Secondary Logo

Journal Logo

Original Research Article

Harness-Supported Versus Conventional Treadmill Training for People with Lower-Limb Amputation

A Preliminary Report

Lamberg, Eric M. EdD, PT; Muratori, Lisa M. EdD, PT; Streb, Robert PhD, PT; Werner, Marc CPO; Penna, James MD

Author Information
JPO Journal of Prosthetics and Orthotics: April 2014 - Volume 26 - Issue 2 - p 93-98
doi: 10.1097/JPO.0000000000000025
  • Free


The number of people living with a major lower-limb amputation (LLA) is expected to double by the year 2050 as the incidence of dysvascular disease increases with an aging population.1 A main focus of rehabilitation will be to ensure that these individuals retain the functional abilities needed to optimize active participation in the community. Typical rehabilitation after LLA focuses on gait and functional mobility retraining, so when discharged, most have regained the capacity to walk functional distances.2 However, walking and balance abilities of community-dwelling individuals after LLA are typically less than optimal compared with age-matched peers, contributing to a higher rate of falls, decreased functional independence, and limited participation outside the home.3 Individuals with LLA walk more slowly, have decreased balance, demonstrate gait disruptions that increase the physiological energy requirements of walking and contribute to long-term musculoskeletal impairments, and have decreased endurance.4–8 However, after initial inpatient rehabilitation after LLA, most do not seek further rehabilitation services despite the continued need.9,10 Currently, there is no evidence addressing whether continued gait training for community-dwelling individuals living with LLA greater than 1 year after amputation can lead to functional improvements.

Repetitive step training on a treadmill with a percentage of body weight supported with a harness (body weight–supported treadmill training [BWSTT]) is a gait training strategy that unloads the lower limbs. This type of training allows for the continued practice of walking in a safe environment. The harness minimizes postural control requirements, making it easier to walk by allowing the development of a more efficient, symmetrical limb strategy. After central nervous system damage, repetitive practice of this new locomotor strategy ultimately leads to experience-dependent plasticity.11 Studies in stroke,12,13 Parkinson disease,14 and traumatic brain injury15 have demonstrated that improvements found when using treadmill training carried over to improved overground gait and endurance abilities. Interestingly, similar positive results from preintervention to postintervention have been shown after orthopedic injury.16

Only one study has been published that examined the use of body weight harness support for individuals with LLA.17 This study, which focused on energy expenditure, found decreased energy costs when walking with a harness, providing partial body unweighing. To date, there are no reports citing the clinical gait effects of providing body weight support for the individual with LLA. For individuals after LLA, use of a supportive harness relieves a percentage of the individual’s body weight, allowing for repeated walking practice with less energy requirement. The goal of this study was to determine whether use of a harness for support during treadmill training would lead to changes in gait symmetry and endurance compared with treadmill training with a harness available for safety but without providing any support. We hypothesized that the supportive harness condition would provide both a safe environment and reduced postural demands, allowing this group of participants to achieve better gait symmetry and walking endurance compared with the group using the harness for safety only.



Participants with LLA between the ages of 21 and 70 years were recruited through flyers, online postings, and advertisements in local newspapers. Inclusion criteria were unilateral transtibial, transfemoral, or knee disarticulation amputation more than 1 year before; ability to walk using prosthesis; comfortably fit with a prosthesis for at least 6 months; ability to tolerate a moderate-intensity exercise program; and currently not receiving physical therapy for gait training. Specific exclusion criteria included cardiac or pulmonary disease that limits ability to exercise, discomfort that restricts ability to walk, and presence of active wounds on the residual limb or contralateral foot. The experimental procedure was approved by the institutional review board of Stony Brook University, and all participants signed consent forms before initiating the protocol. This trial was registered as “Treadmill Training With Lower Extremity Amputees” with registry no. NCT01419288.


The participants were assigned to either treadmill training with harness support (SUP) or treadmill training without harness support (NoSUP) groups. A Biodex Medical Systems (Shirley, New York) Gait Trainer treadmill and Offset Unweighing system including harness were used. The adjustable harness was secured around the participant’s torso and then to the overhead superstructure. The deck of the treadmill is instrumented to provide temporal-spatial information regarding walking.

