Abstract
Quadruple amputation, defined as the loss of all four limbs, is an exceptionally rare condition. Conventional gait training is particularly challenging in this population because the absence of upper limbs limits the ability to grasp assistive devices that are essential for maintaining balance and stability during walking. End-effector robot-assisted gait training offers an alternative rehabilitation strategy that enables upright mobility and task-specific gait training for patients who experience substantial difficulty performing conventional gait training. This approach provides safe, repetitive, and hands-free gait training for individuals with quadruple amputation, resulting in measurable improvements in balance, ground reaction force, and functional mobility. We report two cases of successful amputee gait rehabilitation using an end-effector–type gait robot in two females (aged 72 and 51 years) with quadruple amputation.
-
Keywords: Amputees, Gait, Robotics
INTRODUCTION
Quadruple amputation, the loss of all four limbs, is an exceptionally rare condition [
1]. Patients with bilateral lower-limb amputations usually require parallel bars or crutches during early gait training [
2]. However, individuals with quadruple amputation cannot use these aids because of upper-limb loss, limiting their ability to engage in conventional gait training that relies on upper-limb support for balance and stability. As a result, their rehabilitation progresses more slowly, and their expected final K-level is typically lower [
1].
Robot-assisted gait training (RAGT) may offer a novel alternative. RAGT has been extensively studied for its efficacy in central nervous system disorders such as stroke and spinal cord injury. Two main types of RAGT devices are available: exoskeletal and end-effector systems. Among these, end-effector robots attach primarily to the patient’s feet, allowing greater freedom of movement at the knee and hip joints and enabling so-called destabilization training. This approach facilitates neural circuit reinforcement, enhances postural and balance control, and improves sensory integration [
3,
4]. Its ground-walk mode simulates level-ground ambulation, whereas the stair-ascending mode mimics the increased hip and knee flexion angles required during stair climbing. This system has been validated to generate lower-limb muscle activation patterns similar to those of real walking and stair ascent [
5].
Although RAGT has been predominantly investigated in neurological populations, its application in amputee rehabilitation—particularly in cases of multiple limb loss—has been rarely documented. We hypothesized that end-effector–based RAGT could be especially beneficial for quadruple amputees who are unable to grasp assistive devices. Here, we report two cases of quadruple amputees who underwent end-effector RAGT. This retrospective case review study was approved for exemption by the Institutional Review Board of Ilsan Paik Hospital (IRB No. ISPAIK 2025-09-029). A written informed consent was obtained from the patients for the publication of this case report.
CASE REPORT
Case 1
A 72-year-old female was admitted to the Department of Infectious Diseases on January 1, 2022, with septic shock secondary to a urinary tract infection. After inotropic treatment for hemodynamic support, she developed necrosis of both hands and feet. She underwent right-hand amputation at the metacarpal shaft and left-hand amputation at the metacarpophalangeal joint on September 22, 2022, followed by bilateral transtibial amputations on October 26, 2022.
On December 21, 2022, she was fitted with transtibial prostheses using a patellar tendon-bearing (PTB) socket, shuttle lock suspension, and energy-storing feet (Vari-Flex EVO®, Ottobock) and was admitted for inpatient rehabilitation.
At rehabilitation onset, upper-extremity wounds were unhealed, preventing fabrication of upper-limb prostheses. Thus, she could not grasp parallel bars or a cane, making conventional gait training impractical. RAGT with the Morning Walk
® system (Curexo) was initiated on December 28, 2022, in ground-walk mode (
Fig. 1A). Considering her age and multiple limb loss, the initial rehabilitation goal was K-level 1.
During the first RAGT session, cadence was 25 steps/min with 8% body-weight support. Cadence was gradually increased to 45 steps/min while support was reduced to 0%. Ground reaction force (GRF) data were obtained from the footplate sensors integrated in the Morning Walk
® system. The device records three-dimensional GRF (vertical, anterior–posterior, and medio-lateral components) through 6-axis load cells; however, only the vertical component is provided in the clinical report. The GRF value used in this study represents the average of the peak vertical GRFs measured during the stance phase across all steps in each training session, expressed as a percentage of body weight. Mean GRF improved markedly (
Fig. 2A). Training then advanced to stair-ascending mode at 40 steps/min. After two months, cadence reached 50 steps/min with GRFs of 96% (right) and 106% (left).
