informed consent was obtained from all subjects before participation.

Table 1:
Detailed information about subjects


Two gait analyses were performed on each subject with a 1-week acclimation period between the two tests. The initial gait analysis was performed on the subject wearing the C-Leg, the second test involved the 3R60. Correct sagittal knee axis alignment was ensured by a highly experienced certified prosthetist/orthotist (CPO) with more than 20 years experience who confirmed that both prostheses were properly fitted and aligned. The CPO was particularly attentive to adjustment of the C-Leg such that the line of weight bearing was posterior to the knee center of the prosthetic knee. The subjects wore their habitual socket during both analyses. Each subject also wore his/her habitual prosthetic foot and shoes during both analyses. Thus, the only difference between the two analyses was the knee prosthesis. Subjects were asked to walk with a constant speed of 1.1 m/s in a standard gait laboratory. The speed of 1.1 m/s was chosen because this has been reported to be the mean gait speed of TF amputees,7 and because this parameter has an effect on several biomechanical variables.14 Approximately 15 to 20 trials were recorded for each subject to acquire trials with identical walking speed. Trials with higher or lower walking speed than 1.1 m/s (±0.05 m/s) were discarded. All subjects walked comfortably at this speed. Five successful trials from both the C-Leg and 3R60 analyses were used for statistical analysis.


Fifteen reflective markers were placed on each subject. The modified Helen Hayes' marker setup was used for this study. Markers were placed on the following anatomical landmarks: sacrum, anterior superior iliac spines, lateral femoral epicondyles, and lateral malleoli. Markers were attached to the shoes analogous to the second metatarsals, and calcanei. Additional markers were placed on wands on the thigh and the shank about 1/3 of the segment length from the distal end of the segment. A marker was placed on the prosthesis to indicate the center of rotation for the knee, and another marker was placed on a wand similar to the sound limb.

Vicon 460 (Oxford Metrics, Oxford, England) was used to collect kinetic and kinematic data. Kinetics were measured by two force platforms (AMTI Inc., MA) at a sampling frequency of 1,000 Hz. The positions of the reflective markers in 3D were collected at a sampling frequency of 100 Hz, and the positions were used to calculate the kinematics by the use of a link-segment model. Kinematic data were filtered using a quintic spline algorithm. Gait cycle events were determined by the Vicon system, and the correctness of the events was visually confirmed.


Assessment of gait pattern asymmetry was done by calculating the absolute symmetry index (ASI) as presented by Herzog et al.15 ASI is calculated as follows:

where I is the intact limb, and P is the prosthetic limb. ASI was used to reflect that perfect symmetry is revealed by a value of zero.

Spatial symmetry was calculated by the use of step length for the sound and the prosthetic limb. Prosthetic step length was defined as the distance from toe-off for the prosthetic limb to heelstrike for the prosthetic limb. Step length for the sound limb was defined as the distance from toe-off for the sound limb to heelstrike for the sound limb. Temporal symmetry was calculated by the use of duration of stance phase for the sound and the prosthetic limb.


In this study, the contour of the GRF vectors was displayed, and the area beneath the curve was calculated. Figure 1 shows such two butterfly plots on top of each other. We now introduce a novel parameter for describing gait symmetry by calculating the ratio of the area of the butterfly plot of the sound limb to the prosthetic limb:

Figure 1.:
Butterfly plots from one C-Leg trial for one subject. The dark-shaded butterfly plot shows the prosthetic limb, and the light-shaded butterfly plot indicates the sound limb. The BSR for this trial was 1.31.

Similar to the ASI, a butterfly symmetry ratio (BSR) of one indicates perfect symmetry. However, it must be noticed that it is possible for the ratio to be one even though the butterfly plots are not identical, e.g., if the area is differently distributed.


The two symmetry indexes and the BSR were analyzed by the use of paired t-tests. A linear mixed model was used to test whether differences in duration of stance phase, single support, and step length between the two conditions and the sound and the prosthetic side were statistically significant.

where i indicates the i'th subject, and j indicates the j'th measurement.

In all, 15 comparisons were performed, and Bonferroni-Holm corrections were used to ensure an overall level of significance of 0.05.



The duration of the stance phase was significantly shorter on the prosthetic side compared with the sound side in both C-Leg and 3R60 conditions (Table 2). No significant differences were observed for the duration of stance phase on the sound side in gait with C-Leg and 3R60 (Table 2). This was also the case on the prosthetic side (Table 2). Temporal symmetry was not significantly improved by the C-Leg (Table 3).