All participants completed 12 training sessions (three times a week for 4 weeks) requiring 30 minutes of walking on the treadmill. Rests were provided if needed but not included in the total time spent engaged in walking. Both groups wore the harness throughout training and were permitted to walk using their natural gait pattern. At the beginning and end of each training session, the skin of the residual limb and the contralateral foot was assessed for evidence of breakdown. To determine the initial treadmill walking speed for training, all participants walked at a comfortable pace on the treadmill with the harness secured for safety but without body weight supported. The SUP group started training with 30% of body weight supported. Previous research has determined that if greater than 40% of body weight is supported, the mechanics of gait change.18 The amount of support was systematically reduced by 5% increments, with a goal of no body weight support (full weightbearing) by training session 10. The NoSUP group wore the harness, without any body weight supported, to mimic the environment that the SUP group trained under and for safety. Treadmill speed was increased for both groups in 0.1-mph increments as determined by a member of the research team. Vital signs were monitored throughout training.


The participants were assessed before training (baseline) and at 1 and 4 weeks after training. Assessment sessions were similar. The 6-minute walk test (6MWT) was completed to assess overground functional mobility and walking capacity.19,20 The 6MWT requires participants to walk as far as they could in 6 minutes. Functional mobility was assessed using the Timed Up and Go (TUG) test.21 The TUG requires participants to stand from a chair, walk 3 m, turn, walk back to the chair, and sit. Performance is measured in seconds. In this study, the participants performed a practice trial, and then two trials were completed, with the mean data used for analysis. The self-report Activities-Specific Balance Confidence Scale (ABC), which measures perceived ability to maintain balance during various functional activities, was completed.7,22 A physical examination was performed including measurement of height, weight, leg length, residual limb characteristics, and skin condition of the residual limb and the contralateral foot. To explore changes that may have occurred during training, data from the instrumented treadmill were analyzed from training sessions 1, 6, and 12. This included treadmill speed, the coefficient of variation (CV) of prosthetic-side step length, CV of sound-side step length, and a composite score termed ambulation index (AI). The AI ranges from 0 to 100 (the goal is 100) and is calculated using mean step cycle and time spent on each leg during the course of a training session.23


Mixed-model repeated-measures analysis of variance (ANOVA) was used to test the effects of treatment between the groups over time (baseline, 1 week, and 4 weeks) for the 6MWT, the TUG, and the ABC. Additional mixed-model repeated-measures ANOVA was used to identify changes in temporal-spatial variables throughout training (visits 1, 6, and 12). Tukey’s post hoc comparisons were used when appropriate, and treatment effect size using Cohen d statistic and 95% confidence intervals (CIs) were calculated. At 4 weeks, one subject reported a headache and did not complete the 6MWT. Because of the intention-to-treat analysis set a priori, we replaced the single point of missing data by carrying the previous value from 1 week forward, providing a conservative estimate of the true data. A significance level of p < 0.05 was used.


Twenty-three individuals inquired about the study and were screened for eligibility. Thirteen were excluded as follows: seven were unable to commit to the time requirements and six did not meet inclusion requirements. Written informed consent was obtained from 10 participants. Of the 10, one chose not to return after baseline and one developed unrelated medical complications during training and was excluded from analysis. A total of eight individuals (seven men) with unilateral LLA due to trauma completed this study; four were assigned to the SUP and four were assigned to the NoSUP. Table 1 shows participant demographics.

Table 1:
Participant characteristics

As seen in Figures 1 and 2 (data collapsed across NoSUP and SUP), significant main effects were found, demonstrating that, for all participants, the distance walked during the 6MWT (p = 0.002) and the time to complete the TUG (p < 0.001) improved. However, the mode of treadmill-based locomotor training (SUP vs. NoSUP) did not result in significant changes for either the 6MWT (p = 0.93) or the TUG (p = 0.99). Post hoc analysis revealed that, overall, the distance walked during the 6MWT significantly increased by 25% when comparing 1 week with baseline (mean change, 89.6 m; 95% CI, 50.2–129.0; d = 1.5) and by 32% comparing 4 weeks with baseline (mean change, 112.4 m; 95% CI, 70.0–154.7; d = 1.7). Post hoc analysis revealed that the TUG improved by 13%, a significant change when comparing 1 week with baseline (mean change, −1.2 seconds; 95% CI, −0.4 to −2.0; d = 1.2) and 4 weeks with baseline (mean change, −1.2 seconds; 95% CI, −0.3 to −2.1; d = 0.9). Table 2 reports the group (SUP or NoSUP) data. There was no change in ABC scores between the groups (p = 0.35) or during the course of the study (p = 0.74).

