Translating Technique into Outcomes in Amputation Surgeries

Abstract

The department of surgery at Washington University is putting increased emphasis on outcomes for amputees. This multidisciplinary effort begins with choosing the correct surgery and incorporating the latest technical advances in amputation surgery.

Introduction

There are two million amputees in the United States. Projections estimate there will be 3.6 million amputees by 2050 due to the rise of peripheral vascular disease and diabetes mellitus.^1,2 Despite the 10.6 billion total cost of amputation and postoperative management, surgical amputations often have suboptimal results.³ Wound complications and deep tissue infections are common, often requiring revisionary procedures or costly, morbid re-amputations. Amputees have a hospital readmission rate of 19.5 times per person-year with associated hospital stays of 71.2 days per person-year.³ This significantly impacts amputees’ quality of life, ability to perform ADLs, and profoundly affects longevity. Mortality associated with a major amputation ranges from 15–20% in the first 30 days and approaches 50% over the first five years.^3,4,5 A thoughtful approach to amputations in the context of a larger limb preservation effort can create meaningful differences in long-term quality of life for amputees. At Washington University, incorporation of the latest technical developments in amputation surgery form the initial step in a broader institutional outcomes-oriented approach to amputee care. This article will introduce the modern technical innovations that have been demonstrated to improve clinical outcomes in amputation surgeries.⁶

Transfemoral Amputation

Transfemoral amputations disrupt normal muscle balance between abducting/adducting and flexion/extension muscle groups; leading to 65% greater energy expenditure when ambulating compared to non-amputees. Following transfemoral amputation, residual limb length, orientation of the femur, muscular reattachment, and muscular atrophy impact prosthetic fit and mobility. Transfemoral amputees have 65% greater energy expenditure when ambulating compared to non-amputees. Surgeons aim to restore an energy efficient gait by maintaining the biological, anatomical alignment of the femur.⁸

Before the 1980s, transfemoral amputations sacrificed the adductor magnus and sectioned the hamstring muscles.^7,8,9,10 Resting tension and motor power were lost since the adductor magnus has the greatest moment arm, giving it the best mechanical advantage.⁹ The loss of adduction causes unopposed abduction and a thigh flexion—abduction deformity.^7,8 The anatomical and mechanical axes of the femur become discordant—destabilizing the femur. As a result, the patient lurches sideways with each step. Muscle dis-insertion causes 40–60% of the remaining muscle bulk to atrophy—making prosthetic fit and control difficult.^8,10

To address the energy demands imposed by a lurching gait and address the problems with prosthetic fit, transfemoral amputation evolved to a medial-based flap with preservation of the adductor magnus and myodesis of the adductor to the distal femur.^7,9 Maintaining insertion of the adductor magnus at the linea aspera, will maximize the power and stability of the remaining muscles.⁸ The adductor magnus is ideal for myodesis because it is innervated by both the sciatic and obturator nerves, so it functions in both adduction and extension at the pelvis. Adductor myodesis improves femoral position; soft tissue padding for the distal femur; balanced gait mechanics; gait efficiency, and improved prosthetic fit.^8,10

Medially-based myofasciocutaneous flaps off the obturator artery are preferred because the obturator artery is rarely affected by peripheral vascular disease. The femur is transected proximal to the knee joint; the sartorius, gracilis, and medial hamstrings are transected and included in the muscle anchorage.^8,10 The sciatic nerve is dissected, but is not ligated to decrease neuroma formation.¹⁰ The adductor magnus is myodesed to the lateral cortex of the femur. Adductor myodesis is traditionally performed using Krackow sutures, but recently fiber tape was found to be an easy, quick, and strong alternative.⁷ The quadriceps muscle is reattached to the posterior femur and the hamstrings are anchored to the adductor magnus posteriorly.^8,10

Surgeons aim to create the longest possible residual limb to provide the longest lever arm for transfers, sitting, and increased walking velocity.^9,11,12 Greater length preservation will decrease gait asymmetry and pelvic tilt. Shorter residual limbs result in muscular atrophy, increased trunk excursion, and increased pelvic motion caused by muscular imbalance.^11,12

Osseointegration

Osseointegration was developed in the 1990s in response to persistent challenges related to the socket-limb interface.^{13,14,15,16,17} Traditionally, a prosthetic limb is attached to the patient’s residual limb using a custom designed socket, but 25–33% of patients experience chronic wounds, pinching, perspiration issues, skin irritation, and pain.^13,18 Osseointegration is the direct fixation of an artificial implant into living bone. The implant can withstand normal weight bearing and loading pressures. Offloading the soft tissue surrounding the residual limb, decreases pressure sore and wound formation. Attaching the prosthetic directly into the bone improves osseoperceptive sensory feedback and prosthetic control.^13,16–18