Table 2:
Mean value ± one standard deviation is given for duration of stance phase, single support, and step length on the sound side and the prosthetic side for both interventions
Table 3:
Mean value ± one standard deviation is given for the two symmetry indexes and the BSR

The step length was longer on the prosthetic side than on the sound side for both prostheses, but the difference was not significant (Table 2). No significant differences were observed between the step lengths on the sound side between the two prostheses (Table 2). The step length on the prosthetic side between the two prostheses did not differ (Table 2). The similar step length on the sound side between the two prostheses and on the prosthetic side resulted in no statistical difference between the two prostheses when spatial symmetry was calculated (Table 3).


Single support was significantly longer on the sound side than on the prosthetic side (Table 2). This difference in duration of single support between the two limbs was statistically significant for both prostheses. Table 2 shows that duration of single support of the sound limb was not significantly different between the two prosthetic knees. Similarly, duration of single support was not significantly different between the two prosthetic knees when the prosthetic limbs were analysed.


Table 3 shows the mean BSR for all subjects. The BSR was not significantly improved for subjects when using the C-Leg. Outlines of representative butterfly plots for two subjects are displayed in Figure 2. As can be seen from the figure, the butterfly plot is narrower on the prosthetic side than on the sound side for subject 1. This is the case for gait with both prosthetic knees, and it was the case for three out of five subjects. For subject 1, the angle of the GRF vector at toe-off is clearly more forward oriented on the sound side compared with the prosthetic side where the vector is more vertically oriented.

Figure 2.:
Outline of representative butterfly plots for two of the five subjects. Both data from the sound limb (solid line) and the prosthetic limb (dotted line) are given. Red lines indicate 3R60 trials whereas black lines indicate C-Leg trials. It is seen from the figure that the subjects seem to have adopted a gait pattern, which was not affected by the type of prosthetic knee.

The pattern of the butterfly plot for the sound limb was quite similar between the two tested conditions, and this was also the case for the prosthetic side. The same is seen for subject 4. However, the patterns of the butterfly plots differ remarkably between subjects which is obvious from Figure 2. Clearly, subject 1 has a more asymmetric pattern than subject 4.



One good thing about having so few subjects is that it provides the ability to investigate each subject carefully and separately. In this study, butterfly plots were used for this purpose. Most subjects showed a more forward oriented GRF vector at toe-off on the sound side than on the prosthetic side. The reason as to why the GRF vector is more vertically oriented on the prosthetic side is that the construction of the prosthetic foot does not allow an active plantar flexion, which can produce propulsive forces. We suggest that more advanced prostheses with two active components at the knee and the ankle, respectively, could be developed to solve this problem. This is a very complex task, and if such prosthesis is developed all amputees would be able to produce propulsive forces. Because some of the amputees in this study were able to produce propulsive forces with both limbs, we believe that intensive gait training is able to give the amputee a more symmetric and normal gait without introducing more advanced prosthetic knees. In this context, we believe that information about gait strategy through butterfly plots could be one objective parameter to help assist in gait training sessions as described below.

The differences in the outline of the butterfly plots indicate different gait strategies between the subjects in the study. Subject 1 had a much more forward oriented and larger GRF vector at toe-off on the sound side than the other subjects. Furthermore, this subject had a nearly vertical GRF-vector at toe-off on the prosthetic side. To generate propulsive forces, this subject therefore relies heavily on the function of the sound limb. In contrast, subject 4 showed an almost identical outline of the butterfly plot on the two limbs through the entire stance phase when using C-Leg. This subject therefore had a more symmetric gait than subject 1. The other subjects showed a gait symmetry somewhere between subject 1 and subject 4. However, none of the subjects improved their gait symmetry during the C-Leg condition, which was revealed by the non-significant BSR, and that the outlines of the butterfly plots were almost identical across the two conditions. It is seen from Figure 2 that the subjects seem to have adopted a gait pattern, which was not affected by the type of prosthetic knee. The conclusion that butterfly plots did not reveal improved gait symmetry when subjects used C-Leg corresponds well to the temporospatial symmetry indexes, which also did not reveal improved gait symmetry.

However, we suggest that the butterfly plots and the BSR could function as simple and objective clinical tools in evaluation and rehabilitation to improve the gait function, adaptation to the prostheses, and symmetry in TF amputees. For example, the use of butterfly plots could possibly help TF amputees change their gait pattern to obtain a more symmetric gait. It has been shown that TF amputees in just 4 minutes with real-time visual feedback were able to improve their gait symmetry.16 Among the visual feedback, was a plot of the vertical GRF, which was the subjects' preferred visual feedback. Although this feedback was not really a butterfly plot as both the horizontal GRF and the centre of pressure were excluded from the plot, the result indicates that the butterfly plot seems to be a promising clinical tool.

