Painful Peripheral Neuropathies

Definition of Neuropathic Pain

Neuropathic pain is never a complete diagnosis in its own right but is always a symptom of an underlying neurological disease. From a historical perspective, it is interesting to reflect that the terms neuralgia and neuropathic pain were coined to describe pain of peripheral nerve origin. However, it has long been recognized that these definitions are often inappropriately used, a sentiment expressed by Robert more than 5 decades ago: “The indiscriminate use of the term neuralgia to designate almost every undeterminable, and often not neurogenic, painful affection is a menace to exact medical diagnosis.” Recently, a redefinition of neuropathic pain has been endorsed by the International Association for the Study of Pain (IASP) ( ). In the new definition, neuropathic pain is defined as “pain caused by a lesion or disease of the somatosensory system.” This supersedes previous IASP definitions and, importantly, removes it from the ill-defined term of “dysfunction of the nervous system.” The new, more restrictive definition of neuropathic pain provides a welcomed increase in diagnostic stringency. However, its major weakness is the associated grading system of diagnostic certainty. Under this grading system, classic trigeminal neuralgia does not fulfill the criteria for neuropathic pain, whereas many non-neuropathic conditions could still be categorized as neuropathic pain by virtue of an altered von Frey hair detection threshold in the skin ( ).

It is common clinical experience that patients with neuropathic pain have significant co-morbid conditions and that these conditions have an important impact on the global pain experience. Psychological factors such as changes in mood, anxiety, and altered sleep patterns have all been identified as significant adjuncts of painful neuropathies, and in addition there may be social isolation and reduced employment status ( ). Approximately 60% of patients report at least discomfort as a result of difficulty sleeping, and moderate to severe depression is present in a third of patients and anxiety in a quarter ( Fig. 65-1 ). Newly referred neurology outpatients with pain have a high prevalence of depression, pain is more likely to persist in patients with depression, and depression is more likely to persist in those with co-existent pain ( ).

Diseases of the nervous system commonly cause a variety of secondary consequences such as skeletal deformity, arthropathy, and musculoskeletal changes, all of which may be painful. In these circumstances the pain is likely to be nociceptive in type. Finally, the disease causing the neuropathy may, as in the case of diabetes mellitus, lead to a host of non-neuropathic types of pain, particularly those of vascular, joint, and skin origin. In individual patients it may be clinically difficult to differentiate between the relative contribution of these secondary factors to the overall burden of pain. This underlines the need for careful clinical assessment of all patients with pain and a neuropathy; the pain may not be of a neuropathic type and will demand appropriate investigations and treatment.

A number of questionnaires and pain scales have been devised to assist in the recognition and diagnosis of neuropathic pain, including the Neuropathic Pain Scale (NPS) ( ), the Leeds Assessment of Neuropathic Symptoms and Signs (LANSS) ( ), the Neuropathic Pain Questionnaire (NPQ) ( ), the Douleur Neuropathique en 4 questions (DN4) ( ), painDETECT ( ), and ID-Pain ( ). These instruments can be useful in formalizing the assessment of symptoms and signs and may be a helpful alert for the non-specialist to consider the possibility of a condition causing neuropathic pain in the differential diagnosis of a pain state. Although these instruments are often hailed as being “diagnostic,” they are in fact not and can even be positively misleading in non-painful neuropathies such as Charcot–Marie–Tooth (CMT) disease type 1, where the pain is often of musculoskeletal origin and can frequently co-exist in regions of reported numbness, mild painless paresthesias, and hypoesthesia to touch or pinprick. Therefore, questionnaires are not a shortcut and no substitute for the traditional approach to neurological diagnosis. A positive score on a questionnaire is therefore a potential starting point and not the end of the diagnostic process. In other words, if a pain is thought to be neuropathic in type, a neurological diagnosis must be sought.

Classification of Painful Neuropathies

There are different classification schemes for painful peripheral neuropathies. The anatomical distribution pattern of the affected nerves provides valuable differential diagnostic clues to possible underlying causes. It is therefore common clinical practice to group painful neuropathies into symmetrical polyneuropathies, diseases affecting many nerves simultaneously, typically in a length-related glove-and-stocking distribution; asymmetrical neuropathies with a mono- or multiplex distribution; or processes affecting the brachial or lumbosacral plexuses. Neuropathies commonly associated with pain are listed in Box 65-1 .

