For many centuries limb injuries have been common as a result of man's fascination with war. The outcome of even minor limb injuries inflicted on the battlefield was frequently fatal, in strong contrast to the minor importance accorded to such injuries in modern historical drama! The cause of these deaths was presumably sepsis from contaminated wounds containing devitalized tissue, but it was not until the last century that the important role of debridement was fully appreciated. The signs of increasing limb pain, swelling, and toxaemia were recognized as harbingers of death, and amputation the only cure. Despite the absence of aseptic technique, removal of the limb by incisions through healthy tissue with free drainage was often followed by recovery and healing. Although there are several descriptions of amputation dating back to Hippocrates the technique was popularized by Ambroise Paré around 1575 and was rapidly taken up by other military surgeons and then transferred to civilian practice. In the absence of true anaesthesia it was not possible to perform other than a guillotine type amputation and healing was always by secondary intention, although the later addition of a short flap by Yonge and Lowdham in 1679 made closure easier. To minimize the horrific experience surgeons became adept at rapid amputation, and apocryphal tales describe amputations performed in a few seconds, sometimes with the loss of the assistant's fingers.

The desire to replace an amputated limb with a prosthesis is as old as the procedure itself. There are many examples of false legs, mainly of wood, in ancient drawings and the use of a hook to replace the amputated forearm is also well known. The wooden peg-leg is common to many cultures and relied on the user kneeling on the weight-bearing surface. However there are also many examples of ingenious and complex artificial legs dating from the time of Paré.

The modern era of amputation surgery was heralded by the advent of anaesthesia and later by aseptic surgical techniques. It became possible to perform amputation using flaps fashioned to allow primary closure at a variety of levels. The huge number of amputees that followed the First World War provided a stimulus to the development of improved limb prosthetics, and many of the modern services originated at this time. Over the next few decades it was slowly appreciated that not all indications for amputation could be treated identically, and that patients presenting with infection or gangrene associated with loss of limb pulses had to be treated more cautiously in terms of amputation than patients undergoing amputation for trauma. Thus, Homans (1939) was voicing contemporary surgical opinion about the outcome of amputation in patients with vascular disease when he wrote ‘amputations below the knee can almost never be expected to offer a healthy stump’. Fortunately, this view has been proven incorrect, but the technical demands of amputations performed for limb ischaemia are emphasized by this quotation.

As a consequence of increasing specialization in surgery, limb amputations are performed by different specialists, depending on the indications. Thus orthopaedic surgeons most often perform amputations for trauma, and vascular surgeons usually amputate for vascular disease. Although the techniques used for different indications are broadly similar there are some individual factors that must be considered.

Since traumatic vascular injury is usually repairable, limb amputation is only indicated when there has been massive destruction of tissue, by either crush or blast injuries. The common causes are either motor accident or gunshot wounds in civilian practice. Ideally, the decision to amputate should be made immediately—a late amputation dictated by developing sepsis in an inadequately debrided limb represents a failure of management. Because the vascularity and health of the tissue proximal to the injury is usually good the amputation can be placed as distally as possible, within the limitation imposed by the need for excision of all devitalized tissue. Contamination with soil and dirt is common and in these cases the wound should be left open and closed secondarily after several days. Even if the early swelling of skin flaps prevents full closure by secondary suture, the basic health of the tissue will usually allow good healing by granulation and amputation at a higher level should be considered only if serious sepsis ensues.

Amputation of the limb of a patient with rest pain, sepsis, or gangrene with the aim of primary wound healing is perhaps one of the greatest challenges to a surgeon. Several factors conspire against the healing process. Firstly the tissue at the line of incision has often only slightly better vascular perfusion than the critically ischaemic or frankly gangrenous tissue distally, and the lack of oxygen and nutrients delivered to this tissue means that the ability to form new vessels in granulation tissue and the metabolic processes involved in healing are greatly impaired, as is the ability to fight bacterial contamination. The lack of perfusion may be made worse by the frequent coexistence of poor cardiac function, respiratory disease and anaemia. Furthermore many patients have sepsis already within the limb and there may be considerable oedema as a result of this and other coexisting conditions such as heart failure. Many patients are diabetic, further reducing resistance to infection.

