Clinical manifestations

Presentation and course

	• There is perception of pain in the lost limb.
	• The pain is frequently described as burning, tingling, shooting, or a feeling of pins and needles.
	• Nonpainful sensations such as movement, touch, temperature, and pressure are more frequent, but the patient can tell the difference from phantom pain.

Loss of a part of the body is followed by several phenomena that may be both painful and nonpainful. Nonpainful sensations are more frequent than phantom limb pain and are of 3 types: (1) kinetic (perception of movement), (2) kinesthetic (willed movements), and (3) exteroceptive (touch, temperature, pressure, itch, and vibration). Patients can usually distinguish these from the painful or unpleasant sensations as intense versions of normal exteroceptive sensations. The pain is frequently described as burning, tingling, shooting, or a feeling of pins and needles. This is distinct from the local stump pain associated with wound-healing in the acute stage. Occasionally, the patient may experience a spasm of the residual muscles of a stump associated with a paroxysm of phantom pain. Back pain is also common after amputation and may be more bothersome than phantom limb pain.

Prognosis and complications

Gradual reduction of the frequency and duration of painful episodes occurs in about half of patients. Resolution may take several weeks to 2 years. Aggravation is rare, and only 3% of patients with phantom limb pain report aggravation over course of time. Chronic pain may cause depression; suicide is reported rarely.

Biological basis

Etiology and pathogenesis

	• Phantom limb pain mostly occurs following traumatic amputation, with sudden loss of a limb.
	• Pain is more likely to occur in patients who had pain in the limb prior to amputation.
	• Neuromas may form at the amputation site and give rise to chronic stump pain, which may trigger phantom pain.
	• Brain imaging studies suggests that phantom limb pain might be associated with cortical activation.
	• Interactions between pathophysiological mechanisms in the peripheral and central nervous systems may initiate and maintain chronic phantom limb pain.

The basic cause of phantom limb pain is the loss of a limb, regardless of the pathology that leads to amputation (eg, crush injury, ischemia, or malignant disease). Most of the cases are of traumatic amputation with sudden loss of limb. A gradual dissolution of the limb, such as occurs in leprosy, does not preclude the possibility of phantom limb pain. Phantom phenomena are usually seen in traumatic amputees, but not in the congenital absence of a limb that does not lead to cortical reorganization. Amputation is not essential for the occurrence of a phantom. A painful phantom can be experienced after avulsion of the brachial plexus without injury to the arm.

Postamputation pain is more likely to occur in patients who had pain in the limb prior to amputation. Neuromas may form at the amputation site and give rise to chronic stump pain, which may trigger phantom pain. Other pathological changes at the amputation site that can trigger pain include skin pathology, circulatory disturbances, infection, and bony spurs.

Atypical supernumerary phantom limb pain has been reported in 2 cases of following pontine hemorrhage (61). These patients were treated conservatively, and their symptoms took more than a month to resolve.

The role of emotional factors in the etiology of phantom limb pain remains undefined. Although emotional disturbances play a role in the occurrence and persistence of phantom pain, it remains to be determined if such disturbances are the cause of pain.

Pathophysiology. Several theories have been proposed to explain the mechanisms underlying phantom limb pain. Early explanations dealt with specificity or pattern theories of pain. Later physiological research was based on the framework provided by the gate control theory of pain and focused on identifying peripheral, spinal, and central neural mechanisms. Psychological explanations are based on psychoanalytic or personality theories of chronic pain, which imply that phantom limb pain results from preamputation psychological disturbance. Central memory theory of phantom pain implies a cortical engram for the phantom that can generate pain in the absence of stimuli from the periphery. The role of the central as well as the peripheral nervous system in phantom limb pain will be summarized here briefly.

Role of the brain. One explanation of phantom pain is based on the substantial plasticity of the somatosensory cortex after amputation. This is an adaptive and pain-preventing function and explains some of the nonpainful phantom limb phenomena. A strong relationship has been reported between the magnitude of the phantom limb pain and the amount of cortical reorganization, indicating that phantom limb pain is related to, and is a consequence of, plastic changes in the somatosensory cortex. Use of functional MRI to investigate upper limb amputees with phantom limb pain before and after training in mental imagery showed that reduction in intensity of constant pain correlated with reduction in cortical reorganization (31). A case-controlled MRI study showed increased cerebral blood volume in regions associated with emotion in the cerebral pain network of patients who had undergone unilateral arm amputation after suggesting that these changes were related to neuroplasticity associated with phantom limb pain (50).