The authors emphasize that a quantitative method of comparing butterfly plots, which include the time aspect should be developed. Inclusion of the time aspect will most certainly lead to important knowledge about which part of the stance phase is the most crucial in regards to gait training and rehabilitation.


The subjects participating in this study reported that they had acclimated to 3R60 after 3 days. The comparable pattern of the butterfly plot between the two conditions confirms that the gait pattern normalized within 1 week with the 3R60. Although English et al.17 recommend a period of 3-week acclimation, we believe that the 1-week acclimation period was adequate for this study, and that a longer period would not have changed the results.

Different results regarding gait performance and asymmetry have been reported.5–9 This may be due to misalignment of knee axis to the weight-bearing line. In this study, correct alignment of the knee axis positioned anterior to the weight-bearing line was conditional and was confirmed by a CPO. Typically, it is not stated that the knee center is anterior to the weight-bearing line, which may have affected the results of these studies.6,7,9 We recommend that future analyses of C-Leg gait should clearly mention that a CPO has confirmed that the knee center is anterior to the weight-bearing line. Gait speed was controlled in the study to compare results across subjects. In a number of C-Leg studies, gait speed has not been controlled, which may have affected the results.6,18


The step length was longer on the prosthetic side than on the sound side, but the difference was not significant. The reason as to why the step length is longer on the prosthetic side could be that the propulsive forces are generated by the plantar flexors on the sound side, while this propulsive force cannot be generated on the prosthetic side as the prosthetic foot cannot actively push off. Another reason for the longer step length on the prosthetic side could be the significantly longer stance phase on the sound side compared with the prosthetic side.

Spatial symmetry was not significantly different between the two prostheses as was hypothesized. Similar to this study, Hafner et al.8 found that the step length on the sound side with C-Leg and a mechanical prosthesis was identical. However, they also found that the step length on the prosthetic side increased when subjects used C-Leg, indicating a tendency toward increased asymmetry for C-Leg compared with the mechanical prosthesis. The step length in this study was longer on both the sound side and the prosthetic side compared with the study by Hafner et al.8 The reason for this could be that subjects in this study were habitual C-Leg users, whereas subjects in the study by Hafner et al. had a mechanical prosthesis as their daily prosthesis. Segal et al.7 found that step length on the prosthetic side decreased when subjects used C-Leg, thereby improving spatial symmetry compared with a hydraulic prosthesis. This contrasts with the results of this study in which the use of C-Leg did not improve spatial symmetry. However, step length on both the sound and prosthetic side in our study was comparable with the results from Segal et al.7


We found that the duration of the stance phase was significantly longer on the sound side compared with the prosthetic side. The duration of stance phase on the prosthetic side was similar to data from normal subjects previously collected in our gait laboratory. However, the duration of stance phase was increased by an average of 6% on the sound side compared with normal subjects. The results are comparable with Jaegers et al.1 who found a duration of stance phase of 58% on the prosthetic side whereas it was 63% on the sound side. The longer stance phase on the sound side may be because the subjects rely considerably on the sound side to maintain balance and stability during walking.

Importantly, the duration of stance phase did not differ between the two tested conditions. Temporal symmetry was lower for gait with C-Leg than gait with 3R60, but when tested it did not lead to a statistically significant difference between the two conditions.


Single support was significantly longer on the sound side than on the prosthetic side for both prostheses. The difference in single support between the sound side and the prosthetic side was smallest for C-Leg, but it did not lead to any significant difference between the two prostheses as hypothesized.


The hypothesis that gait symmetry is improved with the C-Leg could not be statistically confirmed in this study. A trend was seen indicating that walking with the C-Leg improved temporal symmetry, and single support on the sound side. Only five subjects participated in the present study, which renders any conclusions based on such a small sample size tentative. However, the number of C-Leg users in the local area was limited to 15 subjects, and thus prevent the recruitment of an appropriate study sample.

The design in this study with inclusion of a homogenous group where subjects use the same prosthesis on a daily basis should clearly reveal if differences are present when they use another type of prosthesis. Other studies have been based on a mix of subjects who used hydraulic-, pneumatic-, and microprocessor-controlled knee prostheses. Some of these studies only allowed a short-acclimation period. Therefore, subjects in these studies might have changed their gait pattern solely due to inexperience with the prostheses. This can possibly have introduced incorrect results in the previous studies, which is the reason why only one group was included in this study. However, a more optimal design than the one applied would have included a group of users with conventional knee prostheses and a group of C-Leg users, where both groups were tested under both interventions. The results could then have been compared across the two groups, within each group as well as compare the sound limb versus the prosthetic limb. However, such a design would include many subjects and be very time consuming because earlier studies have pointed out a minimum acclimation period of 3.5 months when subjects have to acclimate to a C-Leg.6


This study investigated gait symmetry in TF amputees walking with the microprocessor-controlled knee prosthesis C-Leg and the hydraulic knee prosthesis 3R60. Neither temporospatial symmetry, BSR, nor duration of single support revealed a significantly improvement when using the C-Leg. The butterfly plot showed that subjects had clearly different gait strategies, but that the within-subject gait strategy was not influenced by type of prosthetic knee. Therefore, results from the butterfly plots led to the same conclusion as the outcome of the temporospatial parameters that C-Leg did not improve gait symmetry. However, it is suggested that the butterfly plot could be a promising clinical tool in the future to achieve better gait symmetry in TF amputees.