Table 2:
Group data (mean [SD]) at the three assessment sessions
Figure 1:
Mean and SD of the distance walked during the 6-minute walk test before training (baseline), 1 week after training, and 4 weeks after training collapsed across the two training groups. Asterisk denotes significantly different (p < 0.05) than baseline.
Figure 2:
Mean and SD of the time to complete the TUG before training (baseline), 1 week after training, and 4 weeks after training collapsed across the two training groups. Asterisk denotes significantly different (p < 0.05) than baseline.


The mode of treadmill-based locomotor training (SUP vs. NoSUP) did not produce significant differences during training for the composite AI (p = 0.93), treadmill speed (p = 0.28), CV of prosthetic-limb step length (p = 0.80), or CV of sound-limb step length (p = 0.92), However, significant main effects were found over time indicating that, regardless of group assignment, improvements were noted in AI (p = 0.004) and treadmill speed (p < 0.001), with simultaneous reductions in CV of prosthetic- (p = 0.015) and sound-limb (p = 0.004) step length. Post hoc analysis revealed that most of these improvements were significant when comparing training visit 1 with visit 6 and visit 1 with visit 12 (see Table 3).

Table 3:
Group data (mean [SD]) at training visits


This preliminary study showed that participation in an intensive 4-week treadmill walking program improved walking capacity, walking speed, step length consistency, and functional mobility in a group of individuals who were living with an LLA, on average, for more than 25 years. Further, these improvements were maintained after a period of 4 weeks with no training. Contrary to our initial hypothesis, the results demonstrated that, in this sample, there was no added benefit of supporting body weight during treadmill gait training.

Using BWSTT as an intervention to improve functional ambulation has traditionally been used and found to be effective in neurologic rehabilitation.11,16 By supporting a percentage of body weight and reducing postural demands, individuals are able to devote increased time on the task (more time walking). The underlying neurologic theory suggests that, through massed stepping practice, a peripheral stimulus is provided that can bypass a damaged central nervous system and excite spinal cord circuitry to promote rhythmic alternating movements of the lower limbs.

In this study, participation was dependent on the presence of a peripheral (LLA), rather than central, injury. Thus, theoretical benefits of BWSTT were not applicable. Instead, the protocol tested the potential for body weight support to improve walking because less body weight needed to be transferred through the legs, decreasing effort and increasing time spent on the task.

On the basis of results from studies conducted with other patient populations including those with orthopedic injuries, we hypothesized that participants who were provided with support would perform better than those with the harness for safety alone. Although we found improvements for both groups, there were no differences between the groups, indicating no added benefit from partially supporting body weight during walking. It is feasible that because all participants in this study were already considered independent community ambulators, we had a ceiling effect; alleviating some body weight did not have a great effect in reducing postural demands because these individuals were already skilled. Results such as these have been found previously. The addition of body weight support was not superior to treadmill training alone for people with stroke who were independent in walking at the start of treatment.24

As indicated by the results, all participants improved their 6-minute walk distance when compared with baseline. At initial testing, the participants in this study walked, on average, 356 m. This distance is similar to the distance completed in a study assessing reliability of the 6MWT during 2 days of testing for individuals with LLA.10 In that study, 44 participants with mixed LLA level and a mean age of 66 years walked, on average, 338 m.10 Further, investigators were able to determine a minimal detectable change (MDC) score for the 6MWT using a 90% confidence level of 45 m.10 The MDC is the minimum amount of change needed to be 90% confident that the change exceeds typical measurement error. In this study, we found that the change in distance walked during the 6MWT exceeded the sugested MDC when comparing 1 week with baseline (90-m change) and 4 weeks with baseline (112-m change).

Although it is important to establish that the results meet or exceed the MDC previously determined,10 it is also important to know whether the improvements we found are meaningful to the individual. One way to evaluate meaningfulness is to compare the improvement with established minimal clinically important difference (MCID) values. Currently, no study has identified the MCID for the 6MWT in the LLA population. In a study of community-living older adults, the MCID for the 6MWT was determined to be 50 m.25 In a different study, it was determined that, for people with chronic lung disease, the MCID ranged from 54 to 80 m.26 The improvements we found at both follow-up visits when compared with the baseline exceeded the upper range of 80 m, which is indicated above. Thus, we believe that, although the self-report perception of changes in balance as captured through the ABC did not show change, the improvements gained on the 6MWT, as a result of participating in this study, have the potential to make a clinically important difference to the participants. Qualitatively, as reported by one participant who was now able to travel, “I would not have done this before this study”; “[I] Feel like I got my life back.”