According to another theory, dissociation in the primary motor cortex between motor representation that disappears and sensory representation that persists induces phantom limb pain (53). This is the basis of reconciliation of sensorimotor integration of the limb to relieve pain.

Central representation of phantom limb phenomenon in amputees, as studied with single photon emission computerized tomography, suggests that phantom limb pain might be associated with cortical activation involving the frontal, temporal, or parietal cortex. This may imply the possibility of the existence of an ascending polysynaptic pathway that conveys the uncomfortable phantom limb sensation to the cerebral cortex. PET studies have shown that phantom limb sensations are produced by the same central nervous processes that underlie the experience of the body when it is intact, and that the awareness of phantom limb pain is encoded in a thalamocortical network.

Both limb amputation and chronic phantom limb pain are associated with gray matter changes in the cerebral cortex. Phantom limb pain is generally considered to be triggered by loss of sensory input and caused by maladaptive cortical plasticity. According to 1 hypothesis, visual adaptation mechanisms may compensate for the lack of sensorimotor feedback and may, therefore, function as a protection mechanism against development of phantom limb pain (44).

Role of the spinal cord. Central sensitization in the spinal cord increases, and the nerves become hypersensitive due to an increase in the N-methyl-D-aspartate (NMDA) activity in the dorsal horn of the spinal cord, rendering them susceptible to action of substance P and neurokinins followed by an upregulation of the receptors in that area. This change in the spinal cord decreases the inhibitory activity from the supraspinal centers, which contributes to phantom limb pain in combination with nociceptive signals due to increased activity of NMDA in the dorsal horn.

Role of the peripheral nervous system. Following amputation, injured neurons regenerate and form neuromas. These are sensitive to mechanical and chemical stimuli. Spontaneous abnormal firing from the injured regenerating nerve sprouts has also been recorded. Stump pain may act as a generator for the phantom pain.

According to an alternative hypothesis, phantom limb pain is not due to the loss of input from the peripheral nervous system, but rather to exaggerated input, generated ectopically in axotomized primary afferent neurons in the dorsal root ganglia (55). This is the basis for suggesting dorsal root ganglia as the target for treatment.

Interactions between pathophysiological mechanisms in the peripheral and central nervous systems. These are now recognized to be responsible for the initiation and maintenance of chronic phantom limb and stump pain and include the following:

	• Peripheral damage to nociceptive fibers and dorsal root ganglion cells, which become abnormally sensitive to mechanical, thermal, and chemical stimuli.
	• Prolonged sensitization of central nociceptive “second order” neurons in the dorsal horn of the spinal cord that become hyperexcitable and start responding to innoxious stimuli.
	• Degeneration of nociceptive neurons, which may trigger the anatomical sprouting of low threshold mechanosensitive terminals to form connections with central nociceptive neurons.
	• Visual, tactile, and other sensory systems contribute to phantom limb awareness following traumatic limb amputation.

Epidemiology

• Figures for the prevalence vary, but 55% to 100% amputees experience phantom limb pain at some point in their life.

Older studies starting with that of Weir Mitchell in 1872 and continuing over the next 100 years indicate that 80% to 100% of amputees experience phantom limb pain at some point in their lives following amputation. Studies within the last 25 years report phantom limb pain in 55% to 88% of amputees. In a survey of 11,000 out of 450,000 U.S. war veteran amputees, 80% reported having significant phantom pain (51). In a questionnaire survey of British war veteran amputees, 55% reported phantom limb pain and 56% stump pain (57).

Of the 200,000 patients who undergo surgical amputations in the United States each year, 70% experience phantom limb pain after the procedure, and 50% still experience phantom pain 5 years after surgery (06). Results of a United States national survey indicate that phantom pain was reported by 79.9% of amputees and that a large proportion experienced severe pain (16). A 6-month follow-up study after limb amputation reported that phantom limb pain occurred in 78.8% of the survivors and 51.2% had stump pain (46). Phantom limb pain after amputation for malignant disease in children and young adults is common but short lived as the prevalence drops to 10% after 1 year (10). In a cross-sectional study of United States military service members after traumatic limb amputation, 58% experienced phantom pain, and it was frequently associated with neuroma formation (09). In this study, there was a significant association of chronic postamputation pain with symptoms of catastrophizing and post-traumatic stress disorder (PTSD).