It therefore behoves the surgeon to pay attention to these factors and correct those that are correctable, wherever possible. Heart failure, anaemia, chest infection, and limb infection should all be treated, since this may make the crucial difference between success and failure. Unfortunately delay is not always humane or wise, since the toxaemia from an infected ischaemic leg may make factors such as heart failure worsen despite treatment until the limb is removed, and opiates required to treat severe rest pain may make a chest infection untreatable. The judgement to wait or proceed is often difficult and is probably best taken by the surgeon since other physicians may not appreciate the role of the ischaemic limb itself in dictating the general condition of the patient.

Amputation for a tumour affecting a limb is now a relatively rare occurrence, since most tumours can be controlled locally by a combination of radiotherapy and cytotoxic therapy, even if excision is not possible. When amputation becomes necessary it is always a planned procedure, usually in a limb with good vascularity without sepsis and the amputation is usually, therefore, technically quite straightforward. Psychological effects are likely to be the greatest problem in this group, since there is ample opportunity to brood on the forthcoming operation and the subsequent disability is compounded by the fear of recurrence of the tumour.

At one time amputation was considered the best option for a number of congenital deformities, on the basis that this gave superior cosmesis and would ultimately lead to better acceptance by society. However, it has now been realized that even patients with severely deformed limbs usually can obtain better function from the deformed limb, albeit with a number of prosthetic additions, than by amputating the limb and attempting to fit a conventional prosthesis. Furthermore, such patients are unlike normal limb amputees, in that they have never developed a body image of themselves with a full-size limb and do not adjust easily to a full-size prosthesis. A limb remnant represented by a single digit may be used to activate a control for a mechanized transport or prosthesis and should not be removed.

The only common exception to this rule is the deformity caused by congenital absence of the fibula, resulting in shortening of the lower leg. Conversion to a Symes amputation produces an end-bearing stump which may be fitted to a foot prosthesis that can result in remarkably normal function in young children.

Amputation is rarely required for a limb that is severely deformed, useless, or painful from acquired disease. Examples of conditions that sometimes require amputation are neurological diseases leading to a painful flail limb and chronic osteomyelitis, where renal failure due to subsequent amyloidosis is a danger. These indications have become increasingly rare since the following points have been understood. Firstly, even a severely deformed limb can usually function better than a prosthesis. Secondly, a paralysed limb can still provide support following arthrodesis of appropriate joints or by the use of splints and gives better cosmesis. Thirdly, persistent severe pain of central neurological origin may make a limb useless but, unfortunately, removing the limb often does not result in relief since the pain persists in the ‘phantom’ limb.

The site of amputation is influenced by several factors. The overriding general principle is that the least mutilating procedure possible should always be chosen, a principle that is particularly important in the lower limb, where the retention of the knee joint results in much greater mobility after rehabilitation. However other factors are also important, the most obvious being the indication for amputation. In the case of amputation for malignancy the amputation must be performed at a site that is sufficiently proximal to ensure clearance of the malignant process and minimize the chances of recurrence. This usually means removing the limb at a level that includes the joint proximal to the lesion. In the case of amputation for trauma the amputation site should be as distal as possible, subject to the possibility of fitting a usable prosthesis. In the case of amputation for ischaemia the choice of amputation site is a balance between using as distal a site as is compatible with removing dead or near-dead tissue and a proximal site sufficient to ensure healing of the wound.

In the past the ease of fitting the limb for a prosthesis was also a factor. For instance, above or below knee prostheses were easier to fit and functioned better than through knee prostheses. However the steady advance in materials and design of prostheses has meant that this factor is less important than previously, although it is still difficult to get a good cosmetic result from a jointed prosthesis for long thigh amputations such as the Gritti-Stokes.

