This chapter describes surgery and prosthetic options for lower limb amputations in foot and ankle practice, covering immediate traumatic, planned essential (e.g., diabetic necrotic ulcers), and elective amputations (e.g., painful fused ankle), but not the management of congenital abnormalities or stump revision surgery.
Amputation is one of the oldest procedures in surgery. Evidence of digital amputations, perhaps performed for religious purposes, is present in cave art from 36 000 years ago found in modern-day France and New Mexico1. The first medical text on amputation, found in the Hippocratic text On Joints, however, is much more recent, dating from the latter half of the fifth century BC. The Hippocratic author describes the practice of amputating an ischemic limb below the “boundaries of blackening” as a measure of last resort2. The concept of amputating more proximally, through healthy tissue, did not appear for another four hundred years when Aulus Cornelius Celsus, a Roman encyclopedist who may or may not have practiced medicine, published his De Medicina in around 50 BC. In many ways, the principles he outlines hold true to this day:
between the sound and the diseased part, the flesh is cut through with a scalpel down to the bone, but this must not be done actually over a joint, and it is better that some of the sound part should be cut away than that any of the diseased part be left behind. When the bone is reached, the sound flesh is drawn back from the bone and undercut from around it, so that in that part also some bone is bared; bone is then to be cut through with a small saw as near as possible to the sound flesh which still adheres to it, next the face of the bone, which the saw has roughened, is smoothed down, and the skin drawn over it; this must be sufficiently loosened in an operation of this sort to cover the bone all over as completely as possible3.
Celsus was familiar with the use of ligatures on blood vessels, and may have used them during amputations but, interestingly, does not describe them in this context. Barring minor refinements, the technique of amputation remained largely unchanged for nearly 1800 years, until Jean-Louis Petit, who also invented the tourniquet, advocated dividing the skin and muscle at one level and the bone more proximally, the “two-stage circular cut,” in 17184. The use of a soft tissue flap for skin closure, meanwhile, had been described by James Yonge in 1679, a naval surgeon, although he attributed the innovation to “a very ingenious surgical brother, Mr. C. Lowdham of Exeter”5.
The history of prosthetics is similarly ancient, with the first literary account, that of the warrior-queen Vishpla losing a leg in battle and having it replaced with an iron limb, being found in the Sanskrit Rig Veda of around 1800 BC6. Simple prostheses were used in the classical world, but during the renaissance some prostheses of quite remarkable sophistication, such as those designed by the surgeon Ambroise Paré, were made. At this time prostheses were individually bespoke, and the emphasis was upon producing a prosthesis that replicated the anatomy of the missing limb, rather than replacing its function. The pragmatic approach of designing cost-effective, functional prosthetic limbs that could be mass produced did not achieve currency until the battlefields of World War I saw amputations carried out in unprecedented numbers.
The word “amputation” is derived from the Latin “amputatio,” meaning to cut or prune around. Although used in the classical Latin texts, the word did not appear in English until Peter Lowe’s “Discourse on the Whole Art of Cirurgerie” of 1597. Lowe was a Scottish surgeon who traveled to the continent to study medicine in the late 1560s. He completed his training in Paris at a time when the legendary Ambroise Paré dominated the surgical life of the city. Paré had been using the word “amputation” since the 1550s, and although it is not clear whether Lowe studied under Paré, so great was the latter’s stature that it is difficult to see how Lowe could not have been familiar with his work and teaching.
Annual statistics for patients newly referred to the 43 artificial limb clinics in the UK have been collected and published since 1996: initially under the National Amputee Statistical Database (NASDAB) heading and, since 2013, by Salford University7. The proportion of amputations performed for different causes has remained relatively stable although the proportion of below-knee, compared to above-knee, amputations has increased.
The NASDAB 20048 report identified 7000 amputations of the foot or lower limb from hospital activity data. Of these 4800 were referred to artificial limb clinics, of which 75% were due to vascular causes, 9% infection, and just 7% trauma, of which 80% were at trans-tibial or trans-femoral level. Although the incidence of traumatic amputations and congenital abnormalities is relatively low, as the patients have a long life expectancy their prevalence in amputee clinics is relatively high – 36% in the Cambridge unit.
