Secondary Logo

Journal Logo

Original Article

Impact of Increased Prosthetic Mass on Gait Symmetry in Dysvascular Transfemoral Amputees

A Randomized Prospective Double-Blind Crossover Trial

Moylan, Beverly MD, MA; Paner, Ruby BSc; Pauley, Tim MSc; Dilkas, Steven MD, FRCPC; Devlin, Michael MD, FRCPC

Author Information
Journal of Prosthetics and Orthotics: April 2015 - Volume 27 - Issue 2 - p 63-67
doi: 10.1097/JPO.0000000000000056
  • Free


The trend in the prosthetic industry has been to encourage the use of lightweight prostheses. This can occur at a significant financial burden to patients with lower-limb amputation and/or funding programs. Lightweight titanium components (knee, tubing, couplers) for a typical transfemoral prosthesis (endoskeletal design, plastic laminate socket, Silesian suspension) decreases prosthetic mass from 1075 g to 790 g (27% reduction; socket, foot, and cosmetic cover not included) but increases cost from Can $1313 to Can $1912 (46% increase).1,2

This trend began when it was shown that the addition of 2-kg weights to each foot of a normal person increased the energy expenditure of gait by 30%.3 This led to the belief that a lighter prosthetic leg would lower the energy expenditure of amputee gait. As a result, current lightweight transfemoral prostheses are on average approximately one-third the mass of the intact limb, even with the addition of a hydraulic knee unit.4

The optimal mass of a prosthetic limb is unknown. Substantial design efforts have been directed at reducing prosthetic limb mass with the goal of reducing energy expenditure. It has been shown, however, that the addition of weight to the shank does not increase metabolic costs of energy expenditure at multiple walking speeds.4 In fact, because of the importance of terminal swing phase energy transfer, there may be negative metabolic consequences of an excessive reduction in prosthetic mass.5

The greatest changes in mechanical work from both the hip joint transfer sources and hip muscles occur in the leg during swing phase acceleration and deceleration. More specifically, three sources of mechanical work assist in forward propulsion of the prosthetic limb: 1) the prosthetic foot; 2) the hip flexors; 3) and the hip joint.6 In transfemoral amputees, the transition between the acceleration phase and the deceleration phase is the point in swing phase when there is a reversal of energy transfer across the hip joint from transfer into the leg to transfer out of the leg, and into the trunk.6 It appears that amputees are able to effectively conserve the additional mechanical work required to propel a heavier limb during preswing (push-off) and initial swing by recovering this energy during the terminal swing deceleration of the limb, and transferring it to the trunk.6

It has not been shown that the use of a lightweight prosthesis is metabolically or mechanically beneficial for transfemoral amputees.4,6 There is conflicting evidence in regards to patient preference.7,8 One study looking at dysvascular transfemoral amputees found that more than half the participants preferred a weighted prosthesis over a “placebo” weight.8 Another study looking at traumatic transfemoral amputees showed that 8 of 10 participants preferred their original unweighted prosthesis.7

Most lower-limb amputations in North America are performed on older persons as a result of peripheral vascular disease (PVD).9–11 These amputees experience considerable medical comorbidities and have some degree of muscle weakness, yet preferred a heavier prosthesis.8 It is possible that a heavier prosthesis provides a sense of stability.7

Gait asymmetry has been demonstrated to increase the risk of falls in the elderly population.12 If the same holds true in transfemoral amputees, perhaps a more symmetric gait provides increased stability, thus reducing the likelihood of falls. The purpose of this study was to investigate spatiotemporal gait symmetry and patient preference with the addition of mass to a transfemoral prosthesis.



A single-center, prospective, double-blind, crossover trial design with randomization of testing conditions was used. Ethics approval was obtained from the Joint Bridgepoint Health/West Park Healthcare Centre/Toronto Central CCAC/Toronto Grace Health Centre Research Ethics Board. This is a joint institutional review board that operates in compliance with the Tri-Council Policy Statement, International Conference on Harmonisation/Good Clinical Practice (ICH/GCP) guidelines, the Ontario Personal Health Information Protection Act, and Part C, Division 5 of the Food and Drug Regulations of Health Canada.