The Osseointegrated Prostheses for the Rehabilitation of Amputees (OPRA) Implant System was developed by Branemark et al. in 1999.¹⁷ The FDA-approved OPRA system includes three main parts: an implanted fixture, an abutment, and abutment screw. Performed in two stages, the first operation implants the fixture intramedullary into the bone stump to provide rotational stability. After a six-month healing period, anterior and posterior muscles are sutured to the periosteum and subcutaneous fat is removed proximal to the skin opening. The final, critical step of the operation is the formation of a tight seal around the distal bone graft.¹³ Thin, hair follicle free, immobile skin should encircle the abutment to limit soft tissue redundancy and provide vascularized soft tissue coverage at the implant bone interface.¹⁷ Recently, a single stage approach for osseointegration has been developed.^13,17

The principle concern regarding osseointegrated implants is the risk of bone or soft tissue infection requiring subsequent reconstruction.^13,18 Superficial infections, requiring oral antibiotics, have a 18–63% incidence,¹⁸ while deep infections are rare. Most authors believe that because the incidence of deep tissue infection is low, it weighs favorably against the common risks of hyperhidrosis, ulcerations, and bullous disease associated with socket prostheses.¹⁸ Furthermore, using titanium-coated implants decreases bacterial biofilm formation, limiting periprosthetic joint infection.^19,20 Creating a well-vascularized, multilayered soft tissue closure and a tight seal around the stoma decreases irritation, inflammation, infection, and, subsequently, closure complications at soft tissue implant interface. Plastic surgery expertise may further decrease soft tissue complications.¹³

Following osseointegration, patients report improved physical functioning, decreased pain, improved prosthetic use, improved prosthetic control, increased walking ability, increased social interactions, increased independence, and improved health related quality of life.^13,18,21,22 Using the OPRA system, prosthetics may be driven by implanted electrodes passed through the femur which creates a bidirectional interface for intuitive prosthetic control via sensory feedback from epimysial and nerve cuff electrodes.^14,17 Washington University has recently been approved as a pilot center for use of the OPRA device in the United States.

Transtibial Amputation

Transtibial amputation has historically been performed using the Burgess technique. Previously well described, the Burgess technique involves cutting both the tibia and fibula, but the fibula is cut at a more proximal level.²³ The proximal tibiofibular joint is undisturbed.²⁴ A long posterior myocutaneous flap is sutured over the end of the residual tibia, followed by myodesis of the gastrocneumius tendon to the proximal tibial periosteum.^23,25,26

Disruption of the interosseous membrane between the tibia and the fibula results in residual limb instability. The fibula is unable to participate in load transfer, so with prolonged weight bearing, the fibula becomes angulated toward the tibia and results in a conical, pointed residual limb that causes soft tissue breakdown.⁸ Janos Ertl developed the osteomyoplastic amputation to address the discordant motions of the tibia and fibula after disarticulation.^8,24–27 Ertl described the creation of an osteoperiosteal tube by elevating tibial periosteum and affixing it to the residual fibula. This osteoperiosteal sleeve was sutured at the ends of the tibia and fibula and filled with bone graft. This sealed the medullary canal and formed a solid bony synostosis between the tibia and fibula which created a stable platform and increased surface area for load transfer.^8,25–27 The myodesis recreates physiological muscle tension and stabilizes the surface area available for prosthetic fitting.²⁷ Sealing the medullary canal restores normal intramedullary pressure to the tibia and improves circulation to the stump.²⁷ The Ertl procedure is also performed for transtibial amputees with residual limb pain due to an unstable fibula.^6,27

Pinto and Harris modified this procedure by creating a vascularized fibular strut with or without the osteoperiosteal sleeve. Their technique utilizes a 4.5 cm fragment of autogenous fibula that maintains its blood supply via connection to adjacent musculature.²⁸ The fibular strut is fixated across the distal aspect of the remaining tibia and fibula. All of their patients developed solid synostosis of the strut which was radiographically and clinically evident by the stable fibula and decreased pain with weight bearing.²⁹

Currently, the most common “Ertl” transtibial amputation involves the use of a long posterior myocutaneous flap. The residual tibia length is 10–15 cm for optimal prosthetic fit. The fibula is divided 4 cm distal to the tibia; after removal of the distal leg, one centimeter of the fibula is rotated medially and fixated between the distal tibia and fibula with Kirschner wire or headless screw fixation.^8,29,30 The fibular strut can be supplemented with a vascularized periosteal sleeve which is created by raising periosteum off of the anterior surface of the distal tibia.^8,27–29 The periosteum is taken with the cortical bone to maintains its vascular supply—creating a vascularized sleeve. Autogenous bone graft may be placed on the bone bridge or within the osteoperiosteal sleeve.^8,27 The osseous surface is covered with myodesed posterior musculature, Achilles tendon, anterior tibial, and peroneal musculature.^8,27,30,31