A study of demographics of amputees for prevalence of phantom limb pain in New Zealand and Cambodia showed no differences in prevalence despite different environments, indicating that severity of neurologic impact of amputation may override all other risk factors (11).

Prevention

	• Adequate management of preamputation pain
	• Primary targeted muscle reinnervation at time of amputation
	• Adequate treatment of stump neuromas
	• A prosthesis covered with e-dermis that perceives touch and pain

Findings of a retrospective study on amputees in China shows that preamputation pain and the postoperative deafferentation are the risk factors of phantom limb pain and form the basis of recommendations for adequate relief of preamputation pain and avoidance of sudden postamputation deafferentation (60). Preamputation pain can be reduced by epidural or intrathecal infusion of local anesthetics and opioids. None of the patients treated with a combination of local anesthetics and low-dose morphine have been reported to develop phantom limb pain. Infusion through the catheter placed during surgery can be continued for 2 to 3 days postoperatively, and then discontinued in favor of a pharmacologic regimen including opioids along with a tricyclic antidepressant and an alpha-2 adrenergic agonist. The provision of a pain-free perioperative interval using regional anesthetic techniques is likely to reduce the incidence of phantom limb pain. Use of a prolonged postoperative perineural infusion of the local anesthetic ropivacaine 0.5% is reported to be an effective method for preventing the development of phantom limb pain after lower extremity amputation (07). Neuroimaging studies show that brain changes in phantom limb pain, which are based on processes of neuronal plasticity, can be partially reversed by analgesic interventions. A retrospective study concluded that type of anesthesia and the degree of postoperative pain are not related to the development of phantom limb pain, but inadequate management of preoperative pain has an impact on the development of phantom limb pain (39).

Primary targeted muscle reinnervation, ie, rerouting of nerves cut during amputation to surrounding muscle, significantly prevents the occurrence of phantom limb and residual limb pain (08). Originally developed to enable amputees better control of upper limb prosthetics, it has now been used to treat symptomatic neuromas or phantom limb pain in below-the-knee amputees.

A prosthesis covered with e-dermis that perceives touch and pain has been devised to replace the normal sensation of the amputated limb and may prevent the development of phantom limb pain (41).

Differential diagnosis

Confusing conditions

Although the diagnosis of phantom limb pain appears to be obvious with the history of amputation, other causes should be excluded because not all amputees have phantom pain. Causes that need to be differentiated are:

Psychogenic phantom limb pain. This is extremely rare and is characterized by manifestations of mental and emotional problems. The patient usually misinterprets ordinary nonpainful phantom sensations.

Stump pain. A prosthesis covered with e-dermis that perceives touch and pain may appear due to local complications and should be distinguished from phantom pain, although it is recognized as a trigger of phantom pain. Stump pain can be diminished by both intravenous morphine and lidocaine, whereas phantom pain is diminished only by morphine, suggesting that the mechanisms and pharmacological sensitivity of stump and phantom pains are different. This information can be used for differentiating the 2 conditions.

Risk factors for phantom limb pain are related to maladaptive plasticity, whereas risk factors for residual limb pain have components leading to neuropathic pain, such as the amount of neural lesion and previous history of chronic pain (38).

Foreign body in the stump. A foreign body introduced during trauma or overlooked during surgery may be a cause for pain. It can be detected by radiological examination.

Lesions of the brachial plexus or spinal nerve roots. These may occur in previously pain-free amputees and can produce pain that may be localized to the territory of a nerve.

Lesions of the central nervous system. Stroke or spinal cord injury in a patient with amputation may produce central neuropathic pain in the phantom limb, but the characteristic features of this pain enable differentiation.