The 6MWT is a measure of mobility and endurance rather than balance and therefore may not be well reflected in self-report measures of balance. The improvements in the 6MWT distance are likely the result of the increased time spent walking on the treadmill. In a recent study conducted with people with a transtibial LLA that included a range of 1 to 4 hrs of prosthetic gait training focused on reducing existing gait deviations and maximizing performance using various prosthetic feet, the distance the participants walked in 6 minutes did not change.27 Here, we found that when individuals received 6 hrs of focused walking (three times a week for 4 weeks), without instructions to minimize gait deviations, the distance walked in 6 minutes did significantly improve, but we are unaware whether gait deviations overground were altered. Further, it is unclear whether participation in overground walking training at the same intensity would produce similar results or if the safety of the harness and the peripheral stimulus of a moving treadmill is required. In addition, further testing is needed to investigate variables such as duration and intensity of training as well as to determine whether treadmill training with support is beneficial in the LLA population early in the rehabilitation process.

The results found using the data from the instrumented treadmill deck that were collected during the training sessions indicate that the participants improved their walking speed (treadmill walking speed increased on average by 71% when comparing training visit 1 with visit 12). According to the Manual on Uniform Traffic Control Devices for Streets and Highways,28 the minimum interval used at pedestrian crossways to allow safe passage across a street should be 3.5 ft per second (fps). When the participants in this study first walked on the treadmill, they were ambulating at a speed equal to 2.05 fps (1.4 mph). At this speed, it would be difficult to safely navigate across traffic. However, by training visit 12, the participants in this study were able to ambulate at 2.4 mph, which is equivalent to 3.52 fps, allowing these individuals to walk at a minimum safe walking speed.28

Changes in intrasubject variability are one way to document improvement in skill performance.29 In this study, the participants not only increased walking speed but also were able to do so with greater stride-to-stride consistency. In gait, a reduction in stride variability may have real-world consequences. In older adults, increased stride variability is associated with unsteadiness and increased fall risk.30,31 Although no studies have been completed to investigate this association in the LLA population, it is plausible that stride variability may increase risk for fall for these individuals as well.


This study had a small number of participants with various LLA levels and thus makes generalizability to all persons living with LLA challenging. This study, by design, looked at only those who have been living with an LLA for greater than 1 year. Further, it is not known as to the effect of relieving body weight during treadmill ambulation may have when providing rehabilitation services earlier during the rehabilitation process.


This pilot study compared two modes of treadmill-based locomotor training in individuals living with an LLA. Participation in an intensive treadmill locomotor training program that maximizes time spent on the task of walking is an effective means of improving function many years after LLA regardless of the mode of training. Despite the general impression that rehabilitation provides the best results when performed early, less is known about the effect of intervention further away from the initial injury. Providing access to walking programs for individuals with LLA may allow for improvements in ambulation that provide the individual with more opportunities to participate in the community.


The authors thank Raymond McKenna for providing valuable assistance with data analysis and the staff of the GCRC at Stony Brook University for all their assistance.