A convenience sample of patients attending an outpatient amputee clinic as well as those admitted to an inpatient rehab program was screened for study inclusion. Inclusion criteria were as follows: unilateral transfemoral amputee, age 50 years or older, amputation secondary to PVD, at least 2 to 3 weeks postfitting with a prosthesis, and demonstrated ability to be an independent community ambulator, with or without a gait aid. Patients were excluded if they did not meet all inclusion criteria or if there was cognitive impairment sufficient enough to limit the ability to participate. Written informed consent was obtained for all subjects before study initiation.

Demographic and clinical characteristics are included in Table 1. Residual limb length was measured similarly to a previous study by taking the distance from the distal end of the inside of the prosthetic socket to the ischial shelf.8 Prosthetic mass includes the mass of the prosthesis, suspension, shoe, and socks.

Table 1
Table 1:
Demographic and prosthetic data


Similarly to a previous study,8 each subject used his/her own prosthesis during testing. The prosthesis was modified in 3 ways: the addition of a 180-g mass (light placebo condition), 726-g mass (medium condition), and 1633-g mass (heavy condition). The mass was added 14 cm inferior to the knee joint. The mass added consisted of standard physiotherapy weight belts with a vinyl cover (to prevent visual distinction) and Velcro straps. The weight belts held up to 10 removable 180-g weights. The number of 180-g weights was adjusted for each mass condition. The weight belts were attached over the endoskeletal prosthesis. Mass conditions were applied in a random order by the primary investigator who was not involved in any of the data analysis. The subjects were not allowed to touch the weight belts to ensure blinding. The research assistant responsible for processing the data for analysis, the data analyst, and all subjects remained blinded to the mass added until after study completion.

After application of each mass, subjects were given 5 to 10 minutes to walk and familiarize themselves with the new prosthetic mass, followed by a 2- to 5-minute rest before performing each walk test. They had a 5- to 10-minute rest between subsequent mass conditions. On the command “go,” subjects walked the length of the GAITRite mat, stepped off the mat, safely turned around, and walked back the length of the mat. This was considered one full walk. Subjects started each walk at a sufficient distance to allow 4 to 6 footfalls before stepping on the mat. This allowed sufficient distance before and after measurements were taken to minimize the effects of acceleration and deceleration. Subjects were instructed to take an additional six steps after stepping off the mat to eliminate braking. If a subject did brake or pause, the walk was repeated.

The 726-g mass was chosen because it is similar to the mass saved by using lightweight components (titanium components plus lightweight solid ankle cushioned heel [SACH] foot). The rationale for adding a mass of 1633 g was because it is approximately twice the previous mass, and it approximates the mass found to produce the most energy-efficient gait (Vo2/m) in a previous study.13 A placebo mass of 180 g was used because it was approximately equal to the mass of the cosmetic cover (foam fabric Velcro) used to make the 3 masses visually indistinguishable in an earlier study.8 In addition, it has been shown that transfemoral amputees cannot detect the addition of 226 g to the prosthesis and that this did not result in any detectable change in any of the gait parameters measured.14 Each mass was added at a location 14 cm inferior to the knee axis because it has been shown that mass added to the proximal shank results in more energy-efficient gait (Vo2/m) than mass added above the knee or to the distal shank.13 Movement of the mass relative to the prosthesis was not apparent during the walk tests.


The primary outcome measures consisted of absolute and relative spatiotemporal gait asymmetry and patient preference. Absolute gait measures were identified a priori as step time, step length, stance time, single support time and percentage of gait cycle, step width, and heel off/on time of the amputated limb. Degree of asymmetry (DoA)15 was used to express spatiotemporal gait measures of the amputated limb relative to the intact limb. Degree of asymmetry was calculated for the various spatiotemporal gait measures using the following formula: [(intact side value – prosthetic side value)/(intact side value + prosthetic side value)]. Degree of asymmetry greater than zero indicates greater intact side values; DoA less than zero indicates greater prosthetic side values; DoA equal to zero indicates perfect symmetry. Regardless of side of amputation (i.e., left vs. right), all spatiotemporal gait measures were normalized to the amputated limb (i.e., amputated limb vs. intact limb).