The most important requirement for successful amputation is skin healing, preferably by first intention. The skin that is to form the amputation flaps must at least be viable at the outset. Signs of demarcation, fixed staining, or anaesthesia of the flap skin are obvious contraindications to its use. The presence of excoriation, cellulitis, or frank ulceration are relative contraindications, although the likelihood that the phenomenon is entirely localized and that infection has been treated adequately must be considered. The tissues beneath the skin must also be healthy. The presence of deep tenderness and loss of function may signal necrosis of deep muscles or sepsis, which make an amputation unlikely to heal. The skin temperature is more a reflection of the ambient temperature, or the presence of sepsis or systemic illness and is not a reliable guide to subsequent healing. The absence of a pulse at the next most proximal palpation point has been said to be a contraindication to more distal amputation and vice versa, but this is certainly an unreliable sign. The decision about the level and type of amputation should also not be taken during the operation, since the degree of bleeding noted on incision of tissues at operation is a very poor guide to subsequent healing, being greatly affected by anaesthetic agents and transient hypotension.

The decision to undertake a more distal amputation is inevitably associated with a greater risk of failure to heal and the need for reamputation at a higher level. Furthermore, the exact magnitude of risk has been shown to vary considerably from centre to centre, and to make a rational decision it is necessary to know one's own results: a valuable contribution of continuous audit. The final decision should always take into account all of the factors influencing outcome including the physical and mental capacity of the patient to withstand a reamputation and whether the patient is likely to make use of the increased mobility that a joint saving amputation may give, to name just two.

A variety of techniques for assessing the blood supply to the amputation flaps, including thermography, transcutaneous oxygen measurement, cutaneous perfusion measurement, xenon-133 clearance, Doppler ultrasound, and digital plethysmography have been described. There is no doubt that in the hands of enthusiasts these techniques can allow the correct choice of amputation site in a larger proportion of cases than is probably achieved by clinical judgement alone. However, all of the techniques require some degree of expertise in their performance and some are not readily available in many hospitals. No single technique has been shown to be sufficiently superior to the others to attain widespread acceptance, although transcutaneous oxygen measurement is probably the most valuable.

Although general anaesthesia is preferred by most centres it is possible to use local anaesthesia, nerve block, epidural, or spinal anaesthesia in appropriate situations. Some surgeons object to the use of techniques other than general anaesthesia on the grounds that the noise and motion associated with amputation is psychologically distressing to the patient, even in the absence of pain. However, the use of amnesic sedative agents during the procedure is highly effective and most patients remember nothing if such agents are used.

Local anaesthesia is only useful for digital amputation and has the potential disadvantage of causing local swelling of the tissues which may make closure less easy. There is also the theoretical (but in practice, remote) disadvantage of spreading infection. Local anaesthetic combined with adrenaline must never be used as this may further prejudice the blood supply.

Nerve block techniques have the advantage of avoiding the dangers of general anaesthesia and hypotension, but unfortunately tend to be unreliable, and conversion from nerve block to general anaesthesia in the middle of an amputation is distressing not only to the patient but also the theatre staff.

Epidural or spinal anaesthesia is usually very adequate for amputation of the lower limb, producing vasodilation as a by-product which certainly does no harm from the point of view of the procedure. A tourniquet should never be used for patients with vascular disease and is usually unnecessary even in patients with normal vessels. Broad spectrum antibiotics, such as penicillin, with activity against Clostridium spp. must always be given preoperatively.

After careful identification of the patient and the correct limb, the limb should be carefully painted with skin-prep prior to starting surgery, wrapping any gangrenous or infected portion in a plastic bag, sealed to the skin above by adherent non-porous tape. The incision lines for flap formation must always be marked out using a sterile marker.

In modern surgical practice approximately 80 per cent of amputations are performed for ischaemia secondary to vascular disease. The life expectancy of this generally elderly group of patients is limited, mainly due to concomitant cardiovascular and cerebrovascular disease. Approximately 50 per cent of vascular amputees will die within 3 years. The main consideration in performing the amputation is to obtain stump healing. In non-vascular amputations, such as those for trauma or tumour, the patient is often young and a lifetime of prosthetic use must be anticipated. In these patients more advanced procedures to obtain maximal function may be indicated. Since the majority of surgeons perform lower limb amputations for ischaemia most of the techniques described below will be those recommended for this indication. Modifications and additional procedures applicable to amputation in non-ischaemic limbs will be considered as addenda to each technique. In those amputations indicated mainly for non-ischaemic disease, such as those in the upper limb, the reverse order will apply.

The amputation technique should be governed not only by the disease process and anatomical principles but also by the prosthesis available to replace the severed limb. A comment on the prosthesis available is therefore a necessary adjunct to the description of each amputation technique.

