How Does Science Explain Phantom Pain?

Phantom limb pain ─ the pain felt in the missing part of an amputated limb — is a common condition affecting approximately 64% of people with limb amputations. Unfortunately, phantom limb pain is difficult to treat, primarily because its underlying mechanisms are poorly understood.

While some studies suggest it is triggered by peripheral mechanisms, others suggest it is driven by central mechanisms in the brain. In this article, we will discuss several mechanisms proposed to underlie phantom limb pain at each level and how they are used during clinical reasoning to advise on treatment choice.

Peripheral mechanisms of phantom limb pain

The amputation of a limb involves cutting through skin, muscles, bones, and nerves. The muscles and skin heal and mold into a conically-shaped residual limb over time. However, the healing process of a nerve is complex. Following an amputation, the nerves develop a neuroma ─ an abnormal growth of cells at the end of the severed nerve (A in Figure 1). The neuroma develops as an attempt of a nerve to heal itself. However, this abnormal nerve healing can cause further problems. For example, electrophysiological studies have shown spontaneous and exaggerated electrical impulses generated by the neuroma. The brain can interpret these potentially dangerous impulses as pain. Furthermore, studies have shown a positive relationship between the strength of these electrical impulses and phantom limb pain. This means that the stronger the electrical impulses, the more severe the phantom limb pain.

Perhaps, one would expect to feel pain only at the site of the neuroma. However, nerve pain is also characteristically felt in the far end of the limb that is innervated by the affected nerve. For amputees, this can present as phantom limb pain — experienced below the point of amputation.

Other evidence supporting the peripheral origin of phantom limb pain is pain triggered by applying pressure to the neuroma. This type of pain can be reproduced clinically by tapping the neuroma, and patients often experience it when using ill-fitting prosthetic limbs, causing irritation and pressure to the residual limb when walking.

It is worth noting that the onset of phantom limb pain can present as early as one day after amputation — long before a neuroma has developed. Considering this, Vaso and colleagues provided compelling evidence that phantom limb pain is also generated by spontaneous and ongoing electrical nerve impulses, specifically at the dorsal root ganglion (a cluster of sensory neurons relaying sensory information from the periphery to the spinal cord) of the injured nerve (B in Figure 1). It is on this basis regional anesthetics are thought to be most effective in reducing phantom limb pain, particularly when brain mechanisms thought to maintain pain are not predominant.

Brain mechanisms of phantom limb pain

The evidence from functional magnetic resonance imaging (fMRI) studies revealed functional changes in the areas of the brain (C in Figure 1) responsible for generating sensations and movement, in that the area that previously innervated the amputated limb shrinks and gets invaded by neighboring regions. The early studies showed a positive relationship between the extent of this invasion and phantom limb pain severity. However, it has not been confirmed whether the reorganization of the brain areas directly causes phantom limb pain because not all amputees with confirmed brain reorganization have phantom limb pain, and not all amputees with phantom limb pain have confirmed reorganization of the brain areas.

More recently, researchers hypothesized that phantom limb pain is maintained by the random entanglement of sensory, movement, and pain networks resulting from sensory and movement deprivation. The idea behind this theory is that pain networks impose themselves onto deprived sensory and motor networks so that moving the phantom limb will trigger phantom limb pain. Treatments such as graded motor imagery and phantom motor executions aim to sequentially activate the movement networks (below a threshold that would trigger pain) in such a way that movement and pain are uncoupled. These treatments are effective in reducing phantom limb pain in some patients. The lack of 100% efficacy suggests that phantom limb pain may be driven by different mechanisms at different times after amputation, or perhaps, by the combination of these mechanisms at a single point in time.

A link between the peripheral and central nervous systems

The peripheral and central mechanisms of phantom limb pain are not mutually exclusive; mechanisms at these levels may be connected. For example, patients with traumatic amputations, who are typically healthy and do not present with pain before amputation, may experience pain as a result of the peripheral mechanisms described above. However, because the peripheral and central nervous systems are linked, spontaneous electrical impulses from the periphery can contribute to maladaptive changes in the brain areas responsible for processing pain. These changes in the brain are linked to pain maintenance.

On the contrary, patients with painful diabetic neuropathy who experience persistent pre-amputation pain may present with maladaptive changes in the brain areas responsible for processing pain before undergoing a limb amputation. These changes may maintain pain after amputation. For example, in a study by Jensen and colleagues, patients reported phantom limb pain in the same area they experienced pre-amputation pain. In addition, the quality and intensity of the phantom limb pain were similar to that of pre-amputation pain. In this case, phantom limb pain could be driven at different levels of the nervous system at the same time.

The mechanisms underlying phantom limb pain are not well understood at this stage. However, the consensus is that pain is driven by maladaptive changes in the peripheral and central nervous systems. This suggests using a combination of treatments targeting different levels of the nervous system may yield superior pain reduction compared to using one treatment in isolation.