A portable, flexible walkway with embedded pressure-sensitive switches was used to analyze gait. The GAITRite was used to analyze spatial parameters (i.e., step and stride lengths), temporal parameters (i.e., step time), and derived measures of rate in a subject walking at different speeds and various degrees of step symmetry. The validity and reliability of the GAITRite compared with validated gait analysis methods has been demonstrated,16 as well as test-retest reliability in people with stroke.17 The mat is composed of a series of sensors organized in a grid pattern, sandwiched between two layers of vinyl, and gives the visual impression of a carpet runner.

Patient preference was determined by asking the subjects to state their preferred condition after the three walking trials. They were also asked to elaborate on why they preferred this choice.


Descriptive statistics were used to summarize and describe the data. Repeated-measures analysis of variance (ANOVA) was conducted to test for differences in spatiotemporal gait measures across prosthetic mass conditions. Where significant main effects were detected, comparisons between paired experimental conditions were conducted using paired sample t-tests for within-subject comparisons. The Holm’s sequential Bonferroni procedure was used to control for the increased risk of type I errors associated with multiple pairwise comparisons. Analyses were conducted using SPSS version 21 for Windows.


Ten subjects met study inclusion criteria and completed the study. Baseline demographic data are recorded in Table 1. The average age was 64 years. Eight participants walked using a gait aid. These were the gait aids used by the participants on a regular basis and included single-point canes, two-wheeled walkers, and four-wheeled rollators. There were no adjustments made to the components of the individual prostheses; thus, the six subjects who ambulated on a daily basis with a locked knee maintained a locked knee for testing. There was a wide range of time from initial prosthesis fitting—from 3 weeks to 53 months.


Repeated-measures ANOVA was used to test for any differences in spatiotemporal gait measures across prosthetic mass conditions on both the amputated and intact sides. The results are summarized in Table 2. Overall, there was no change in velocity across any of the three weight conditions (F2,18 = 0.13, p = 0.88). No differences were detected between the amputated and intact sides for step time (F2,18 = 2.38, p = 0.12), step length (F2,18 = 0.99, p = 0.39), single support time (F2,18 = 0.45, p = 0.64), or step width (F2,18 = 1.22, p = 0.32). With the exception of step width, the results were similar when expressed in DoA. Of note, step width on the GAITRite is measured from the midline midpoint of the current footprint to the midline midpoint of the previous footprint on the opposite foot. This can result in different measures of step width between the right and left foot of a given individual, particularly where the forward progression of one foot is much shorter than the contralateral foot. Therefore, separate measurements for the step width of the intact and prosthetic legs were obtained.18 No differences were detected for step time (F2,18 = 2.99, p = 0.08), step length (F2,18 = 1.30, p = 0.30), or single support time (F2,18 = 0.34, p = 0.71), although step width did differ across mass conditions (F2,18 = 4.15, p = 0.03). Follow-up paired sample t-tests demonstrated a significant difference between the light and medium conditions (−0.005 ± 0.01 vs. 0.002 ± 0.004, t9 = 2.73, p = 0.02), indicating greater lateral step displacement on the prosthetic side for the light condition, but greater lateral step displacement on the intact side for the medium condition. Step width DoA for the heavy condition (−0.0002 ± 0.005) did not differ from the light (t9 = 1.57, p = 0.15) or medium conditions (t9 = 1.40, p = 0.20). Comparisons between experimental conditions were conducted using paired sample t-tests for within-subjects comparisons to assess DoA.

Table 2
Table 2:
Mean (SD) spatiotemporal gait data for the intact and prosthetic legs


Ninety percent of subjects preferred the mass-added conditions. Of the 10 subjects, five preferred the medium-mass condition, four preferred the heavy-mass condition, and only one preferred the light-mass condition (the placebo condition).