A ‘racquet’ shaped incision is recommended (Fig. 1) 399. The tissues should be severed from the incision line directly down to the shaft of the phalanx and dissection of tissues should then continue proximally as close to the bone as possible, to avoid the possibility of damaging the digital vessels laterally and thus compromising the blood supply to the flaps and other digits. It is not usually wise to amputate just the distal phalanges for vascular disease and the dissection should continue up to the metacarpophalangeal joint, which is disarticulated to allow removal of the digit. The head of the metatarsal should then be removed by the careful use of nibbling forceps preventing further damage to the surrounding tissues. The bone should be nibbled back until the tissues can be seen to be falling in to cover approximately half of the cut end of the bone. Tendons should be pulled down with forceps, cleanly severed as short as possible, and allowed to retract. The incision should not be sutured closed but is best packed lightly with ribbon gauze, to limit bleeding, and allowed to heal by granulation. Prostheses are not usually required.

The usual indication for toe amputation in non-ischaemic limbs is for trauma. The wounds can be closed primarily provided that there is no danger of contamination. Amputation of distal phalanges may be contemplated for cosmetic reasons.

This amputation performed in an ischaemic limb allows surprisingly good function, despite the loss of the great toe, but is undoubtedly the amputation with the highest failure rate. The important requirement is an adequately vascularized flap of sole skin, allowing application of the technique shown in Fig. 2 400. It is uncommon to have an adequate vascular supply in the presence of gangrenous toes, and for this reason the amputation is not often performed for vascular disease. However the correct indications do occur sometimes in diabetic patients who have localized digital gangrene in the presence of foot pulses. If insufficient vascularized sole skin is available to ensure primary closure the amputation should not be performed as secondary closure often fails in patients with vascular disease. The dorsal transverse incision is first made over the metatarsals as shown (Fig. 2) 400, incising through the dorsal tendons down to bone. The metatarsals are divided using a bone cutting forceps, or more satisfactorily, a power saw, and then the distal bones are elevated to allow careful dissection posteriorly close to the periostium and eventual separation and removal of the forefoot, leaving behind the sole, which should be kept as long as possible at this stage. Each metatarsal shaft should then be nibbled back for a distance of 1.5 cm using bone nibbling forceps to allow the tissues to fall in over the bone ends. The exposed tendons should then be removed from the sole flap by pulling down with forceps and then cleanly severing the tendon as short as possible. After ensuring haemostasis the flap is brought forward over the metatarsals and cut to the appropriate length to allow closure without tension, using deep interrupted absorbable sutures and an atraumatic skin closure (either steristrips or subcutaneous prolene inserted without instruments is preferred by the author).

In patients without ischaemic disease it is possible to leave the flaps open if gross contamination has occurred due to trauma, and use skin grafts later if full closure cannot be obtained.

A simple shoe filler is usually sufficient to provide a good functional result.

Owing to the poor healing and inevitable equinus deformity caused by unopposed contraction of the ankle flexors these amputations are not recommended in either vascular or non-vascular patients.

Syme's operation is the best amputation for patients with extensive non-ischaemic damage to the forefoot. The stump that is produced has the advantage of allowing considerable weight bearing, and thus allows considerable independence in the home without a prosthesis. It may suffice alone in areas where prostheses are not available. The defects of the procedure used to be difficulty in fitting a foot prosthesis because of lack of clearance from the ground, which has now been overcome, and poor cosmesis of the prosthesis in females, due to the thickness at the ankle: this is still a problem. There is a difference between the technique that can be used in children and in adults although the incisions used are similar (Fig. 3) 401. In children the amputation uses the same heel flap to cover the end of the stump, but the amputation itself can consist of simple disarticulation of the talus, followed by removal of the malleoli, leaving behind most of the joint surface. In adults the incision is made as shown in Fig. 3 401. The anterior incision is then deepened through the anterior tendons to allow disarticulation of the talus from the joint. The Achilles tendon is divided posteriorly and the talus and calcaneum are then dissected free from the heel by sharp dissection, staying close to the periosteum to avoid damage to vascular structures. The dissection is particularly difficult over the thin skin at the back of the heel, and there is a danger of ‘buttonholing’ the skin, which can again be avoided by staying close to the periostium. A variant operation described by Boyd is to leave a slice of the calcaneum attached to the heel skin, which can then be attached to the proximal cut bones in a similar manner to the Gritti-Stokes operation. This may lessen the trauma to the tissue but lengthens the stump.