The diversity of clinical conditions known to produce peripheral neuropathic pain is astonishing, and it has been difficult to identify a common denominator for the pain in conditions as heterogeneous as post-traumatic and diabetic neuropathy.

This has led to the development of a symptom-orientated diagnostic approach to neuropathic pain conditions that supplements the etiology-based classification scheme. It recognizes the fact that neuropathic pain is usually manifested as a composite of several distinct pain symptoms ( ). A symptom-orientated approach does not negate the fact that distinct neuropathies behave differently clinically and that some neuropathic disease states may predispose to certain constellations of pain symptoms (e.g., touch-evoked pain in post-herpetic neuralgia [PHN]). The rationale of this approach recognizes several principles:

• 1.

  Clinically distinct pain symptoms such as ongoing stimulus-independent pain may be caused by similar, if not identical neural mechanisms, even if the underlying neuropathies differ. For example, nociceptive afferents that develop ongoing activity after axon injury could mediate ongoing pain regardless of the precipitating or sustaining neuropathic cause.

• 2.

  More than one pain mechanism is usually present in an individual patient. It is not uncommon that patients can clearly differentiate between the different qualities of stimulus-independent and stimulus-induced pain, and these different symptoms are probably caused by diverse neuronal changes.

• 3.

  Some symptoms, such as mechanical hypersensitivity, can be explained by several distinct neural mechanisms, which may even co-exist in an individual patient.

A symptom-based approach to painful neuropathies can be useful for dissecting the underlying neural mechanisms, and this knowledge may eventually be harnessed for the development of novel analgesic drugs that differentially target these mechanisms. However, a purely descriptive symptomatic approach to neuropathic pain that is not accompanied by a rigorous clinical differential neurological diagnostic assessment and appropriate investigations is negligent.

Epidemiology of Neuropathic Pain in Peripheral Neuropathy

Several studies have investigated the prevalence of neuropathies, but from these investigations it is often not possible to obtain information about the prevalence of pain. The difficulties surrounding the definition of neuropathic pain have compounded epidemiological surveys. Furthermore, the outcomes of epidemiological surveys obviously depend on the tools (questionnaires, clinical examination, outcome of appropriate investigations) that are used to make a diagnosis. Neuropathies represent one of the most common neurological disorders, with a prevalence of 2.4% in the general population that rises to 8% with age ( ). Estimates of the point prevalence of neuropathic pain in the general population are as high as 5–7% ( , , Institute of Medicine 2011 http://download.nap.edu/catalog.php?record_id=13172 ).

A complicating factor is that even within an etiologic entity such as diabetic polyneuropathy, chronic neuropathic pain develops in only a minority of patients. Thus, the point prevalence of chronic painful peripheral polyneuropathy is on the order of 10–20% of patients with diabetes mellitus ( ). Moreover, it is still unclear why the symptoms within an etiologically defined population of patients can be extremely diverse. In fact, it is one of the challenges of the field to understand the factors that determine why painful symptoms develop in some individuals only with apparently identical conditions and what determines the range of the initial symptoms.

Range of Painful Symptoms and Signs in Peripheral Neuropathies

No single symptom or sign is pathognomonic for neuropathic pain. The symptoms may be divided into those that are unprovoked ( stimulus independent ) and those that are provoked by maneuvers ( stimulus induced ) such as skin stimulation, pressure over affected nerves, or changes in temperature. The term deafferentation syndrome is often used for diseases characterized by extensive and complete disconnection of peripheral nerves from their target, such as with amputations or plexus lesions. In these situations the pain is stimulus independent. In contrast, when the connections in the periphery are partially retained or are at the borders of a completely deafferentated zone, the pain often has stimulus-independent and stimulus-induced components.