Subjects were asked to explain their preferred mass condition. Those who preferred the heaviest mass said, “[The] knee worked better,” it “took pressure off my other leg,” and they had “better control of the leg. When it’s too light, the leg goes all over the place.” One subject preferred the medium mass because “It’s like driving a BMW. The leg just worked better.” Another said that “the leg is more steady...more straight...” The subject who preferred the placebo mass said it “just felt better.”


Stability is “the property of a body that causes it, when disturbed from a condition of equilibrium or steady motion, to develop forces or moments that restore the original condition.”19 The intrinsic and extrinsic risk factors for falling (i.e., age, chronic disease, gait and balance instability, decreased vision, altered mental status, and medication use) suggest that amputees secondary to PVD are vulnerable.20

Repeated-measures ANOVA showed no differences in spatiotemporal gait measures across prosthetic mass conditions. However, there was a statistically significant difference in the DoA in step width, with the heaviest condition being closest to parity. Similarly, studies suggest that increasing the inertial loading of a lightweight prosthesis might increase the economy of prosthetic gait and change kinematics toward a more normal pattern.21 A systematic review looking at the relation between inertial loading and energetics and kinematics of lower-limb prosthetic gait concluded that inertial loading of a lightweight prosthesis was not beneficial to gait, and decreasing the weight should not be the main goal of prosthetic design.21

Using the medium mass, there was a slightly wider step width on the intact side. However, as lower-limb amputees place significantly more of their body mass through the intact side,22 they are much more stable with their center of mass (CoM) over the intact leg.

A wider base of support was created with the lightest mass condition. It is reasonable to assume that this increases the lateral excursion of the CoM. However, CoM was not measured directly. Increased oscillation of CoM is a marker of instability,12 which could increase the risk of falls.

This study was limited in that 8 of 10 participants used a gait aid, and this may have masked the true occurrence of asymmetry. However, gait aid use is common in elderly amputees with PVD. The GAITRite assessment does not permit the calculation of joint kinematics or kinetics. Therefore joint angles, moments, and powers cannot be determined. The GAITRite system also defines some gait parameters using local instead of global coordinate systems and various pressure centers rather than surface markers to define displacements, both of which may result in some discrepancies in data from those of other studies. However, these discrepancies have been shown to be negligible in practice. Furthermore, the GAITRite does not allow us to analyze gait above the floor. For example, we cannot comment on what was happening at the hip (circumduction, hiking) or explain the reasons for a wide-based gait. Gait laboratory analysis, including insights into metabolic cost, joint loads, and ground reaction forces, would be important to consider in future work. Further research could also assess a 2-minute walk test across mass conditions, which has been shown to correlate with both 6- and 12-minute walk tests and is a marker for community ambulation.23 Our small sample size and the short duration of our intervention were also limitations. It is unclear whether prosthetic mass has a different impact over longer walking distances. However, these subjects with amputations secondary to PVD are not walking long distances and are looking for a prosthesis that will help them ambulate within their homes and short outdoor distances (i.e., from a parking lot to a medical appointment). Further research allowing the subjects time at home with the different mass conditions could expand our results to long-distance mobility.

It has been suggested that the location of the added mass may be more important than the amount added.13 In our study, the mass was added at a fixed distance, regardless of the height of the patient, which could have different impacts on the moments of inertia. In addition, the distribution of the varying prosthetic components available to patients may not completely coincide with the location of the added mass in this study.


Short-term increases in prosthetic mass do not adversely affect spatiotemporal gait measures in dysvascular transfemoral amputees. However, the heaviest mass brought gait width closer to parity, which is the ideal outcome. In addition, 90% of subjects preferred an added mass condition over a lighter prosthesis. These results do not support a shift toward a lighter prosthesis in dysvascular transfemoral amputees.

Furthermore, as gait symmetry is a marker for falls, more evidence is needed to confirm that a heavier prosthesis increases stability. Additional research is also needed to determine the optimal mass of a transfemoral prosthesis.