The 2013 British documentation of amputations9 concluded that the incidence of lower limb amputation is up to 8 to15 times higher in diabetic patients, compared to non-diabetic patients. Further, the number of people with diabetes in the UK has increased from 1.4 million to 2.9 million since 1996 and is likely to reach 5 million by 2025. Furthermore, up to 70% of people die within five years of having an amputation as a result of diabetes. The risk of death within 30 days of a lower limb amputation may be as high as 17%, reflecting the patients’ multiple comorbidities.
In planning an amputation, the primary aim is to produce a stump that will heal well and comfortably, and will accept the prosthesis best suited to the patient’s functional needs. At what level and by what technique this is best achieved will, of course, depend upon many factors: age, health and functional demands of the patient, the pathology necessitating amputation, the state of the limb in question, the state of the patient’s other limbs, and so on. In general, however, more distal amputations allow better function with higher walking speeds and lower energy consumption, while more proximal amputations offer more generous soft tissue cover, better tissue perfusion and therefore more reliable wound healing. As a rule of thumb, therefore, amputations should be carried out at the most distal level that offers a high probability of wound healing. Preserving limb length should not prejudice successful soft tissue cover. To do so, especially in diabetics, can often lead to “nibbling up the leg” in a series of unsatisfactory, and ever-more proximal, amputations.
There is, of course, more to planning an amputation than the technicalities of producing a satisfactory residual limb. Planning for the prosthesis is also of crucial importance, not just in terms of prosthesis fitting and suspension, but in determining the level of the amputation. This is because there is an inherent tension between limb preservation and prosthetic options. In simplistic terms, the more limb that is retained, the less length available to accommodate prosthetic components if the limb–prosthesis composite is to match the length of the contralateral limb. There is, of course, also the profound psychological impact of the procedure to be considered. For these reasons patients considering undergoing elective amputation surgery should ideally be assessed by a multidisciplinary team including a specialist in amputee rehabilitation and a psychologist, as well as the surgeon. Good communication between the different members of the team is essential to optimize patient outcome. Some patients find meeting amputees in a support group helpful in setting their expectations of the journey upon which they are to embark.
It is also important that patients considering undergoing an amputation, as opposed, perhaps, to limb reconstruction, should be aware that amputation is seldom the final procedure to be performed on the limb. Further surgery and stump revision may be required with the passage of time.
Elective amputations should be planned in collaboration with the prosthetic service that will be providing amputee rehabilitation in the long term, as there may be local variation in prescribing and limb-fitting practice. Contact should be made with that service as soon as possible after emergency amputations to allow patient expectations to be managed in a consistent way from the earliest stage. There are 43 NHS prosthetic and amputee rehabilitation services in the UK, many with satellite clinics10, and a small number of private firms, which mainly cater for people with compensation claims following injury.
Flaps should be kept thick with ample vascular supply. Excessive soft tissue dissection and undermining should be avoided. The aim should be to produce a sturdy soft tissue envelope for the stump. The various flap options at each level have been defined, but an atypical or unusual flap may be preferable to a more proximal level of amputation.
As a general rule, resection of muscles at least 5 cm distal to the level of bone resection will allow sufficient cover for the stump. Myodesis (suturing of muscle to bone) or myoplasty (suturing of muscle to its antagonist or fascia) confers greater soft tissue stability of the stump, reduces the risk of contractures, and reduces the rate of muscle wasting following the amputation.
Most amputations are performed under tourniquet, except for cases of severe infection or ischemia where assessment of tissue viability is crucial.
Small vessels may be cauterized, but major structures should be ligated. It can be a useful practice to undertake this proximal to the level of bone resection to facilitate re-exploration of the stump and revision surgery if necessary. The tourniquet should be deflated and careful hemostasis carried out prior to skin closure.