Once the posterior flap has been dissected free the distal foot is removed, leaving the heel flap long at this stage. The exposed distal tibia and fibula are dissected free of adherent tissues staying on the periosteum. A transverse cut is then made approximately 1 cm above the ankle mortice. It is important that this cut is truly transverse, to prevent uneven pressure later when weight bearing. It is also important to leave some element of the malleolar prominence on each side, since this allows stabilization of the prosthesis and prevents rotation. After cutting tendons short and allowing them to retract and ensuring haemostasis the flap is brought over the cut end of the bone and trimmed to a suitable length without tension. The wound is closed by suturing the deep tissues of the heel pad to the deep fascia anteriorly with a suitable atraumatic skin suture.

Syme's amputation requires a relatively good blood supply to the heel pad and most patients that need to lose the foot from vascular disease do not have sufficiently good blood supply to allow this procedure. However, some diabetics with forefoot infection or gangrene may have normal ankle pulses and a Symes amputation can be successful.

This is the most common amputation performed for vascular disease, and possibly the most demanding technically. Preserving the knee joint is of such major benefit in terms of rehabilitation that below knee amputation should usually be undertaken in all patients with non-reconstructible ischaemia distal to the calf, unless there are clear indications that the operation is certain to fail. Such contraindications would include necrosis of the potential posterior flap skin, or a patient who is certain not to benefit from the advantages of a below-knee prosthesis, for example where the knee has been arthrodesed. The operation described is a variant of the Burgess long posterior flap technique. A more complex technique, based on similar principles, is the skew-flap technique (see bibliography), which probably has similar healing rates but produces a more ideal conical stump.

The important principles of the technique are shown in Figs. 5 to 8 403,404,405,406. The stump should be as short as possible to maximize the chances of healing but long enough to allow optimal use of a below knee prosthesis: 9 cm from the knee joint line is ideal in thin legs, 11 cm in fat legs, and 7.5 cm can be regarded as the absolute minimum. A long posterior flap is used which should initially be kept as long as possible, and only trimmed prior to final suturing. The flap is based on half the circumference of the leg to maximize the blood supply from above. After outlining the flaps the skin incision is deepened through fascia, marking the tibia anteriorly to denote the later line of bone section. The muscle of the anterior and lateral compartments is divided to expose the fibula, ligating the anterior tibial vessels. The fibula is divided with a lateral bevel 1.5 cm proximal to the tibial line. The tibia is divided at the marked line with an anterior bevel and the lower leg removed by dissecting the posterior flap away from the bones. Bone dust should be washed away to prevent spur formation. Bleeding can be controlled by folding the flap forward and applying pressure. The bulk of the flap should be reduced by complete removal of the soleus muscle (Figs. 6, 7) 404,405, which does not contribute to the blood supply of the flap. If the lateral portions of the gastrocnemius muscle bellies are particularly prominent these should also be trimmed to reduce ‘dog-ears’. The posterior tibial and peroneal vessels are individually ligated whilst removing soleus. The flap is then brought forward to the tibia, cut to a suitable length and sutured, deep fascia to periostium and deep fascia, over a vacuum drain. Skin closure should be performed with meticulous care of the skin. The author's preference is for subcuticular prolene inserted by hand needle, without the use of instruments, and covering steristrips (Fig. 8) 406.

A longer stump may be of considerable benefit in young patients with non-ischaemic limbs: 15 cm is ideal. Short anterior or lateral flaps may be used to keep the scar away from the prosthetic bearing surfaces. More complex myodesis procedures may also allow better function. A fibula bone bridge can be inserted between the tibia and fibula to allow the Achilles tendon to be brought over and sutured to holes in the anterior tibia. The anterior and posterior tibial muscles can be sutured together over the interosseus membrane.

Most below-knee prostheses rely on a patellar tendon bearing socket with either strap or suction suspension (Figs. 9, 10) 407,408. A variety of ankle joints with varying degrees of ankle and toe flexion are available.

