Phantom limb pain (PLP) and residual limb pain (RLP) are pervasive problems following limb amputation. Inherent to limb amputation is the transection of peripheral nerve fibers, typically followed by resection or ligation. This injury results in inherent changes to the structure of the individual neurons. Specifically, the axon swells at the transected area creating an “end bulb,” which quickly gives rise to terminal sprouting. The resultant knot of the axon terminal and newly formed sprouts is referred to as a neuroma.1 In addition, those sprouts that remain free of the neuroma may spread into nearby tissue, increasing their sensitivity to otherwise benign stimuli.
In addition to these mechanical changes in the peripheral nervous system, changes in the central nervous system also occur. Following amputation, the lack of sensory input to segments of the sensory cortex allow them to become responsive to adjacent sensory input. This maladaptive plasticity of the sensory cortex has been suggested as the primary source of phantom pain.1
It is often difficult to clearly segregate RLP and PLP as the experience of either generally makes the individual more sensitive to the other. These pain experiences can be tremendously disruptive throughout the process of rehabilitation, compromising wellness during waking hours and sleep quality during resting hours. The management of amputation-related pain is often pursued through pharmaceutical approaches.2 These may include both opioid and non-opioid solutions such as over-the-counter pain relievers (Tylenol, Motrin, Aleve), antidepressants (nortiptyline), and anticonvulsants (Gabapentine, Neurontin, Lyrica).2
However, in view of the inconsistent responses to pharmaceutical approaches and the range of associated undesirable side effects, a number of non-pharmaceutical solutions to PLP have been attempted. This article provides an overview of the established and developing approaches to mitigating limb pain following amputation.
Mirror therapy was introduced by Ramachandran et al. over 20 years ago in consideration of the theory of learned paralysis. According to this theory, the brain continues to send motor signals to the amputated limb but doesn’t receive any sensory signals in return to confirm the movement. This creates an illusion of paralysis that becomes painful for some individuals. To address this illusory paralysis, mirror therapy attempts to synchronize the outgoing motor commands with visual feedback suggesting the desired movement, created by observing the sound side limb through a mirror. This reflection of the voluntary movement from the sound side limb creates a counter illusion of pain-free movement of the phantom limb.
From its inception, mirror therapy has proven a successful means of reducing PLP for some individuals. However, as one of the more studied means of reducing PLP, a recent systematic review of some 20 publications reported the efficacy of mirror therapy on PLP, but only with low levels of evidence.3
Among the shortcomings of mirror therapy is that it is minimally immersive, challenging the abilities of many individuals to perceive the reflected movements as those of the phantom. In addition, in many cases the aberrant size and position of the phantom limb fail to match those of the reflection. To address these deficits, variations of mirror therapy have been developed using virtual reality and augmented reality, which have the added benefits of increased immersion and customized images. While the supportive literature is smaller than that associated with mirror therapy, the results were somewhat similar. More specifically, existing studies suggest improvements in PLP with virtual and augmented reality using both quantitative and qualitative measures, but these were all lower levels of evidence, confined to case studies and case series.4
In contrast to the visual perceptions associated with mirror therapy and its augmented variations, transcutaneous electrical nerve stimulation (TENS) represents an alternate noninvasive approach to pain management. TENS represents a minimally invasive form of neuromodulation in which electrical stimulation is used to stimulate sensory nerves to modulate sensory perceptions. As implied by the name, TENS is applied at the surface of the skin and directs electrical current to subdermal sensory nerves. The resultant experience replaces the perception of pain with the sensation of paresthesia and has been used to treat a wide range of peripheral pain syndromes with mixed observed efficacy. As with the other non-invasive approaches presented thus far, while there is a small body of evidence suggesting the efficacy of TENS in managing PLP, systematic review has reported mixed results with no high-quality evidence to support or refute its efficacy.5
A specialized approach to transcutaneous stimulation has been explored in the form of noninvasive closed-loop neuroprostheses. Explored in a few studies thus far, this technique uses transcutaneous stimulation on the residual limb of the user while he or she trains with a myoelectric prosthesis. More specifically, sensors are placed in the prosthetic hand to measure the resultant prosthetic grip strength. These signals are then sent to a microcontroller that converts the pressure data into electrical stimulation patterns that are applied transcutaneously using eight small electrodes attached to a cuff worn over the patient’s forearm. With training, patients learned to interpret these signals to regulate their functional grip strength.6 While these studies have tended to focus on the functional benefits of the intervention, they have also reported reductions in PLP, with one recent trial of eight users of transradial prostheses reporting a reduction in PLP severity of nearly 50 percent with the use of transcutaneous sensory restoration.6