By definition, any transected nerve will form a stump neuroma. A variety of techniques have been described in the hope of preventing formation of painful neuromas: electrocautery, perineural closure, silastic capping, and burying the nerve stump in bone or muscle, to name but a few. None has won widespread acceptance and most surgeons now content themselves with pulling the nerve gently down into the wound and cutting it cleanly – well proximal to the level of bone resection – and allowing the nerve to retract into the stump.
Bone resection is classically carried out with a Gigli-type saw to avoid osteonecrosis, although a power saw with saline cooling may also be used. Periosteal stripping should be kept to a minimum to avoid bone overgrowth. Sharp bone edges and prominences should be resected and rasped to a smooth edge, especially in areas of poor soft tissue cover, such as the anterior tibia.
In an emergency, such as trauma or overwhelming infection, amputation of all or part of a limb may be carried out to save life or to preserve as much of the viable segment of the limb as possible. In such circumstances the luxury of detailed prosthetic and reconstructive planning is seldom available and the tissues available often determine the operative technique.
Even under these conditions some general principles should be followed:
1. All non-viable tissue should be removed.
2. The limb stump should be of sufficient length to accept a prosthesis easily but not so long as to make it difficult to accommodate a prosthesis distal to it.
3. The stump should be stable but well padded with good soft tissue cover over bony prominences.
4. If possible the stump should be sensate and motor control of the residual limb maintained with, for instance, a stable working knee joint.
In some cases in which trauma or sepsis has rendered the attainment of satisfactory soft tissue cover with good primary wound healing over a stump of satisfactory length unlikely, open amputation may offer an alternative to the fashioning of a short residual limb or amputation at a higher level. In an open amputation full wound closure is not attempted at the first operation. Rather the soft tissues are closed at a second procedure, with tissue transfer by a plastic surgeon, or the wound healed by secondary intention usually with a negative pressure dressing. This reduces the likelihood of persistent infection and wound failure but leads to a more protracted recovery with delayed ambulation.
Surgery and Prostheses at Each Level of Amputation
Forefoot amputations for diabetes, infection, or trauma are common. They require little or nothing in the way of prostheses and, as a rule, amputation of a single toe causes little disability. In the long run it may facilitate the development of malalignment of adjacent digits, such as hallux valgus arising after second toe amputation. In theory, this could be prevented by ray amputation, but against that one has to weigh the risk of transfer metatarsalgia developing as a result of the loss of a metatarsal head. In contrast to the hand, therefore, the role of central ray amputation in the foot is relatively limited: the interdigital gap caused by loss of a toe is largely cosmetic but loss of a load-sharing metatarsal head may cause significant functional problems secondary to the disruption of the metatarsal arcade. The exception is the lateral border amputation of the fifth metatarsal, such as for an ulcer of the fifth metatarsal head. In this procedure the fifth toe and most of the fifth metatarsal can be excised through a laterally based tennis-racquet incision. It is important that the insertion of peroneus brevis is not sacrificed, as this will lead to the development of hindfoot varus. Skin viability can be an issue, so the flaps should be kept as thick as possible.
Amputation of the great toe, or of all the lesser toes, with preservation of the metatarsal heads, usually causes little difficulty with slow walking, but can cause problems with brisk walking or running due to a loss of push-off in terminal stance. In the first ray, in particular, it is therefore desirable to try to preserve the proximal 1 cm of the proximal phalanx, and therefore some function of flexor hallucis brevis (FHB), if possible.
The level of amputation of an individual toe is usually determined by the extent of tissue damage, be it by trauma or infection. All non-viable tissue should be removed and a long plantar and shorter dorsal flap should be fashioned of healthy tissue. The flexor and extensor tendons should be cut cleanly and allowed to retract into the proximal tissues. The bone end should be smoothed after resection and the flaps closed traditionally with non-absorbable sutures.