The Ertl amputation remains controversial. Opponents argue that the risks associated with longer operative and tourniquet times are not outweighed by the supposed benefits. The wound complication rates and physical function scores are equivalent to the standard procedure.^25,31 Inclusion of the bone bridge increases the surgical revision rate, implant removal, screw fixation problems, and neuroma formation.^{26,27,30,31,32} Advocates argue that the increased surface area of the weight bearing region is better able to dissipate force. This results in less pain, improved ambulation, and more prosthetic options.^26,27,33 A higher proportion of U.S. service members who had the Ertl procedure were able to return to active duty.^25–27 However, among vascular patients, only one case series found improved ambulation rates.^27,31 We conclude: younger, more active individuals, and U.S. military personnel who seek to return to active duty may benefit from the osteomyoplastic bone bridge reconstruction while older patients, with medical comorbidities, may not experience additional functional gains and should be screened with caution.

Osseointegration for Transtibial Amputees

Osseointegration to improve prosthetic use, mobility, ambulation, and quality of life among transfemoral amputees has been described.^6,13–22,34 Osseointegration is not widely performed for transtibial amputees, but a recent cohort of nine transtibial amputees who underwent osseointegration reported increased mobility, increased walking ability, and increased prosthetic use.³⁵ Even among patients with peripheral vascular disease, historically a contraindication for osseointegration, all patients were able to walk unassisted one year postoperatively and reported improved quality of life.^34,35

Agonist-Antagonist Myoneural Interface

MIT’s Center for Extreme Bionics developed a new construct aimed at improving myoelectric prosthetic control among amputees: the agonist-antagonist myoneural interface (AMI).^6,36,37,38 Recent studies have shown that the dynamic relationships within agonist-antagonist muscle pairs are fundamental to natural sensations of joint movement.^35,36 In a traditional amputation, residual muscles are arranged isometrically to provide maximum padding for the prosthetic socket. Eliminating the dynamic relationships between agonist-antagonist pairs prevents muscle spindles and golgi tendon organs from communicating proprioceptive information to the CNS.²⁶ The AMI construct consists of an agonist and antagonist muscle tendon surgically connected in series: when the agonist contracts, the antagonist is stretched.^36–38 Thus, mechanoreceptors in both muscles communicate the position, speed, and torque of the joint.^36–38

A tendon harvested from the tarsal tunnel mechanically links the two muscles that compose the AMI. The tarsal tunnels are fixated to the tibia and act as a pulley such that force produced in one muscle causes stretch in the other muscle.^36,38 When provided with a bionic limb consisting of powered ankle and subtalar joints electrically connected to the two AMIs, patients are able to move the bionic limb by contracting the AMI muscles associated with the intended motion.³⁶ The AMI controlling the subtalar joint is composed of the tibialis posterior and the peroneus longus muscles; this AMI is responsible for prosthetic inversion and eversion movements. The AMI controlling the bionic ankle joint is composed of the lateral gastrocnemius muscle and tibialis anterior muscles; this AMI is responsible for prosthetic plantar flexion and dorsiflexion. Since each muscle within the AMI is harvested with its own nerve, the AMI acts as a bidirectional interface between the nerve and bionic prosthetic. Implanted electrodes interface with sensors within the bionic limb to generate a movement command for the motors within the prosthetic. ^6,36

This surgical procedure is now described as the “Ewing Amputation” which incorporates the two AMIs described into the transtibial residuum. Patients who have undergone the Ewing Amputation have improved stability, improved motion path efficiency, improved gait symmetry, natural reflexive behavior during stair ambulation, and improved performance on torque control tasks when compared to traditional transtibial amputees.^36,38 None of the patients experienced muscle atrophy of the residual limb. This may be because the AMI creates nonisometric fixation of muscles which allows for concentric, eccentric and isometric muscular contraction. As a result of increased muscle loading, muscle mass increases.^36,38

Knee Disarticulation

The knee disarticulation (KD) amputation is disproportionately underutilized due to concerns surrounding wound healing complications, issues with prosthetic fit, and concern about a bulbous stump.^39,40 Many surgeons worry that the long flaps necessary to close the incision over the femoral condyles will cause wound healing complications.⁴¹ However, the metabolic cost of walking is greater for proximal amputations due to the additional weight of two prosthetic joints.⁴² Especially among patients with PVD, self-selected walking speed and cadence decrease with proximal amputations, while oxygen consumption increases. Thus, the KD is an energy efficient alternative to transfemoral amputation.^42,43