Associated or underlying disorders

1. Phantom limb pain is a sequel of limb amputation, which may be performed for the following:

2. Severe limb trauma, eg, severe crush injury where a limb can no longer be salvaged

3. Malignant tumors such as osteosarcoma

4. Gangrene due to ischemic peripheral vascular disease

5. Severe infections that do not respond to treatment, eg, necrotizing fasciitis

6. Gangrene and osteomyelitis in diabetic patients

Diagnostic workup

	• History of primary disease and description of pain with assessment by a pain scale
	• General physical examination and check of stump for neuroma
	• MRI scan of stump

The patients should be evaluated prior to amputation with particular emphasis on assessment of pain and psychological status.

Postamputation assessment should include a detailed history and description of the pain. In addition to a neurologic examination, a general examination should evaluate the primary disease. The stump should be checked carefully to detect a neuroma by Tinel sign. Examination of the gait and the back should be performed particularly if the patient has back pain as well.

Laboratory examination should include an MRI scan of the stump that may detect a neuroma.

Management

	• Several treatments, pharmacological as well as nonpharmacological, have been used for the management of phantom limb pain.
	• Several neurosurgical procedures have been used.
	• Neuromodulation including deep brain and spinal cord stimulation

Various procedures that have been used for the management of phantom limb pain are listed in Table 1.

Table 1. Methods of Treating Phantom Limb Pain

Pharmacotherapy
	• Nonsteroidal antiinflammatory drugs/acetaminophen • Opioid analgesics • Antiepileptic medications
		- Carbamazepine - Clonazepam - Gabapentin - Topiramate
	• Benzodiazepines
		-Midazolam
	• Antidepressants
		- Amitriptyline
		- Fluoxetine
	• Topical capsaicin • Botulin type A injection • NMDA antagonists: ketamine, memantine • Calcitonin • Tizanidine
Anesthetic procedures
	• Local anesthetic injection into stump neuroma • Preoperative epidural anesthesia continued postoperatively • Continuous postoperative regional anesthesia or patient-controlled intravenous analgesia • Intrathecal fentanyl
Nonpharmacological methods
	• Acupuncture • Therapeutic touch • Electroconvulsive therapy • Stump percussion • Psychological approaches Biofeedback Virtual and augmented reality: mirror therapy, graded motor imagery • Proper fitting of prosthesis with sensory function
Neurosurgical procedures
	• Stump revision • Excision of stump neuroma • Targeted muscle reinnervation • Sympathetic block/sympathectomy • Spinothalamic tractotomy • Anterior cingulotomy • Procedures that are ineffective, performed rarely, or no longer recommended
		- Posterior rhizotomy - Epidural spinal cord stimulation - Thalamotomy - Postcentral gyrectomy - Frontal leucotomy
Neuromodulation
Anodal M1 transcranial direct current stimulation (tDCS)
	• Deep brain stimulation • Dorsal root ganglion stimulation • Spinal cord stimulation • Transcutaneous electrical nerve stimulation • Transcranial magnetic stimulation
Multidisciplinary approaches

Pharmacotherapy. Various analgesic medications have been used for the management of phantom limb pain. There is no consensus on guidelines for the pharmacological management of phantom limb pain, but tricyclic antidepressants, antiepileptic drugs such as gabapentin, opioids, local anesthetics, and N-methyl-D-aspartate (NMDA) receptor antagonists such as ketamine have been recommended for the treatment of phantom limb pain (17). In the acute state, the infusion of calcitonin and oral opioid analgesics has proven to be helpful, whereas established phantom limb pain may respond to antidepressants and anticonvulsants. A Cochrane review of clinical trials reported that memantine and amitriptyline were ineffective for treatment of phantom limb pain (03). An update of Cochrane review of phantom limb pain trials concluded that short- and long-term effectiveness of botulinum toxin A, opioids, NMDA receptor antagonists, anticonvulsants, antidepressants, calcitonins, and local anaesthetics for clinically relevant outcomes including pain, function, mood, sleep, quality of life, and treatment satisfaction remain unclear (02). Morphine, gabapentin, and ketamine have demonstrated favorable short-term analgesic efficacy compared with placebo. Larger and more rigorous randomized controlled trials were suggested for making stronger recommendations about which medications would be useful for clinical practice, but such trials had not been reported as of the end of 2018. A review did not reach definite conclusions about the pharmacological management of phantom limb pain based on the available evidence, but the N-methyl-D-aspartate receptor antagonists, such as memantine, demonstrated consistent positive results (20).