1. Assistive Devices Plan Administrative Manual. Ontario Ministry of Health and Long Term Care. 2012.
2. Otto Bock Orthopedic Industry of Canada General Price List 2012. Otto Bock Orthopedic Industry of Canada. 2012.
3. American Academy of Orthopaedic Surgeons. Atlas of Limb Prosthetics: Surgical and Prosthetic Principles. St. Louis: Mosby-Year Book, Inc.; 1981.
4. Czerniecki JM, Gitter A, Weaver K. Effect of alterations in prosthetic shank mass on the metabolic costs of ambulation in above-knee amputees. Am J Phys Med Rehabil 1994; 73(5): 348–352.
5. Beck JC, Czerniecki JM. A method for optimization of above-knee prosthetic shank-foot inertial characteristics. Gait Posture 1994; 2(2): 75–84.
6. Gitter A, Czerniecki J, Meinders M. Effect of prosthetic mass on swing phase work during above-knee amputee ambulation. Am J Phys Med Rehabil 1997; 76(2): 114–121.
7. Hekmatfard M, Farahmand F, Ebrahimi I. Effects of prosthetic mass distribution on the spatiotemporal characteristics and knee kinematics of transfemoral amputee locomotion. Gait Posture 2013; 37(1): 78–81.
8. Meikle B, Boulias C, Pauley T, Devlin M. Does increased prosthetic weight affect gait speed and patient preference in dysvascular transfemoral amputees? Arch Phys Med Rehabil 2003; 84(11): 1657–1661.
9. Ebskov B, Ebskov L. Major lower limb amputation in diabetic patients: development during 1982 to 1993. Diabetologia 1996; 39(12): 1607–1610.
10. Ebskov L. Lower limb amputations for vascular insufficiency. Int J Rehabil Res 2014; 14(1): 59–64.
11. Moxey PW, Gogalniceanu P, Hinchliffe RJ, et al. Lower extremity amputations—a review of global variability in incidence. Diabet Med 2011; 28(10): 1144–1153.
12. Hausdorff JM, Rios DA, Edelberg HK. Gait variability and fall risk in community-living older adults: a 1-year prospective study. Arch Phys Med Rehabil 2001; 82(8): 1050–1056.
13. Skinner HB, Mote CD. Optimization of amputee prosthetic weight and weight distribution. Rehabil Res Dev Prog Rep 1989; 26: 32–33.
14. Godfrey CM, Brett R, Jousse AT. Foot mass effect on gait in the prosthetic limb. Arch Phys Med Rehabil 1977; 58(6): 268–269.
15. Highsmith MJ, Schulz BW, Hart-Hughes S, et al. Differences in the spatiotemporal parameters of transtibial and transfemoral amputee gait. J Prosthet Orthot 2014; 22(1): 26–30.
16. McDonough AL, Batavia M, Chen FC, et al. The validity and reliability of the GAITRite system’s measurements: a preliminary evaluation. Arch Phys Med Rehabil 2001; 82(3): 419–425.
17. Kuys SS, Brauer SG, Ada L. Test-retest reliability of the GAITRite system in people with stroke undergoing rehabilitation. Disabil Rehabil 2011; 33(19–20): 1848–1853.
18. GAITRite Electronic Walkway. Technical Reference Document Number: WI-02-15. Rev. L 05/06/2013. 2015.
19. Merriam-Webster Online Dictionary 2014. Available at:
20. Miller WC, Speechley M, Deathe B. The prevalence and risk factors of falling and fear of falling among lower extremity amputees. Arch Phys Med Rehabil 2001; 82(8): 1031–1037.
21. Selles RW, Bussmann JB, Wagenaar RC, Stam HJ. Effects of prosthetic mass and mass distribution on kinematics and energetics of prosthetic gait: a systematic review. Arch Phys Med Rehabil 1999; 80(12): 1593–1599.
22. Summers GD, Morrison JD, Cochrane GM. Amputee walking training: a preliminary study of biomechanical measurements of stance and balance. Int Disabil Stud 1988; 10(1): 1–5.
23. Reid L, Thomson P, Besemann M, et al. Going places: Does the two-minute walk test predict the six-minute walk test in lower extremity amputees? J Rehabil Med 2015; 10–1916.

amputees; gait; peripheral vascular diseases; prostheses and implants; mass

© 2015 by the American Academy of Orthotists and Prosthetists.