1. Ziegler-Graham K, MacKenzie EJ, Ephraim PL, et al. Estimating the prevalence of limb loss in the United States: 2005 to 2050. Arch Phys Med Rehabil 2008; 89: 422–429.
2. Munin MC, Espejo-De Guzman MC, Boninger ML, et al. Predictive factors for successful early prosthetic ambulation among lower-limb amputees. J Rehabil Res Dev 2001; 38: 379–384.
3. 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: 1031–1037.
4. Gailey R, Allen K, Castles J, et al. Review of secondary physical conditions associated with lower-limb amputation and long-term prosthesis use. J Rehabil Res Dev 2008; 45: 15–29.
5. Genin JJ, Bastien GJ, Franck B, et al. Effect of speed on the energy cost of walking in unilateral traumatic lower limb amputees. Eur J Appl Physiol 2008; 103: 655–663.
6. Houdijk H, Pollmann E, Groenewold M, et al. The energy cost for the step-to-step transition in amputee walking. Gait Posture 2009; 30: 35–40.
7. Miller WC, Speechley M, Deathe AB. Balance confidence among people with lower-limb amputations. Phys Ther 2002; 82: 856–865.
8. Gard SA. Use of quantitative gait analysis for the evaluation of prosthetic walking performance. J Prosthet Orthot 2006; 18: 93.
9. Pezzin LE, Dillingham TR, Mackenzie EJ, et al. Use and satisfaction with prosthetic limb devices and related services. Arch Phys Med Rehabil 2004; 85: 723–729.
10. Resnik L, Borgia M. Reliability of outcome measures for people with lower-limb amputations: distinguishing true change from statistical error. Phys Ther 2011; 91: 555–565.
11. Finch L, Barbeau H, Arsenault B. Influence of body weight support on normal human gait: development of a gait retraining strategy. Phys Ther 1991; 71: 842–855; discussion 855–856.
12. Barbeau H, Visintin M. Optimal outcomes obtained with body-weight support combined with treadmill training in stroke subjects. Arch Phys Med Rehabil 2003; 84: 1458–1465.
13. Duncan PW, Sullivan KJ, Behrman AL, et al. Body-weight–supported treadmill rehabilitation after stroke. N Engl J Med 2011; 364: 2026–2036.
14. Miyai I, Fujimoto Y, Yamamoto H, et al. Long-term effect of body weight-supported treadmill training in Parkinson’s disease: a randomized controlled trial. Arch Phys Med Rehabil 2002; 83: 1370–1373.
15. Brown TH, Mount J, Rouland BL, et al. Body weight–supported treadmill training versus conventional gait training for people with chronic traumatic brain injury. J Head Trauma Rehabil 2005; 20: 402–415.
16. Hesse S, Werner C, Seibel H, et al. Treadmill training with partial body-weight support after total hip arthroplasty: a randomized controlled trial. Arch Phys Med Rehabil 2003; 84: 1767–1773.
17. Hunter D, Smith Cole E, Murray JM, Murray TD. Energy expenditure of below-knee amputees during harness-supported treadmill ambulation. J Orthop Sport Phys Ther 1995; 21: 268–276.
18. Visintin M, Barbeau H. The effects of body weight support on the locomotor pattern of spastic paretic patients. Can J Neurol Sci 1989; 16: 315–325.
19. Directors ATSBo. ATS statement: guidelines for the six-minute walk test. Am J Resp Crit Care Med 2002; 166: 111–117.
20. Lin SJ, Bose NH. Six-minute walk test in persons with transtibial amputation. Arch Phys Med Rehabil 2008; 89: 2354–2359.
21. Schoppen T, Boonstra A, Groothoff JW, et al. The Timed “up and go” test: reliability and validity in persons with unilateral lower limb amputation. Arch Phys Med Rehabil 1999; 80: 825–828.
22. Miller WC, Deathe AB, Speechley M. Psychometric properties of the Activities-Specific Balance Confidence Scale among individuals with a lower-limb amputation. Arch Phys Med Rehabil 2003; 84: 656–661.
23. Biodex Medical Systems, Inc. Gait Trainer 3. Application/Operation Manual. Shirley, NY; 2011: 7–12.
24. Visintin M, Barbeau H, Korner-Bitensky N, Mayo NE. A new approach to retrain gait in stroke patients through body weight support and treadmill stimulation. Stroke 1998; 29: 1122–1128.
25. Perera S, Mody SH, Woodman RC, Studenski SA. Meaningful change and responsiveness in common physical performance measures in older adults. J Am Geriatr Soc 2006; 54: 743–749.
26. Wise RA, Brown CD. Minimal clinically important differences in the six-minute walk test and the incremental shuttle walking test. COPD 2005; 2: 125–129.
27. Gailey RS, Gaunaurd I, Agrawal V, et al. Application of self-report and performance-based outcome measures to determine functional differences between four categories of prosthetic feet. J Rehabil Res Dev 2012; 49: 597–612.
28. Federal Highway Administration. Manual on Uniform Traffic Control Devices. 2009 Ed. Washington, DC; 2012.
29. Magill RA. Motor Learning and Control: Concepts and Applications. 9th Ed. New York, McGraw Hill; 2011.
30. Hausdorff JM, Edelberg HK, Mitchell SL, et al. Increased gait unsteadiness in community-dwelling elderly fallers. Arch Phys Med Rehabil 1997; 78: 278–283.
31. 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: 1050–1056.

amputee; functional outcomes; gait training; physical therapy; rehabilitation; transfemoral; walking

© 2014 by the American Academy of Orthotists and Prosthetists.