Opioids. Sequential pharmacotherapy may include opioids, but long term use of morphine should be avoided because of risk of addiction. Opioids show efficacy in the treatment of phantom limb pain and may potentially influence cortical reorganization.

Patients refractory to other treatments have been shown to respond to oral methadone in anecdotal reports; controlled clinical trials would be appropriate to verify this observation.

Antiepileptics. Various antiepileptic medications have been used. These are used mainly for peripheral neuropathic pain and are seldom tried for phantom limb pain. Anecdotal reports indicate some benefit from use of carbamazepine and clonazepam. Gabapentin is a useful adjunct to pain management in children and young adults with phantom limb pain. Topiramate has also been reported to decrease phantom limb pain. In a study of phantom limb pain following amputation in U.S. war veterans, the most common medical treatment was gabapentin, although some patients reported greater pain relief from self-treatment methods such as distraction and relaxation techniques (27).

Botulinum toxin type A. Phantom limb pain has been successfully treated by injection of botulinum toxin type A in the amputation stump (25). A randomized, double-blinded pilot clinical trial of botulinum toxin type A versus lidocaine/methylprednisolone resulted in immediate decrease of residual limb pain but no relief of phantom limb pain with either treatment (59).

Midazolam. Midazolam potentiates gamma aminobutyric acid and enhances the inhibitory action of glycine receptors at spinal neurons. Relief of acute exacerbation of phantom pain following intravenous midazolam has been reported but it is of short duration.

Ketamine. This is a NMDA receptor antagonist. Ketamine, administered intravenously to patients with established stump and phantom limb pain, has been reported to result in adequate control of pain.

Capsaicin. This is applied locally for relief of stump pain.

Calcitonin. Intravenous salmon calcitonin 200 IU may be an effective treatment for phantom limb pain.

Tizanidine. Tizanidine is an agonist at alpha2-agonist receptor sites and is approved for the management of spasticity associated with multiple sclerosis and spinal cord injury. Its use as an analgesic is an off-label indication. No reliable report has shown efficacy of tizanidine for phantom limb pain.

Dextromethorphan. This oral medication has been reported to be effective for relief of phantom pain in cancer amputees but efficacy has not been confirmed in controlled studies.

Memantine. Memantine, an orally administrable N-methyl-D-aspartate receptor antagonist, does not provide significant clinical benefit in patients with chronic phantom limb pain.

Anesthetic procedures. Optimized epidural analgesia or intravenous patient-controlled analgesia, starting 48 hours prior to amputation and continuing for 48 hours postoperatively, has been shown to decrease phantom limb pain at 6 months (26).

Nonpharmacological methods.

Acupuncture. The use of acupuncture for relief of pain is described in another MedLink Neurology article; this method has been applied to relieve phantom limb pain as well. A case report describes complete relief of phantom limb pain with sessions of acupuncture in a primary care setting (13). Although there is anecdotal evidence for the success of acupuncture in treating phantom limb pain, controlled clinical trials are needed to provide more definitive evidence for efficacy of acupuncture.

Physical therapy. A pilot study on traumatic amputees has shown that phantom limb exercises plus prosthetic training are more effective in reducing phantom limb pain than routine prosthetic training plus a general exercise program (54).

Mirror feedback therapy. In this approach, the patients are told to use the mirror in a way that the image of the mirrored healthy limb seems to appear in the place of the missing extremity. It is based on the influence of sensory and motor training effects on the pain experience with cortical organization. Preoperative mirror therapy may reduce the incidence of phantom limb pain. The PACT trial was a multicenter randomized controlled trial in which participants with phantom limb pain following lower limb amputation were randomly assigned to care as usual, traditional mirror therapy, or mirror therapy supported by telerehabilitation. Results showed that traditional mirror therapy over 4 weeks was not more effective than sensorimotor exercises without a mirror in reducing phantom limb pain, although significant effects were suggested in some subgroups (48).

A single-blinded, randomized, controlled trial showed that graded motor imagery (GMI) is better than routine physiotherapy and may be a viable treatment for reducing phantom limb pain (30).