A large proportion of forefoot amputations are carried out in diabetics, usually with infected ulcers and often osteomyelitis. Traditionally full primary closure, with or without a delay, has tended to be avoided in favor of healing by secondary intention, split skin grafting, or even the use of free-tissue transfer. In recent years vacuum-assisted closure, also known as negative-pressure wound therapy (NPWT), has been increasingly used and has been reported to increase the amount of the foot that can be usefully retained and to be less expensive and less labor intensive than conventional dressings11. More recently, however, Shaikh et al.12 showed good results following amputation combined with primary closure in infected diabetic feet so long as meticulous tissue handling was observed and the surgery was combined with judicious use of antibiotics and tight diabetic control.
The transmetatarsal amputation can produce a highly acceptable functional and cosmetic result if tissue perfusion is reasonable, but is not recommended in the face of ischemia, unless combined with vascular reconstruction. Ideally the bones should be contoured to replicate the anatomic forefoot cascade as far as possible to facilitate even weight bearing. The bone ends should be beveled to prevent irritation and breakdown of the underlying soft tissue due to excessive pressure. A long plantar flap is most commonly performed, but equal dorsal and plantar flaps can be effective. The flaps should be fashioned slightly longer on the medial side to provide cover for the greater depth of the foot on that side.
Many forefoot amputees manage well without prosthetics, but if an amputee chooses a toe filler and insole can be supplied: this may be molded to provide total contact, if pressure relief under the first metatarsal head is the priority, or be chosen from a range of thin, flat carbon fiber plates if energy return while wearing normal shoes is more important. A bespoke silicone toe (Figure 5.1) may be attached to an insole for patients who wish to have normal-looking feet in open-toed sandals. All diabetic forefoot amputees should be referred for custom insoles to reduce the risk of further diabetic ulceration or pressure points.
Figure 5.1 Bespoke high-definition silicone toes, fixed to sandals, following forefoot amputation.
Lisfranc and Chopart documented their techniques for partial amputations 200 years ago. The advantages of such amputations include the preservation of normal limb length, knee and some ankle function, as well as the production of a good end-bearing residual limb. These procedures can be carried out under regional anesthesia with a popliteal block. Disadvantages include the late development of equinovarus contractures, poor wound healing, particularly in diabetics and vasculopaths, and limited prosthetic options.
Evidence comparing the outcome of partial foot amputations to trans-tibial amputation (TTA) is poor. In many series, partial foot amputations have not been shown to result in better functional outcome than TTA, but amputees are such a heterogeneous group that it is difficult to draw any definitive conclusions13. Moreover, Millstein and colleagues reported retrospectively that in a group of 260 partial-foot amputations resulting from industrial accidents, tarsometatarsal and midtarsal levels gave the most successful outcomes14. Despite this, functional outcome from amputation at this level is often disappointing as the residual foot is large enough to limit the prosthetic options but too small to function as a foot. These poor outcomes may just reflect the problems of the patient cohort for whom this amputation is selected, rather than of the procedure itself. Brown found similar mortality rates but better ambulation scores in diabetics who had undergone midtarsal amputation for osteomyelitis or non-healing ulcers compared to matched controls who had undergone TTA. Neither group did really well, but the partial foot amputations seemed to be the best of a bad lot15.
On balance, partial foot amputations are perhaps most useful following acute, unreconstructable, forefoot trauma, in which cases it is difficult to justify a trans-tibial amputation as a primary procedure. They can also be used for frail, high-risk patients of low functional demand to whom the greater range of prosthetic options offered by a TTA are of little benefit, and in whom it is desirable to minimize the surgical and anesthetic insult.
In the non-diabetic partial foot amputation, only bone that is protected by innervated plantar skin should be preserved. Insensate plantar skin is of limited functional usefulness and vulnerable to ulceration. Grafted skin is insufficiently durable to survive long term on the sole of the foot.