Many terms may be used by patients with neuropathies to describe their painful neuropathic sensations. Because some neuropathic pain symptoms are not experienced in normal situations, patients sometimes use bizarre and ornate verbal descriptors. The most commonly described spontaneous symptoms are deep aching in the extremities and a superficial burning, stinging, or prickling pain. Verbal descriptors from the McGill Pain Questionnaire that are used significantly more frequently by patients with neuropathic pain than by those with non-neuropathic pain include “electric shocks,” “burning,” “tingling,” “itching,” or “prickling,” whereas descriptions such as “dull,” “heavy,” and “tiring” are more often reported by patients with non-neuropathic pain ( ). Patients also report paroxysmal, shock-like lancinating pain, sometimes radiating through an entire limb. Signs of cutaneous hypersensitivity such as touch-evoked pain are more common in neuropathic pain states than in non-neuropathic pain states ( ). In addition, investigations of patients suffering from different neuropathies demonstrate patterns of sensory abnormalities. Table 65-1 shows that the symptoms and signs can be similar in neuropathic pain conditions, such as painful polyneuropathy and PHN, but that there can also be substantial differences when comparing these conditions with non-freezing cold injury (NFCI; also known as trench foot neuropathy) or complex regional pain syndrome (CRPS). However, the diagnosis of neuropathic pain on the basis of the patient’s description of symptoms in the absence of confirmatory findings on clinical examination and investigations is often not possible. One study of patients who were referred to a tertiary neurological center with the suspected diagnosis of neuropathic pain concluded that superficial ongoing pain and brush-evoked pain, but not cold or pinprick hyperalgesia, were more frequently found in patients with definitive and probable neuropathic pain than in patients unlikely to have neuropathic pain ( ). In the following descriptions of the different neuropathies, the major painful complaints typical of each condition are given, but it should be emphasized that within a single etiological or pathological diagnostic category, considerable variation in symptoms occurs in different individuals.

Type of Fibers Affected in Painful Neuropathies

The relationship of painful symptoms to morphological and electrophysiological changes in peripheral nerves has been a subject of interest for many decades, particularly since the introduction of nerve biopsy and the development of clinical electrophysiological techniques ( ). Overall, the human evidence converges on several principles:

• 1.

  Neuropathies characterized by rapid degenerative change are more likely to be painful. In some acute neuropathies, inflammatory changes might affect sensory neurons irrespective of whether they are lesioned ( ).

• 2.

  The co-existence of degenerative and regenerative changes appears to be an important factor ( ).

• 3.

  Ischemia in nerves may exacerbate paresthesias and pain secondary to peripheral damage, and in certain circumstances, severe ischemia, as in vasculitis, is the cause of neuropathies that may be very painful ( ).

• 4.

  Pain may be evoked by excitation of the nervi nervorum in mononeuropathies or in certain polyneuropathies ( ).

Neuropathies involving small fibers, with or without large-fiber involvement, are often painful. Many studies converge to suggest that axonal injury involving a portion of the nociceptive fibers in a peripheral nerve is the single most important causal factor for neuropathic pain:

• 1.

  There are several painless neuropathies in which small-diameter sensory neurons are largely spared that strongly endorse this view. This includes most types of CMT diseases (also known as hereditary motor and sensory neuropathy [HMSN]) type 1, where demyelination of large myelinated axons features as the dominant pathology, or ataxic neuropathies, such as Friedreich’s ataxia, where loss of the cell bodies and axons of large myelinated fibers is commonly observed.

• 2.

  The universal loss of nociceptors, as in congenital insensitivity to pain with anhidrosis (CIPA; also known as hereditary sensory and autonomic neuropathy [HSAN] type 4), does not cause pain, and this concurs with the view that severe loss of small-diameter axons in acquired neuropathies can often be painless.

• 3.

  Neuropathic pain typically develops only when sensory nerve fibers are affected. Thus, selective axonal or demyelinating lesions of motor fibers, as in motor neuron disease or multifocal motor neuropathy with conduction block, are generally painless.

• 4.

  Lesions of unmyelinated sympathetic fibers, such as in pure autonomic failure or Ross’s syndrome (segmental anhidrosis), are also not associated with pain.

• 5.

  Quantification of unmyelinated fibers in sural nerve biopsy specimens or assessment of epidermal nerve fiber (ENF) density in skin biopsy specimens reveals substantial damage to unmyelinated fibers in the majority of painful neuropathies ( ).