Virtual reality. Virtual reality is a sophisticated replication of the mirror box approach. In this method, motion data are captured directly from a patient's stump and transformed into goal directed, virtual action. A variant of this is a method where the virtual limb responds directly to myoelectric activity at the stump, and the illusion of a restored limb is enhanced by augmented reality (40). It was effective for relieving pain in a patient who had been refractive to all other therapies (40).

Mental imagery, which involves limb laterality recognition, imagined movements, and mirror movements, reduces pain and disability in patients with phantom limb pain, but the mechanism of the effect is not clear. A “virtual” limb, the mirror image of an intact limb, may inhibit the cortical reorganization of motor and somatosensory cortex following loss of limb. A virtual reality study in which participants completed pre- and posttreatment measures to evaluate changes in phantom limb pain and phantom sensations and rate helpfulness, realism, immersion, adverse experiences, and treatment satisfaction showed significant reductions in pain intensity and phantom sensations (49). Rehabilitation using a virtual reality approach may be particularly effective for phantom limb pain associated with distorted phantom limb movement and body representations, eg, clamping and gnawing, compared with typical neuropathic sensations, such as shooting, burning, and dysesthesia (42).

Local anesthetic injection. Contralateral injections of bupivacaine in myofascial hyperalgesic areas has been shown to attenuate phantom pain in the affected limb in a double-blinded crossover study (12).

Far infrared ray therapy. The underlying mechanism of effectiveness of such far infrared therapy treatment has not been elucidated. One patient with severe phantom limb pain has been successfully treated by local application of far infrared therapy to the site of amputation (23). In 3 other patients, far infrared therapy was applied directly to the phantom limb site instead of the stump and was found to be effective (22).

Biofeedback. Two types of feedback have been used for the management of phantom limb pain: temperature feedback and electromyographic feedback. Although temperature feedback has been used for patients who complain of burning pain, there are no guidelines to match a specific biofeedback technique with the type of pain. Biofeedback is not used often for phantom limb pain and there are no controlled studies to evaluate its efficacy.

Psychological approaches. Although phantom limb pain is not considered to be a psychiatric disorder, persistent severe pain can have psychological sequelae. Psychological interventions can lessen pain intensity and help the patient cope better with pain. Coping strategies have been evaluated using semistructured interviews and analysis of patients' drawings of their body images and have shown a significant association between coping strategies and pain. Patients who better cope with the loss suffer less from phantom limb pain.

There is growing evidence for the effectiveness of mind-body therapies for phantom limb pain even though it is empirical; in addition to mirror visual therapy, various approaches include targeting of cortical reorganization, autonomic nervous system deregulation, stress management, and coping ability (36).

Proper choice of prosthesis. A poorly fitting prosthesis may exaggerate the phantom sensation due to local trauma and can be corrected by refitting. A prosthesis with feedback function is a promising therapeutic tool to reduce phantom limb pain (14). Use of myoelectric prosthesis in upper-extremity amputees is associated with reduced phantom limb pain and even reduced cortical reorganization.

Clinical trials of nonpharmacological methods. A systematic review of randomized clinical trials revealed limited evidence supporting the effectiveness of both hypnosis in the short term and graded motor imagery in the short-to-medium term, as well as limited evidence that a single session of mirror therapy has no immediate effect on phantom limb pain (05).

Neurosurgical procedures. Several classical neurosurgical procedures have been applied for the treatment of phantom limb pain (58). Some of the older procedures have been abandoned, and newer techniques of functional neurosurgery have been applied for the relief of this problem. The general trend has been away from ablative procedures and more towards stimulation techniques. Brief remarks on some of these techniques are as follows:

Excision of stump neuroma. This is usually the first surgical procedure. The stump is often revised, and the neuroma is excised. Several techniques have been used to prevent neuroma reformation. Surgery on the stump is of limited value, and there is no indication for a more proximal amputation only because of phantom pain.

Coblation of sciatic and femoral nerves. Coblation is controlled ablation of soft tissues by a nonheat driven process using bipolar radiofrequency. Coblation of femoral and sciatic nerves has been reported to achieve 80% relief of both stump and phantom limb pain in a case and may be a more effective approach than excision of neuroma (62). This requires further investigation.

Image-guided percutaneous nerve cryoablation is feasible and safe and may be an effective therapeutic option for patients with phantom limb pain (45).