In a somewhat related approach, repetitive transcranial magnetic stimulation (rTMS) has also been used as a noninvasive treatment for PLP. In this approach, instead of transmitting transcutaneous electrical stimulation to peripheral nerves, magnetic pulses are administered to specific areas of the patient’s head to deliver focused stimuli to the brain. While used predominantly to treat depression, additional work has suggested that rTMS can also provide temporary relief of chronic neuropathic pain.1 A relatively recent randomized, double-blinded trial of 54 individuals with lower-limb amputations showed that the application of rTMS to the primary motor cortex contralateral to a subject’s amputation over two weeks reduced the intensity of his or her PLP for up to 15 days after treatement.7 These findings built upon an earlier trial of 27 individuals with unilateral amputations and chronic PLP who received rTMS to the ipsilateral motor cortex on five consecutive days and experienced significant relief of their PLP for at least two months after treatment.8 Further supplemented by a number of case studies, the noninvasive nature of rTMS shows promise as a modality to remediate PLP. However, long-term efficacy has yet to be evaluated.1
Invasive Electrical Stimulation
Transitioning to more invasive pain management approaches, there are a range of invasive electrical stimulation approaches targeting different anatomic locations of the central and peripheral nervous system. Of these, electrical spinal cord stimulation (SCS) has the longest track record of use with over five decades of application for leg and back pain.1 In its classic application, electrodes are implanted along the midline of the thoracic spinal cord, delivering stimulation to the dorsal columns.1 Stimulation of these structures creates a paresthetic sensation that reduces the pain that would otherwise be experienced. Aggregated data suggest that SCS systems are implanted in over 50,000 people each year with a strong safety record and a low rate of complications.1 The most common complication, occurring in 15-20 percent of SCS implantations, is the migration of the electrode leads, relocating the area of paresthetic coverage, thus mitigating the associated pain relief and requiring surgical revision.1
With respect to PLP and RLP, a relevant limitation of SCS is the challenge of generating paresthetic coverage over the distal aspect of the limbs. A small number of studies have demonstrated a level of effectiveness with SCS in managing post-amputation pain, though these studies generally did not differentiate between PLP and RLP.1 Available publications suggest a success rate of 50-60 percent. Ultimately, the limitations of SCS in the management of PLP are such that the approach has not received widespread clinical adoption.1
Using similar electrical stimulation but focusing that stimulation to a different segment of neural anatomy, dorsal root ganglion (DRG) stimulation represents a newer approach to electrical stimulation that may carry better promise in the mitigation of PLP. Emerging evidence suggests that the location of the electrical leads in DRG are positioned in a more stable interface and thus are less prone to migration.1 While still viewed as an emerging modality, a single study of eight participants with limb amputation observed some initial success with all subjects electing to proceed from a percutaneous application to a fully implanted system, and several participants reporting a decrease in PLP over the two-year follow-up period.9
Moving from the spinal cord to the DRG, it is unsurprising to observe the application of electrical stimulation to the associated peripheral nerves with peripheral nerve stimulation (PNS). Similar to TENS, the surgical implantation of the electrodes allows a more accurate targeting of specific nerves. In addition, PNS permits access to deeper nerves that are inaccessible via TENS as well as access to proximal nerves in the residual limb that previously innervated the amputated limb.1 Further, the permanent implantation results in a more stable neural interface, improving the consistency of daily stimuli and response.1 These benefits are offset by the obvious invasiveness of the surgical implantation, which require episodic revisions to replace batteries.
While the evidence for PNS in the treatment of PLP is still emerging, pilot studies have shown some promise. In the most recent of these, 14 of 16 subjects who completed in-clinic testing responded to stimulation, reporting at least 75 percent parasthetic coverage and clinically significant pain relief. Nine of the 14 subjects successfully completed the two-week home trial that followed, reporting reduction in their mean daily worst PLP, average RLP and PLP, RLP and PLP interference, and Pain Disability Index.10
Invasive Neural Sensory Restoration
The neuroprosthetic construct of mitigating PLP through the restoration of sensory input was described earlier in some noninvasive modalities. This approach has received additional exploration in surgically invasive approaches of true neural integration. The enhanced fidelity of implanted leads with neural sensory stimulation has been associated with substantial improvements to motor control as well as embodiment of the prosthesis.1 In addition, this approach appears to reduce PLP via cortical reorganization in the somatosensory cortex.1 This has been attempted through a number of different electrode types, each with different strengths and weakness. These include intrafascicular, epineural, and regenerative electrodes, each of which will be discussed in greater detail.
As suggested in the name, intrafascicular electrodes are designed to penetrate through the epineurium of peripheral nerves to achieve intimate contact with both sensory and motor neurons. Early reports on the sensory restoration associated with such electrodes have repeated a decrease of general pain experiences with anecdotal reports of decreased PLP.1 Such beneficial observations are tempered by concerns that the insertion of such electrodes causes mechanical damage to the target neurons followed by glial scarring.1 Further, the stability of the signals received by such electrodes over time has been questioned.1
Epineural electrodes provide a less invasive approach to peripheral nerve interfacing. Rather than piercing the epineurium, these electrodes sit outside the epineurium, creating a more stable and potentially less destructive interface, but with less selective stimulation.1 As with intrafascicular electrodes, epineural electrodes have been associated with reductions in PLP as they have enabled a progressive embodiment and beneficial repositioning of the phantom limb, with some subjects reporting an eventual elimination of PLP.