• 6.

  Finally, gain-of-function mutations in ion channels that are prevalently expressed in nociceptors, such as Na _v 1.7, cause pain ( ).

Pathophysiological Findings in Patients with Neuropathic Pain

Although it is clear that changes in central nervous system properties are crucial for the development of painful symptoms in peripheral neuropathy, several lines of independent evidence converge to indicate that changes in the excitability of primary nociceptive afferents are the single most important factor in the generation and maintenance of chronic neuropathic pain in humans. These experimental studies in patients are entirely in agreement with the neuropathological observations mentioned earlier that implicate small-diameter primary afferent fibers as the primary culprits in peripheral neuropathic pain:

• 1.

  Pain is often abolished or at least significantly reduced by local anesthetic block of damaged peripheral nerves ( ) or affected skin ( ), thus indicating that neural activity arising in the nerve or possibly even in the receptive endings contributes to the pain.

• 2.

  Stimulus-independent pain and some forms of stimulus-induced pain sensations persist during a differential nerve fiber block that eliminates conduction in myelinated non-nociceptive afferents ( ).

• 3.

  Psychophysical experiments indicate that the magnitude of ongoing pain correlates with the levels of nociceptor activity ( ).

The evidence from microneurographic investigations generally supports the notion of an abnormal sensitivity of primary sensory neurons in patients with painful nerve lesions by demonstrating abnormal activity and reduced thresholds in cutaneous afferents ( ). Abnormal ectopic activity in myelinated mechanosensitive fibers has been recorded in patients with traumatic nerve lesions, entrapment neuropathies, and radiculopathies ( Fig. 65-2 ). A classic study by provided evidence that ectopic excitation can occur at multiple sites in damaged sensory neurons. Ongoing activity and mechanical sensitivity were recorded proximal to a nerve neuroma in an amputee with phantom limb pain. Following local anesthetic blockade of the nerve distal to the recording site, impulses evoked by mechanical stimulation of the neuroma were abolished, but ongoing activity at the recording site continued, thus suggesting that this residual activity arose from the dorsal root ganglion (DRG) ( Fig. 65-3 ).

Microneurographic studies of unmyelinated nociceptors in patients with polyneuropathy have shown that individuals with pain have a significantly higher percentage of ongoing activity than do those without pain ( ). This difference was particularly pronounced in the population of mechanically insensitive nociceptors, which are thought to be important for central sensitization. Furthermore, nociceptors with ongoing activity were much more likely to discharge multiple times after a single electrical stimulation, a response that is unusual in healthy subjects or patients with painless neuropathies ( , ). This abnormal spiking with natural stimulation provides an explanation for the hyperalgesia in these patients.

Abnormal nociceptors have also been recorded in patients suffering from erythermalgia (also known as erythromelalgia), a condition characterized by painful, red, hot extremities. Nociceptors displayed ongoing activity that is not normally observed in nociceptors, and there was evidence of sensitization of mechanically insensitive afferents to non-painful tactile stimuli ( ) ( Fig. 65-4 ). The inherited form is caused by a gain-of-function mutation in the gene encoding the voltage-gated sodium channel Na _v 1.7 ( ). In heterologous expression systems the mutations can produce a hyperpolarizing shift in activation and a slowing of deactivation ( ) ( Fig. 65-5 ).

Clinical Features

The cardinal clinical features of a peripheral neuropathy are weakness or wasting of the affected muscles, hypoesthesia, loss of or attenuated tendon reflexes, and impaired autonomic functions. It is often possible to demonstrate this with a careful neurological examination of patients with neuropathic pain. However, none of these findings is specific for peripheral nerve disease, and consequently the neurological examination is only the initial step in a diagnostic process that makes use of a range of appropriate investigations. Pain or hyperalgesia in the absence of neurological symptoms or signs can be caused by peripheral nerve disease, but such diagnosis has to be treated with considerable suspicion until investigations provide definitive confirmatory evidence of nerve involvement.