We thank Christer Levin, CPO, from Sahva for carrying out the alignment of the prostheses in this study, and Klaus Kähler Holst, MSc, Department of Biostatistics, University of Copenhagen, for guidance and advice on how to carry out the statistical analysis.

We also thank Otto Bock Scandinavia AB, Box 623, 601 14 Norrköping for lending 3R60 prosthetic knees to Sahva in order to carry out this study.


1. Jaegers SM, Arendzen JH, de Jongh HJ. Prosthetic gait of unilateral transfemoral amputees: a kinematic study. Arch Phys Med Rehabil 1995;76:736–743.
2. Vaughan CL, Davis BL, O'Connor JC. Dynamics of Human Gait. Cape Town: Kiboko Publishers; 1999.
3. Boonstra AM, Schrama JM, Eisma WH, et al. Gait analysis of transfemoral amputee patients using prostheses with two different knee joints. Arch Phys Med Rehabil 1996;77:515–520.
4. Nolan L, Wit A, Dudzinski K, et al. Adjustments in gait symmetry with walking speed in trans-femoral and trans-tibial amputees. Gait Posture 2003;17:142–151.
5. Nimmervoll R, Kastner J, Kristen H. The C-Leg experience—a gait analysis comparison with the conventional prosthetic knee joints. Orthopädie-Technik 2003;54:117–125.
6. Kaufman KR, Levine JA, Brey RH, et al. Gait and balance of transfemoral amputees using passive mechanical and microprocessor-controlled prosthetic knees. Gait Posture 2007;26:489–493.
7. Segal AD, Orendurff MS, Klute GK, et al. Kinematic and kinetic comparisons of transfemoral amputee gait using C-Leg and Mauch SNS prosthetic knees. J Rehabil Res Dev 2006;43:857–870.
8. Hafner BJ, Willingham LL, Buell NC, et al. Evaluation of function, performance, and preference as transfemoral amputees transition from mechanical to microprocessor control of the prosthetic knee. Arch Phys Med Rehabil 2007;88:207–217.
9. Johansson JL, Sherrill DM, Riley PO, et al. A clinical comparison of variable-damping and mechanically passive prosthetic knee devices. Am J Phys Med Rehabil 2005;84:563–575.
10. Goujon H, Bonnet X, Sautreuil P, et al. A functional evaluation of prosthetic foot kinematics during lower-limb amputee gait. Prosthet Orthot Int 2006;30:213–223.
11. Graham LE, Datta D, Heller B, et al. A comparative study of conventional and energy-storing prosthetic feet in high-functioning transfemoral amputees. Arch Phys Med Rehabil 2007;88:801–806.
12. Willingham LL, Buell NC, Allyn KJ, et al. Measurement of knee center alignment trends in a national sample of established users of the Otto Bock C-Leg microprocessor-controlled knee unit. J Prosthet Orthot 2004;16:72–75.
13. Boccardi S, Chiesa G, Pedotti A. New procedure for evaluation of normal and abnormal gait. Am J Phys Med 1977;56:163–182.
14. Lelas JL, Merriman GJ, Riley PO, Kerrigan DC. Predicting peak kinematic and kinetic parameters from gait speed. Gait Posture 2003;17:106–112.
15. Herzog W, Nigg BM, Read LJ, Olsson E. Asymmetries in ground reaction force patterns in normal human gait. Med Sci Sports Exerc 1989;21:110–114.
16. Davis BL, Ortolano M, Richards K, et al. Realtime visual feedback diminishes energy consumption of amputee subjects during treadmill locomotion. J Prosthet Orthot 2004;16:49.
17. English RD, Hubbard WA, McElroy GK. Establishment of consistent gait after fitting of new components. J Rehabil Res Dev 1995;32:32–36.
18. Schmalz T, Blumentritt S, Jarasch R. Energy expenditure and biomechanical characteristics of lower limb amputee gait: the influence of prosthetic alignment and different prosthetic components. Gait Posture 2002;16:255–263.

amputees; artificial limbs; rehabilitation; gait; biomechanics

© 2010 American Academy of Orthotists & Prosthetists