Jacques Lisfranc de St Martin described disarticulation through the TMTJs, and the crucial step of dividing the plantar ligament running from the medial cuneiform to the base of second metatarsal, in 1815. It is widely believed that Lisfranc’s work was inspired by his service in Napoleon’s ill-fated Russian campaign of 1812, during which he observed a large number of fracture-dislocations of the midfoot, particularly in cavalrymen in whom the foot could easily be twisted violently in the stirrup. Sadly the reality is a little more prosaic. Lisfranc did not join the army until 1813, and served not in Russia but Germany at the Battle of the Nations. Moreover, when he presented his paper16 in May 1815 he identified “decay” and “crush injuries” as the main indications for his procedure.
Tarsometatarsal amputation can provide adequate functional outcome in some patients. Careful evaluation of the muscle balance around the foot is essential, specifically tendo Achillis tightness, and the function of the tibialis anterior and peronei should be assessed. Midfoot amputations significantly shorten the lever arm of the foot, so intraoperative tendo Achillis (TA) lengthening is often required. The insertions of tibialis anterior and the peronei should be preserved, or reattached if they have been removed during bone resection. These measures should prevent the muscle imbalance that, over time, can lead to progressive equinovarus deformity, difficulty in prosthetic fitting, and eventual loss of function. Special postoperative clamshell-type orthoses are useful during recovery to achieve maximal function. Postoperative immobilization of the limb in a cast may be necessary to allow the tendon reattachments to heal, as well as reducing edema and wound problems.
Chopart’s amputation through the talonavicular and calcaneocuboid joints was described a decade prior to Lisfranc’s. It removes the forefoot and midfoot while preserving the talus and calcaneus. Amputation through the midtarsal joint produces a foot with a very short lever arm, so rebalancing the TA, peroneal tendons, and tibialis anterior and posterior is especially important if long-term deformity is to be avoided. As with other partial foot amputations, plantar weightbearing skin is best used as a flap for closure; it is secured to bone to prevent shear, with bursa formation.
Prosthetic options following partial foot amputations are inevitably constrained by the limited space available in a shoe. The body weight is taken through the remaining sole of the foot, usually on a molded total-contact insole with a foam toe filler. This may be held in place on the stump by a soft leather bootee, which may fit inside the amputee’s ordinary shoes. Few of these amputees walk outside the house, but for those with greater ambitions, a stiff, full-length sole plate can be added to improve push off at the end of stance phase. This has only a modest effect unless supported by a rigid lever arm extending up the lower leg, in the form of an AFO (Figure 5.2) or, more effectively, a full prosthetic socket, which is necessary to climb stairs or ladders. There is no space for any energy-storing prosthetic components, and the socket is so bulky that shoes usually have to be supplied as well. An alternative approach, for cosmesis rather than function, is a bespoke solid silicone foot (Figure 5.3). These may be sculpted and colored to look very lifelike, but are much heavier and more expensive than the other devices. Fit people hoping to regain high activities after forefoot trauma may wish to consider the greater prosthetic options and performance of a TTA.
Figure 5.2 Molded ankle–foot orthosis with toe filler for a partial foot amputation. This provides maximum stability at the end of stance phase and when climbing stairs. It may require a larger shoe than on the remaining foot.
Figure 5.3 Low-definition silicone prosthesis for hind-foot amputation.
James Syme described his amputation in 184317, to avoid the considerable dangers of TTA in the nineteenth century. Harris18 subsequently revisited it. The Syme amputation is an ankle disarticulation, with removal of the calcaneus and talus and preservation of the heel skin and fat pad. This highly specialized tissue is used to form a durable end-bearing flap for ambulation. The heel skin is sutured to the distal tibia to prevent posterior and medial migration of the fat pad, which can compromise function. Construction of a successful Syme stump allows ambulation for short distances without a prosthesis.