Nerve Conduction Studies, Electromyography, and Evoked Potentials

Clinical neurophysiological investigations are some of the most important investigations to confirm the diagnosis of a neuropathy ( ). Nerve conduction studies in combination with electromyography (EMG) can make the important differentiation between demyelinating or axonal neuropathies and the presence or absence of conduction blocks. However, it is important to realize the limitations of these techniques. In routine practice, nerve conduction studies are restricted to the distal branches of a few major nerves in the extremities. Somatosensory evoked potentials (SEPs) or magnetic evoked potentials (MEPs) can be helpful in the diagnosis of proximal lesions. The main drawback of all these techniques is that they are in principle restricted to the assessment of large myelinated fibers, which are not the culprit in neuropathic pain. Laser-evoked potentials (LEPs) ( ), contact heat–evoked potentials (CHEPs) ( , Chao 2008), and pain-related electrically evoked potentials (PREPs) ( ) are capable of testing the function of Aδ fibers. These techniques are often not available outside tertiary referral centers and do not allow topographical differentiation between peripheral and central lesions. Even though EMG is a sensitive technique for the detection of axonal lesions that can be applied to virtually all skeletal muscles, it provides information only about the motor system. Therefore, abnormal findings on neurophysiological examination can provide positive evidence of nerve disease, whereas normal findings on examination can exclude major nerve damage but cannot fully exclude minor nerve damage or selective damage to thin myelinated or unmyelinated sensory nerve fibers, which are the main source of peripheral neuropathic pain. Microneurography, the only technique that permits direct assessment of the function of small fibers in humans, is generally available just in the research setting.

Nerve Biopsy

Nerve biopsy is an important diagnostic tool to establish the cause of a neuropathy. However, the invasive nature and the suitability of only a few nerves in routine practice are the major limitations of this technique. The two major pathological processes in peripheral neuropathy are axonal degeneration and segmental demyelination. Division of polyneuropathies into either of these pathological categories is somewhat artificial since both processes are usually present, albeit in varying proportions. The findings of axonal degeneration, affecting distal parts of the axon, the cell body, or both, the details of segmental demyelination, remyelination, and onion bulb formation, and the pathological reactions of unmyelinated fibers in different neuropathies are described extensively elsewhere ( ). Some problems with morphological studies of standard nerve biopsy specimens are that examination of transverse sections by light microscopy will fail to recognize segmentally demyelinated axons and thus will underestimate the myelinated fiber population. An increase in the density of small fibers does not necessarily imply a selective loss of large fibers since regeneration will increase the population of small fibers. By relating axonal diameter to myelin sheath thickness by electron microscopic examination and by using certain criteria for differentiating the sprouts of myelinated and unmyelinated fibers, this problem can be overcome to some extent ( ). Unfortunately, not all studies of nerve specimens include electron microscopic examination, and this essentially precludes the analysis of unmyelinated neurons.

Very few electrophysiological and pharmacological studies have been performed on isolated human nerves, which have essentially remained research investigations. took long multifascicular biopsy specimens of the sural nerve and compared compound action potentials with morphological changes in normal volunteers and those with various neuropathies. In Friedreich’s ataxia, a substantial reduction in Aβ potential with preservation of Aδ and C potential correlated well with the morphological decrease in the large myelinated fiber population. Conversely, in dominantly inherited amyloidosis, the absent C-fiber potential, the greatly reduced Aδ potential, and the only moderately reduced Aβ potential correlated well with a near absence of C fibers and reduced small myelinated fiber population on electron microscopy. Similar good correlations were found in two types of hereditary sensory neuropathies and in uremic neuropathy, but not in chronic relapsing inflammatory neuropathy. This was thought to be due to extensive segmental demyelination and remyelination and the resultant dispersion of large-fiber action potentials. In aggregate, these observations established that reasonable predictions about fiber population could be made from physiological observations, except when segmental demyelination was a prominent feature. Grafe and colleagues showed that in unmyelinated axons of the human sural nerve, the action potential propagation of many C fibers was accomplished by tetrodotoxin (TTX)-resistant sodium currents ( Fig. 65-6 )( ). Furthermore, there was a good correlation between the expression of TTX-resistant sodium channels and voltage-dependent calcium channels on nociceptive fibers expressing the capsaicin receptor transient receptor potential vanilloid 1 (TRPV1) ( ).