In non-ischaemic limbs where the knee joint is non-functional or when the limb is damaged in such a way that a below-knee amputation is not possible this operation has several advantages: it is quick and easy to perform, using the lateral flaps shown in Fig. 11 409; the stump is end-bearing, has good leverage, and the bulbous bony shape can be used to suspend the prothesis. Modern polycentric knee prostheses have overcome many of the difficulties of limb-fitting and the cosmetic result is reasonable.

The flaps required to cover the bulbous end of the femur are of considerable length and use less-well vascularized lateral skin. Healing may thus be poor in patients with vascular disease. A below-knee amputation will heal as well and if the ischaemia is considered inadequate for a below-knee amputation a knee disarticulation is unlikely to heal either. This operation is not usually performed in patients with vascular disease, except where retention of the knee joint is contraindicated and a fast bloodless amputation is required.

Prostheses which are available tend to be rather bulky and of similar design to those used after Gritti-Stokes amputation with external joints (see Fig. 13 411). However the bulbous end can be used to allow suspension without a thigh corset and the socket can be end-bearing.

The Gritti-Stokes amputation is a variety of through-knee amputation that has the advantage of using relatively well vascularized anterior skin, and so is useful for patients with or without vascular disease. The operation is useful for the few patients with ischaemic limbs who appear to have good perfusion below the knee but do not have the posterior flap of skin available to allow below-knee amputation (perhaps because of local infection or ulceration) and in those patients where retention of the knee joint is of no advantage. Gritti-Stokes amputation gives a long lever with partial end-bearing, although not as good as that obtained by knee disarticulation, and although the end is bulbous direct suspension of the prosthesis using the bulbous end is often not possible. Prostheses are good functionally but cosmetically inferior to those available for patients with above-knee amputation.

The basic points of the technique are shown in Fig. 12 410. The key point is the fixation of the patella remnant to the cut end of the femur. Some authorities recommend drilling the femur and using wire sutures for this purpose but the author has found that six strong absorbable sutures between the strong tissues at the periphery of the patella to the periosteum and surrounding tissues of the femur are adequate to ensure bony union. It is also helpful to make the cut through the femoral condyles proximal enough to ensure there is relatively little tension and with a slight backward slope that encourages the patellar remnant to stay attached (Fig. 12) 410.

The principles of the technique for patients with ischaemic limbs are shown in Fig. 14 412. Important points to note are that although a long femoral stump has advantages for leverage of a prosthesis this should not be at the expense of poor healing, as is often the case in excessively long stumps where the bone may be inadequately covered with soft tissues and the skin is stretched. In any case the site of division of the femur should not be more than 12 cm from the knee joint line to allow room for the knee mechanism in the prosthesis. The minimum length of stump that can be used to fit an above-knee prosthesis is 7.5 cm below the adductor muscle insertion. If the amputation must be shorter than this and it is hoped eventually to fit a prosthesis, a hip disarticulation should be performed, as this allows better fitting of modern prostheses and reduces weight.

The bone end should be rounded off and covered by suturing the deep fascia of the posterior flap, with its attached muscle, to the same structures anteriorly, enclosing a vacuum drain. The author prefers a subcuticular polypropylene suture with additional adhesive strips for closure.

Where amputation is performed in young patients with good vasculature the addition of more complex myodesis techniques to suture and overlap opposing groups of muscles using drill holes in the distal femur produces a stump that is a conical shape with better function. However these techniques do not work well in ischaemic limbs and may result in devitalization and failure of healing.

There is insufficient space to describe in detail the dissection required for these operations, which are usually performed in non-ischaemic patients, and the reader is directed to the references at the end of the chapter. The principle is to remove the required extent of the limb leaving a posterior flap with a layer of viable muscle that can be used to cover and cushion the deep structures (Fig. 17) 415.

Hip disarticulation and/or hemipelvectomy are rarely required for ischaemia, and when they are required infective gangrene is the usual cause. It is often not possible to produce the ideal flaps described for elective operations, and it is a question of using the best vascularized non-infected flap of skin available to cover the amputation site. Vascularized free flaps may be necessary. Mortality is high and serious consideration should be given in the elderly to non-intervention to allow death with dignity.