Anterior and posterior flaps provide soft tissue cover in a Syme amputation. The posterior flap is marked on the sides of the foot by lines dropped from the tips of the medial and lateral malleoli to the sole of the foot. An oblique line joins these lines across the sole, as the lateral malleolus lies posterior to the medial. The tissues are divided down to bone. The anterior flap is marked with a line taking the shortest distance across the front of the ankle. The extensor tendons are divided, the ankle entered, and the medial and lateral collateral ligaments divided from inside the joint. The posterior flap is then developed by subperiosteal dissection along the plantar surface of the os calcis as far as possible, taking care not to buttonhole the skin. On the dorsal surface, the posterior capsule of the ankle is divided and dissection continued along the superior surface of the os calcis staying close to the bone. A bone hook may be used to draw the os calcis forward to facilitate this process. Alternating dorsal and plantar dissection of the calcaneum is continued until the foot has been removed. The medial and lateral plantar and anterior tibial vascular bundles are ligated and divided, and the nerves are cut under tension. The malleoli are removed with an oscillating saw to produce a flat distal tibia, and the edges of the bone smoothed with a rasp. Some malleolar flare should be maintained for cosmesis and to allow suspension of the prosthesis. The posterior flap should be sutured to the tibia so that the heel pad is held centrally under the tibia. This can be accomplished in several ways, including tenodesis of the tendo Achillis to the posterior margin of the tibia through drill holes, transfer of the tibialis anterior and extensor digitorum tendons to the anterior aspect of the fat pad, or removal of the cartilage and subchondral bone to allow scarring of the fat pad to bone with suturing of the plantar fascia to the anterior tibial periosteum. The flaps are closed and a rigid dressing applied and molded to hold the fat pad centrally.
Although reported outcomes are adequate at this level, many investigators emphasize the need for frequent and accurate prosthetic fitting and unloading of the “weightbearing” stump for comfort19. Prosthetic fitting can be challenging, as the distal stump may have a larger circumference than the proximal part, particularly after muscle wastage. Thus the socket must be close fitting around a thin proximal calf, but it must expand distally to allow the wider tibial flare through when it is applied. Historically this was achieved by leaving a large window in the hard socket, which was closed by lace-up leather flaps or by a rigid plate held on with straps. It is now usually achieved with a molded hard foam liner, made from Pe-Lite®. The external surface is built up over the thin middle section to make it cylindrical. It is also split longitudinally to allow it to gape while being pulled over the stump.
These prostheses are cosmetically unpopular as they are much bulkier around the ankle than a trans-tibial prosthesis. The small space available between the end of the stump and the ground limits the options for prosthetic ankles and feet, which reduces function compared to a trans-tibial prosthesis. Ideally, all the body weight should be taken through the heel pad at the end of the stump. If this is not tolerated and the end has to be offloaded, the socket has to be longer and the proximal socket will have to be a patellar tendon bearing at the knee. To accommodate the extra length, the contralateral shoe may have to be built up, further impairing the overall appearance.
Trans-tibial amputation is by far the most common of the major lower limb amputations. Historically it was mainly used in trauma and peripheral vascular disease, the transfemoral amputation being favored for diabetics because of supposedly better wound healing. In recent years, however, diabetes has become the most common indication for TTA.
The rise in popularity of the TTA is perhaps best explained by the fact that it is the most proximal level of amputation in the lower limb at which most amputees can expect to attain near-normal function. This is largely because preservation of the knee joint and limb length mean that when walking energy consumption, in both children and adults, is significantly less than following a transfemoral amputation20–21. Moreover, ambulant energy consumption has been shown to be no different in trans-tibial amputees to those who have undergone more distal amputations, such as a Syme. Overall use of prostheses and functional outcomes are significantly better in trans-tibial than transfemoral amputees.
In order for a TTA to be successful, certain preconditions should be met. The patient should have control of knee movement, particularly the quadriceps, the knee should be stable and pain free, and the stump will need to well perfused, preferably sensate, and of appropriate length. The use of free tissue transfer and microvascular techniques to improve soft tissue cover should be considered to allow a trans-tibial, as opposed to a more proximal, amputation.