Before RAGT, she could stand only a few seconds. After two weeks, she ambulated 80 m using an elbow-support walker, and after five months achieved supervised ambulation (
Fig. 3A). Her goal was upgraded from K-level 1 to 3. After complete wound healing, silicone cosmetic prostheses for both upper limbs were fabricated on April 12, 2023. She now ambulates independently on level ground and ascends and descends stairs while holding rails using both upper-limb prostheses.
A 51-year-old female developed necrosis of both hands and feet following inotropic treatment for hypoglycemic shock in 2021. Initially managed conservatively, the necrosis progressed. Upon presentation to our orthopedic department on June 19, 2023, she exhibited autoamputation with necrosis of the right ankle and wrist, open autoamputation of the left subtalar joint, and autoamputation of the 2nd–5th proximal phalanges of the left hand. On June 26, 2023, she underwent bilateral transtibial and right transradial amputations.
She was prescribed transtibial prostheses with PTB socket, shuttle lock suspension, and energy-storing feet (Vari-Flex EVO®). A Muenster-type cosmetic hand was fabricated for the right transradial amputation. She was fitted with both upper- and lower-limb prostheses on August 31, 2023, and began inpatient rehabilitation with an initial goal of K-level 1.
RAGT in ground-walk mode was initiated on September 1, 2023 (
Fig. 1B). During the first session, the cadence was set at 30 steps/min with 0% body-weight support. These parameters were later increased to a cadence of 35 steps/min and a step length of 40 cm. However, as stump volume reduction progressed and the wearing time of the PTB socket increased, she reported worsening bilateral knee pain attributable to underlying osteoarthritis. Hyaluronic acid injections were administered to the left knee on September 11 and to the right knee on October 10.
Due to her knee pain, she was unable to press the footplates firmly during RAGT, resulting in decreased GRF values. After the 12th session, there was no significant increase in GRF; however, as the patient no longer reported knee pain, training was continued with an increased cadence (
Fig. 2B). The cadence was intentionally kept low during the initial sessions to reduce mechanical stress on the knees while the patient experienced osteoarthritic pain. After the pain subsided following intra-articular hyaluronic acid injections and improved prosthetic adaptation, the cadence was increased to 50 steps/min to correspond with the patient’s enhanced walking capability and tolerance. After she began walking on level ground under supervision, RAGT was progressed to the stair-ascending mode.
Before initiation of RAGT, the patient could stand independently but required minimal assistance for ambulation. After two weeks of training, she progressed to ambulation with a Lofstrand crutch, and after three weeks, she was able to walk under supervision (
Fig. 3B). Her functional recovery exceeded initial expectations, and her K-level goal was subsequently revised from Level 1 to Level 3. At present, she can ambulate independently with variable cadence and ascend and descend stairs without support.
DISCUSSION
Conventional rehabilitation for lower-limb amputees typically includes early strengthening, standing and balance training with parallel bars, and progressive gait re-education with prostheses [
2]. This process usually begins with manual assistance, followed by ambulation with a walker and later a cane, before advancing to stair climbing and uneven-surface training. Rehabilitation for multiple amputees, however, is considerably more complex due to the simultaneous loss of both upper and lower extremities, which prevents the use of assistive devices and necessitates modification of standard rehabilitation strategies. Only a few reports have described successful rehabilitation in such cases. Shin et al. [
6] reported a 13-week stepwise rehabilitation program for a triple amputee with right knee disarticulation, left transfemoral amputation, and right elbow disarticulation. Kitowski and Leavitt [
7] described a 5-month multidisciplinary rehabilitation program focusing on systematic gait training in a quadruple amputee with bilateral transradial and transfemoral amputations. Recently, Lee et al. [
8] reported the effect of end-effector–type RAGT in a patient with bilateral transtibial amputation and cerebellar ataxia, who demonstrated improved balance, walking capacity, and activities of daily living after a three-week intervention. To date, however, there have been no reports describing the application of RAGT in patients with quadruple amputations.