The end-weight bearing stump produced by the KD allows for direct load transfer which is the physiologic method of weight bearing produced by the 30 bones of the foot.⁹ The weight bearing surface in a KD is the knee joint, which is less stiff and has greater surface area than the osseous stump produced by a transfemoral or transtibial amputation. In a knee disarticulation, the direct transfer prosthetic socket only needs to suspend the prosthesis; in a transosseous amputation the surface area of the bone is small so the bone must be “unweighted” by a socket that distributes the force over the entire surface of the residual bone. Distributing force over a larger surface area is protective against soft tissue breakdown. ^8,9

The surgical technique utilizes a long posterior myofasciocutaneous flap incorporating the gastrocnemius muscle bellies to improve blood supply and padding to the distal stump.^40,42,44 The patellar tendon is removed, and the posterior fat pad is excised. The knee joint capsule and collateral ligaments are divided circumferentially. The cruciate ligaments are released from their attachments; the medial and lateral hamstring tendons are divided; and the popliteal vessels are clamped, divided, and ligated while the perforating vessels remain intact. The tibial and peroneal nerves are drawn distally, cut, and allowed to retract into the soft tissue to prevent neuroma formation.⁴² Muscles that cross the knee joint remain intact and are sutured to the distal end of the stump. The patellar tendon is sutured to the cruciate ligaments and the medial and lateral hamstring tendons are sewn to the knee joint capsule. The iliotibial band is sutured to the lateral aspect of the capsule to improve abduction and hip extension through the gluteus maximus.⁴² For optimal wound healing, blood supply to the gastrocnemius is preserved by protecting the medial and lateral sural branches of the popliteal artery, and perforators between the gastrocnemius and overlying skin.^44,45

Mazet and Hennessy designed a technique to reduce the bulkiness of the KD stump by trimming the medial, lateral and posterior condyles and adding a patellectomy to make a conical distal stump.^41,45 Flaring of the condyles usually results in a bulbous stump which makes donning a prosthetic difficult; by trimming the condyles and discarding the patella, bulbousness is decreased, so shorter tissue flaps are needed to close the incision, improving wound healing.^45,46 Further refined by Burgess, who removed 1.5 cm of the distal end and raised the prosthetic knee joint, cosmetic appearance of the limb is improved because the center of the prosthetic knee joint is brought closer to the level of the contralateral knee.^40,42,43,47

Attinger modified this technique by using a dorsal flap. Instead of discarding the patella, the patella is advanced to cover the end-weight bearing surface of the femur such that the posterior surface of the patella is in direct contact with the distal end of the femur.⁴⁰ The quadriceps tendon is sewn over the cruciate ligament, and the hamstring muscles are myodesed to the quadriceps tendon. Preserving the patella is beneficial for weight bearing during transfers and knee ambulation for patients who are not prosthetic candidates.⁴⁰

Knee disarticulations preserve adductor muscle insertion which improves motor control; maintains the anatomical alignment of the femur; and decreases muscle atrophy.^46,49 Balancing the flexors and extensors within the residual limb decreases the risk of developing hip flexion contractures.^{40,41,43,44,48} Patients use prosthetics earlier, and walk sooner.^46,47 The longer lever arm improves walking stability, prosthesis control, and increases proprioception.^{40,41,43,44,45,49} For non-ambulatory patients, the longer lever arm of knee disarticulation provides easier sitting balance in a wheelchair, trunk stability, mobility in bed, and ease of transfers, while decreasing the risk of developing ulceration and contractures. ^{40,41,44,45,50,51,52}

Using these techniques, patients with PVD have 60–84% primary healing rates, 8–9% revision rate, 0–21% re-amputation rate, and ambulation rates upwards of 75%.^{40,43,44,45,49,50} The conically-shaped stump allows the patient to use a vacuum suction prosthesis that does not need to be removed each time the patient goes to the bathroom. This restores independence to amputees.^{40,41,43,46,47,50} By combining the posterior myocutaneous flap and the Mazet technique up to 81% of patients ambulate.^40,43,44,45

Conclusion

Multiple technical considerations are available to improve long term outcomes for amputees. A thoughtful approach to level of amputation should be taken, incorporating through-joint amputations when feasible to improve direct weight-bearing and independence with transfers. Newer developments such as osseointegration and agonist-antagonist myoneural interface are now being offered at Washington University, and are poised to improve the ease of use and quality of prosthetic ambulation.