There is no consensus as to the optimum level for TTA and, of course, the state and viability of the local tissues should be taken into consideration. In general, amputation through the distal third of the tibia is best avoided as the soft tissue cover will be poor and there are reduced prosthetic options. Traditionally, the optimum level for a TTA in an adult was regarded as preserving one inch of tibia per foot in height (= 2.5 cm per 30 cm). In most adults, therefore, bone sectioning will occur at 10 to 15 cm distal to the tibial tubercle. This remains our preferred practice, if the soft tissues allow. If there is doubt we recommend close liaison between the surgeon and prosthetist as to the minimum stump length required for the planned prosthesis. If a short trans-tibial stump is felt to be the best option, excision of the head of the fibula may allow the stump to fit more snugly within the socket. If it is not feasible to fashion a stump of acceptable length then through-knee or transfemoral amputation has to be considered.
A variety of techniques for TTA have been described but the long posterior flap remains the most widespread. With this a long posterior myocutaneous flap, based upon the blood supply from the gastrocnemius, provides the soft tissue cover. In most cases, particularly vasculopaths, this is the most reliably perfused flap, but in some instances alternative flap arrangements, such as skew and sagittal flaps, may be considered. In general, however, these alternative techniques have not been shown to improve outcome when compared to the posterior myocutaneous flap22.
In the long posterior flap technique, the junctions of the flap bases are marked, medial and lateral, just under two-thirds of the way back to the mid-posterior line at the level of planned bone resection. From these points a short anterior flap measuring 2 cm and a long posterior flap are mark. The posterior flap should be 1 cm longer than the diameter of the limb at the planned level of amputation and should have parallel, not tapering, sides when laid out flat. To avoid tapering the flap it is useful to elevate the limb and check the flap from behind before starting the dissection. The anterior flap should be raised with the skin, deep fascia, and the anterior tibial periosteum as a single composite layer at the level of the tibial section. The anterior tibial vessels should be ligated. The tibia can be sectioned with a short bevel on the anterior surface. The cut surface should be smoothed with a rasp as any sharp edges may cause pain. The fibula is sectioned 1 to 2 cm proximal to the level of tibial resection. In trauma cases with comminuted or segmental fractures, internal fixation can be useful to preserve length. The tibia is then retracted anteriorly and the posterior soft tissues are divided with an oblique cut running distally. When the desired level is reached, the cut is turned through 90° to divide the deep fascia and skin. The major vessels should be identified, transfixed, and ligated at a level proximal to the tibial bone cut, while nerves should be cut cleanly under gentle tension and allowed to retract into the soft tissues. The posterior flap is then fashioned with the skin and fascia left adherent to the underlying gastrocnemius. In most cases the soleus is removed from the flap, but some or all of its fibers can be retained if the flap would otherwise lack muscle bulk. The flap is wrapped around the distal tibia to gauge length; it should reach the anterior periosteum. The optimum length is marked and the flap cut to it under no tension. If necessary the corners of the flap can be tapered to prevent a bulbous shape. The tourniquet should be deflated at this point and meticulous hemostasis achieved. In order to stabilize the soft tissue envelope, and reduce muscle migration, the gastrocnemius can be sutured to the anterior periosteum or a formal myodesis with trans-osseous sutures performed. The flap should be secured with the knee in full extension to prevent flexion contractures.
In recent years, there has been interest in bone-bridging techniques to achieve a synostosis between the distal tibia and fibula. This can be a useful technique in revision surgery for those amputees suffering from tibiofibular instability and has the theoretical advantage of producing a better end-bearing limb. For primary amputations, however, the practical advantages appear limited: Kingsbury et al. could demonstrate no improvement in gait23, while Keeling et al. found no improvement in running or walking distance, prosthesis wearing, or general well-being24. Moreover, complication rates, particularly infection and non-union of the bone-bridge, are higher than for a standard technique25. A synostosis may also complicate subsequent revision surgery. If it is considered desirable to use a bone-bridging technique, the synostosis can be created either using a periosteal sleeve or fibular strut graft. The latter may be anchored with screws or transosseous sutures with or without a compression device.