In the present cases, we implemented end-effector–type RAGT. In contrast to exoskeletal robots, which attach directly to the patient’s hip, knee, and ankle joints and therefore require a full range of motion in these joints, end-effector robots generate the gait cycle through programmed trajectories of footplates to which only the patient’s feet are attached. Because transtibial prostheses fabricated with PTB socket typically incorporate approximately 5° of anterior tilt, achieving proper alignment within an exoskeletal system can be challenging. Consequently, the end-effector type is more suitable for patients using transtibial prostheses.
Our two cases suggest that end-effector–based RAGT can improve balance and gait function while reducing the overall rehabilitation period in patients with multiple limb loss. In Case 1, the patient achieved ambulation with an elbow-support walker after two weeks of RAGT and independent walking after five months. In Case 2, functional walking was restored after only three weeks of RAGT. These outcomes indicate that RAGT may accelerate gait recovery and shorten rehabilitation duration even in quadruple amputees who cannot use conventional assistive devices because of upper-limb loss.
In conclusion, our cases demonstrate that end-effector–based RAGT offers a feasible and effective alternative for patients unable to perform conventional gait training due to upper-limb loss. This approach may facilitate earlier mobility, reduce therapist burden, and broaden rehabilitation possibilities for individuals with severe multiple limb loss.
Nevertheless, this report has several limitations. Because it includes only two patients with distinct etiologies and functional baselines, the generalizability of the findings is limited. The results should therefore be regarded as exploratory, providing preliminary evidence that may guide future pilot and multicenter studies investigating end-effector RAGT in amputee rehabilitation.
Fig. 1.Robot-assisted gait training in ground-walk mode in Case 1 (A) and Case 2 (B).
Fig. 2.Changes in ground reaction force (GRF) and cadence during robot-assisted gait training (RAGT) in Case 1 (A) and Case 2 (B).
Fig. 3.Ambulation on level ground under supervision in (A) Case 1 and (B) Case 2.
REFERENCES
- 1. Athani B, Pillai IN, Prasad J D, Shaji A. Rehabilitation of a quadruple amputee with leukemia: a challenging endeavor. Prosthet Orthot Int 2025;49:559-64.
- 2. Madou E, Sureshkumar A, Payne MW, Viana R, Hunter SW. The effect of exercise interventions on gait outcomes in subacute and chronic rehabilitation from lower-limb amputation: a systematic review and meta-analysis. Prosthet Orthot Int 2024;48:128-48.
- 3. Shin JC, Jeon HR, Kim D, Cho SI, Min WK, Lee JS, et al. Effects on the motor function, proprioception, balance, and gait ability of the end-effector robot-assisted gait training for spinal cord injury patients. Brain Sci 2021;11:1281.
- 4. Shin JC, Jeon HR, Kim D, Min WK, Lee JS, Cho SI, et al. Effects of end-effector robot-assisted gait training on gait ability, muscle strength, and balance in patients with spinal cord injury. NeuroRehabilitation 2023;53:335-46.
- 5. Ahn Y, Hong J, Shim D, Choi JO, Rha DW. Comparing the lower-limb muscle activation patterns of simulated walking using an end-effector-type robot with real level and stair walking in children with spastic bilateral cerebral palsy. Sensors (Basel) 2023;23:6579.
- 6. Shin JC, Yang HE, Yu SJ, Kim NY, Yoon SY. Stepwise rehabilitation of the triple amputee combined with dysfunction of the sound limb. Ann Rehabil Med 2014;38:116-21.
- 7. Kitowski VJ, Leavitt LA. Rehabilitation of a quadruple amputee. South Med J 1968;61:912-4.
- 8. Lee SJ, Song C, Lee JS, Jeon HR. Robot-assisted gait training using an end-effector type robot in a patient with bilateral transtibial amputation and cerebellar ataxia. PM R 2024;16:300-2.
, Yunji Oh, MD1