The usual indication is trauma and the general principle is to preserve as much viable tissue as possible, particularly preserving palmar surface skin and as much of the thumb as is practicable. Multiple injuries require specialist attention. Medial and lateral flaps should be used, which may be closed unless there has been gross contamination. If all the fingers and the palm of the hand are lost the wrist should be preserved since useful grip function can be obtained by use of an opposition plate.

The usual indication is for trauma. Equal length extensor and flexor flaps are fashioned as shown in Fig. 19 417, allowing bone section at around midforearm, if possible. Longer stumps than this have little advantage and tend to heal poorly. The minimum length that can work a forearm prosthesis is 5 cm below the biceps insertion. The extensor and flexor muscles are then sutured together under gentle tension over the bone ends, including the deep fascia with each bite.

A variety of forearm prostheses are available (Figs. 20–22) 418,419,420. Some use mechanical hands with the grip mechanism worked by a shoulder strap and cable (Fig. 20) 418, or an electrically driven grip activated by muscle depolarization (Fig. 21) 419. However, many patients find these cosmetic prostheses both heavy and functionally poor. The most functional prosthesis remains a split hook (Fig. 22) 420 with special tools which can be fitted for specific tasks.

Equal length anterior and posterior flaps are used (Fig. 19) 417. The bone should ideally be divided no closer than 9 cm above the medial epicondyle, to allow room for an elbow mechanism. The minimum length of humerus that allows an above-elbow prosthesis is 5 cm below the anterior axillary fold. However, provided that there are not over-riding reasons such as clearance of malignancy even a non-functional amputation across the humeral neck should be retained in preference to disarticulation, since it gives better cosmesis. The anterior muscle flap should be sutured over the bone end to the posterior muscle flap, with bites that include the deep fascia.

Although cosmetically acceptable prostheses are available, some with active grip using a shoulder activated cable (Fig. 23) 421, the functional result is poor and these limbs are rarely used functionally if the patient has a normal contralateral arm.

Virtually the only indication for these procedures is amputation for malignancy. There is insufficient space to discuss the details of the dissection required, but the incision for both is placed so as to allow a predominantly posterior flap to be brought forward to cover the amputation site, with an extension of the incision anterosuperiorly to allow access for proximal ligation of the limb vessels (Fig. 19) 417. The operation results in considerable deformity which is difficult to disguise with prostheses that allow normal fit of clothing and retain function.

Although functional prostheses are available (see below) they are weighty and have poor function, so that many patients opt for a cosmetic shoulder filler only.

A number of complex operations have been described in the past to allow better function of arm prostheses. Most of these have now been abandoned and those that are still performed are usually not worth considering in a patient who has one normal arm. The formation of a skin-lined tunnel in the biceps muscle (‘kineplastic tunnel’) can sometimes be used to activate the terminal device of a forearm prosthesis. Patients that have lost both forearms can be given grip function by Krukenberg's operation, which separates the radius and ulna to act as separate jaws of a pincer. Although surprisingly good function can be obtained, the cosmetic penalties are considerable and are accepted by few patients.

There are considerable differences in opinion on the dressing, bandaging, and prosthetic fitting of amputation stumps. It is the author's opinion that these differences stem from the experiences of different specialties dealing with either predominantly amputations performed for vascular disease or for non-vascular conditions, and the management of these groups should probably be different.

Amputations for vascular disease should be dressed with a loose-fitting stump bandage, without compression. The bandage will tend to fall off unless supported by multiple longitudinal strips of adhesive tape, but compression may produce local necrosis of skin that is balanced on the edge of ischaemia, and a tourniquet effect, which is easily produced by inexpert bandaging, can be disastrous. Although the dressing is best left undisturbed if all is well, the stump must be viewed if there is any question of infection. The use of a plaster cast, suggested by some authors, is not recommended since it impedes the ease of access to the stump. Gentle active exercise should be encouraged by the physiotherapist, within the limits of pain. Some authorities recommend immediate fitting of a prosthesis and encourage rapid progression to walking within a few days. In the author's opinion this method may be possible in a centre with prosthetists able to attend at all times and the staff able to supervise walking properly, but for most hospitals such facilities are not available and this approach is potentially dangerous. The first goal is wound healing and the author recommends that the stump be left until primary healing of the wound is well under way, usually 2 weeks or more, before trying a few steps under supervision in a pneumatic pylon to regain bipedal balance. Subsequent progression to an adjustable training prosthesis encourages early learning of correct gait with moulding and shrinkage of the stump and the final prosthesis can often be fitted by 6 weeks.