Regenerative electrodes represent a third means of sensory restoration still early in development. In this approach, the hollow, ring-like electrodes are implanted at the terminal end of a transected nerve, permitting the nerve to grow through the cavity of the ring, resulting in intimate contact with the electrical contacts. While this approach may yield both high selectivity and long-term stability, available data is confined to animal studies. However, these have shown promise, with transected nerves regenerating as hypothesized through the ringed electrodes.1
Invasive Nerve Reinnervation
As with the aforementioned various approaches to sensory restoration, the impacts of targeted muscle reinnervation (TMR) on post-amputation pain experiences were secondary to other primary aims. Specifically, TMR was developed to increase the number of motor inputs available for controlling an externally powered prosthesis. Briefly, in TMR the severed nerves associated with an amputation are individually dissected and then coapted into a newly divided motor nerve in a targeted muscle belly.11 While the impact of this procedure on pain was initially uncertain, retrospective evaluation has suggested a reduced pain experiences with regard to RLP and PLP.11
With the functional benefits of TMR now well established, recent literature has begun to canonize the impacts of TMR on PLP and RLP. For example, a recent randomized controlled clinical trial compared the efficacy of two approaches to neuroma management in 28 patients with chronic PLP. Half of the subjects underwent standard treatment, in which the neuroma was excised along the length of the nerve until healthy nerve fascia was observed and then tunneled into the deep aspect of a nearby muscle without tension.11 The remaining subjects underwent TMR as previously described.
One-year follow-up data demonstrated a clear improvement in the mitigation of PLP and RLP with TMR with no comparable improvements in the standard of care subjects.11 More specifically, using a ten-point numerical rating scale, PLP scores decreased an average of 3.2 points with TMR while the standard of care subjects averaged a 0.2 point increase. Similarly, RLP scores decreased an average of 2.9 points compared to a decrease of only 0.9 points for those subjects receiving the standard neuroma resection procedure.11
Given the initially unanticipated benefits of TMR on post-amputation pain, several surgeons have begun to see pain prevention as a primary benefit of TMR, even for those subjects that may not benefit from the enhanced control mechanisms enabled by the procedure. Exploring the position that TMR could be beneficial when performed at the time of the primary amputation rather than as a post-amputation procedure to reduce existing limb pain, a recent publication has retrospectively analyzed the pain experience of a large cohort of patients managed either with or without TMR as part of their primary amputation.12 Comparing 51 subjects from multiple participating sites who had received TMR at the time of primary amputation against 438 subjects who did not, the former reported less PLP and RLP than the untreated controls.12 Specifically, while the untreated control group reported average PLP scores and RLP scores of 5 and 4 out of 10, respectively, those who underwent primary TMR reported an average score of 1 for both pain types.12 These findings were echoed across several other measures of pain and pain interferences.12 Collectively, the findings caused the authors of the trial to conclude simply that “preemptive surgical intervention of amputated nerves with TMR at the time of limb loss should be strongly considered to reduced pathologic phantom limb pain and symptomatic neuroma-related residual limb pain.”12
A more recent alternative approach to nerve reintegration has been described as regenerative peripheral nerve interfaces (RPNIs). Conceptually similar to TMR, RPNIs permit the reintegration of more nerves as these are integrated into a number of small autologous muscle grafts, typically harvested from the vastus lateralis, rather than relying on the limited number of muscle bellies located in the immediate area of the transected peripheral nerves.13 Once integrated with the new peripheral nerves, the grafts act as bioamplifiers for the descending motor commands of the residual nerves. This technique was recently performed on two patients with upper-limb amputations and 14 patients with lower-limb amputations, who reported an average reduction of 53 percent in PLP and 71 percent in RLP as much as 15 months post-grafting.13
A range of creative solutions have been put forward as possible means of mitigating RLP and PLP following amputation. These have included therapeutic illusions created using simple mirror boxes and more nuanced virtual reality systems; non-invasive electrical (TENS) and magnetic (rTMS) stimulations; more invasive stimulation of the spinal cord, dorsal root ganglia, and the peripheral nerves; both noninvasive and surgically invasive means of sensory restoration, including a range of implanted microelectrodes; and surgical reinnervation of peripheral nerves to prevent the formation of neuromas in the first place. Given the pervasive nature of post-amputation pain experiences, clinical prosthetists should be aware of the broad range of solutions that have been put forward to address this common challenge.
Phil Stevens, MEd, CPO, FAAOP, is in clinical practice with Hanger Clinic, Salt Lake City. He can be contacted at [email protected].
Petersen, B. A., A. C. Nanivadekar, S. Chandrasekaran, and L. E. Fisher. 2019. Phantom limb pain: peripheral neuromodulatory and neuroprosthetic approaches to treatment. Muscle & Nerve 154-67.