Sympathectomy. Operations on the sympathetic system have been performed in the past, based on the assumption that the pain is sometimes accompanied by autonomic manifestations and that blocking autonomic ganglia may relieve pain. This procedure is performed rarely these days.

Spinothalamic tractotomy. This was 1 of the earlier neurosurgical procedures to be performed for phantom limb pain and did provide relief for a reasonable percentage of such patients in earlier series. In long-term follow-up, however, the pain frequently recurs. This procedure is now used only rarely.

Anterior cingulotomy. Stereotactic bilateral cingulotomy has been performed for several decades for psychiatric disorders, addiction, and chronic pain, including phantom limb pain. The procedure is safe, but seizures have occurred as a complication, which can be controlled by antiepileptic medications.

Targeted muscle reinnervation. This procedure transfers amputated nerves to nearby motor nerves for prevention of neuroma. A review of retrospective studies showed near-complete resolution of neuroma pain after treatment as well as complete prevention of chronic pain if targeted muscle reinnervation is used as a prophylactic measure during the amputation (32). In a randomized controlled trial, targeted muscle reinnervation was shown to reduce postamputation pain relative to conventional neuroma excision and muscle burying. Patients who could not be randomized also underwent targeted muscle reinnervation, and a prospective study after 1 year demonstrated that improvement in residual limb and phantom limb pain parameters in major limb amputees leading to the recommendation should be considered as a first-line surgical treatment option for chronic amputation-related pain in patients with major limb amputations (35).

Neuromodulation. Various techniques of neuromodulation are used, some of which are noninvasive, whereas others are minimally invasive.

Anodal M1 transcranial direct current stimulation (tDCS). A systematic review of randomized, controlled trials and metaanalysis of research studies of neuromodulation for phantom limb pain have shown that excitatory M1 stimulation-specifically, anodal M1 tDCS, has a significant short-term effect in reducing pain scale scores (43).

Deep brain stimulation. This procedure has been performed for the past 30 years and has been useful in relieving chronic pain. Phantom limb pain has also been satisfactorily treated with deep brain stimulation.

Dorsal root ganglion stimulation. This approach can be tried in patients that do not respond to motor cortex or spinal cord stimulation and is based on the concept that dorsal root ganglion may be the site of generation and/or maintenance of phantom limb pain. In an open trial on 8 patients, 1 had complete relief and 3 had only partial relief of phantom pain (15). In cases where stimulation is effective, a stimulator system can be implanted.

Motor cortex stimulation. Motor cortex stimulation after electrode placement has been shown to be effective for treating peripheral deafferentation pain of the phantom limb.

Spinal cord stimulation. This has been tried in patients where deep brain stimulation is not successful in relieving phantom limb pain. A systematic review found some evidence of efficacy of spinal cord stimulation for phantom limb pain, but due to the relatively small number of patients in each study, further research is needed to demonstrate these benefits (01).

Transcutaneous electrical nerve stimulation. This is still used frequently, even though the relief of pain is temporary and incomplete. The advantages of this method are the lack of complications and the possibility of repetition as often as desired. However, because of a lack of adequately controlled trials with rigorous methodology, transcutaneous electrical nerve stimulation cannot be adequately assessed as a treatment for phantom limb pain (37).

Transcranial magnetic stimulation. This is emerging as a possible treatment for phantom limb pain as evidenced by several case reports of its successful use. Transient analgesic effect induced by repetitive transcranial magnetic stimulation may be due to a temporary interference with the cerebral representation of the differentiated limb. In one case, following a traumatic limb amputation, repetitive transcranial magnetic stimulation for phantom limb pain, was applied over the sensory cortex corresponding to the area of amputation and resulted in significant relief of pain (19). Another patient with severe phantom limb pain has been reported to have been successfully treated with a low frequency repetitive transcranial magnetic stimulation over supplementary motor complex in the cerebral hemisphere contralateral to the involved limb (29).

Outcomes

These are discussed along with some of the methods of treatment.

Special considerations

Anesthesia

Phantom limb pain recurrence or exacerbation may be induced by regional anesthesia, including spinal anesthesia, epidural anesthesia, and peripheral nerve block. This can be successfully managed by midazolam (52).

Phantom limb pain may be prevented by proper anesthetic management of patients with preamputation limb pain.