Refinements such as preoperative physiotherapy and counselling, postoperative compressive stump bandages, and plaster casts with fitted prostheses to allow immediate mobilization are applicable to young amputees, particularly those undergoing amputation for malignancy, and remarkable results can be obtained in this group.

Most patients, particularly the elderly, need to use a wheelchair some or all of the time around the house, and the home must be visited at an early stage to assess the modifications necessary for a wheelchair to be used before the patient returns.

Primary haemorrhage should be a rarity if blood vessels are ligated, but venous oozing is common and should be prevented from collecting and causing swelling and stump breakdown by insertion of a vacuum drain in all amputations. Infection is common, and is usually due to Staphylococcus aureus but other organisms such as Clostridium spp. may occur, particularly in ischaemic limbs. Strict aseptic technique, avoidance of excessive dressings and use of appropriate preventative antibiotics preoperatively and for at least 5 days postoperatively is recommended. Extensive infection despite these measures usually denotes inadequate blood supply and reamputation to a higher level is needed. Small areas of skin necrosis in the suture line may remain uninfected and subsequently heal, helped by wedge excision surgery if necessary, but extensive necrosis with pain should not be left until it becomes infected: early reamputation is a better option.

Provided that skin healing is obtained most stumps can be fitted with a usable prosthesis and the site of the scar is usually immaterial. The exception is a scar that is adherent to a bony prominence since this will eventually ulcerate. Sometimes a thick skin graft or myoplastic flap can be swung in to cover the area, usually requiring the assistance of the plastic surgery department. Otherwise this problem may prevent satisfactory prosthesis use.

Some stumps with redundant tissue develop epidermoid cystic change at pressure points that often become infected. These can be excised, but unless the fundamental problem of chafing and maceration by the prosthesis is cured recurrence is inevitable. Terminal chronic oedema with verrucose hyperplasia occurs in stumps that are not supported sufficiently at the distal end.

Stump neuromas, sometimes up to golfball size, are common, but unless they lie under a point of pressure and a trigger point is demonstrable they rarely need excision. Persistent stump or phantom pain is a problem in some patients. Sometimes ischaemia is the obvious cause but in others this is not the case. Phantom pain will usually lessen with time and techniques such as ultrasound therapy and transcutaneous electrical nerve stimulation may be helpful, combined at times with psychiatric support and antidepressive therapy.

Burgess EM, Romano RL. The management of lower extremity amputees using immediate postsurgical prostheses. Clin Orthop 1968; 57: 137–56.

Burgess EM, Romano RL, Zettl JH, Schrock RD. Amputations of the leg for peripheral vascular insufficiency. J Bone Joint Surg 1971; 53A: 874–90.

Gray DWR, Ng R. Anatomical aspects of the blood supply to the skin of the posterior calf: technique of below-knee amputation. Br J Surg 1990; 77: 662–4.

Greenhalgh RM, Jamieson CW, Nicolaides A, eds. Limb salvage and amputation for vascular disease. Philadelphia: Saunders, 1988.

Kostuik JP, Gillespie R, eds. Amputation surgery and rehabilitation: the Toronto experience. New York: Churchill Livingstone, 1981.

Moore WS, Malone JM, eds. Lower extremity amputation. Philadelphia: Saunders, 1989.

Murdoch G, Donovan RG, eds. Amputation surgery and lower limb prosthetics. Oxford: Blackwell Scientific, 1988.

Robinson KP, Hoile R, Coddington T. Skew flap myoplastic below-knee amputation: a preliminary report. Br J Surg 1982; 69: 554–7.

Royal College of Surgeons. Report of a symposium on amputations and prosthetics held at the Royal College of Surgeons of England. Ann R Coll Surg 1967; 40: 203–88.

Vitali M, Robinson KP, Andrews BG, Harris EE. Amputations and prostheses. 2nd edn. London: Balliere Tindall, 1986.