Research over the past 15 years has helped to clarify the anatomy and physiology of itch, the clinical features of neuropathic itch syndromes and the scientific underpinning of effective treatments. Two itch-sensitive pathways exist: a histamine-stimulated pathway that uses mechanically insensitive C-fibres, and a cowhage-stimulated pathway primarily involving polymodal C-fibres. Interactions with pain continue to be central to explaining various aspects of itch. Certain spinal interneurons (Bhlhb5) inhibit itch pathways within the dorsal horn; they may represent mediators between noxious and pruritic pathways, and allow scratch to inhibit itch. In the brain, functional imaging studies reveal diffuse activation maps for itch that overlap, but not identically, with pain maps. Neuropathic itch syndromes are chronic itch states due to dysfunction of peripheral or central nervous system structures. The most recognized are postherpetic itch, brachioradial pruritus, trigeminal trophic syndrome, and ischaemic stroke-related itch. These disorders affect a patient’s quality of life to a similar extent as neuropathic pain. Treatment of neuropathic itch focuses on behavioural interventions (e.g., skin protection) followed by stepwise trials of topical agents (e.g., capsaicin), antiepileptic drugs (e.g., gabapentin), injection of other agents (e.g., botulinum A toxin), and neurostimulation techniques (e.g., cutaneous field stimulation). The involved mechanisms of action include desensitization of nerve fibres (in the case of capsaicin) and postsynaptic blockade of calcium channels (for gabapentin). In the future, particular histamine receptors, protease pathway molecules, and vanilloids may serve as targets for novel antipruritic agents.
We have all experienced an itch (pruritus). It is a common phenomenon of remarkable complexity that has received little attention in neurological circles. In the past 15 years, important advances have occurred in the study of itch. The purpose of this paper is to review these developments and their clinical implications.
Itch is an irritative sensation provoking a scratch response that provides at least temporary relief. It has been thought to serve a self-protective function, protecting the skin against potentially harmful agents (Ständer et al., 2003). It normally originates in the skin and transitional tissues, such as the conjunctiva and anal mucosa (Handwerker, 2010), which it serves to protect from irritating stimuli, but it can also arise from central or peripheral neurological lesions, as when it follows a stroke or occurs with a peripheral neuropathy. The sensation may be localized or widespread, commonly outlasts any inciting mechanical stimulus, and is sometimes accompanied by skin changes. ‘Itchy skin’ (alloknesis) is itch that is augmented by light touch or brushing of the skin around an itching source. Similarly, hyperknesis is itch that is enhanced when prick is applied to the same surrounding skin area. Itch is a multidimensional perception with psychological aspects, as exemplified by its induction by thought of an irritating stimulus (Ikoma et al., 2006). The scratch phenomenon, which may be painful in other instances, may paradoxically be pleasurable in the setting of itch. Other motor responses to itch include rubbing or kneading the skin. The local application of cold may also alleviate itch.
An important distinction is between acute and chronic itch. Acute itch is the sensation experienced when itch-inducing stimuli (‘pruritogens’) contact the skin, and usually is relieved by pain (including scratch) in the surrounding area. Chronic itch is the persistent sensation that results from various causes in which pain does not relieve itch, and may actually be perceived as itch. Table 1 shows some common causes of chronic itch. By analogy to pain, two main mechanisms of chronic itch have been proposed. Peripheral sensitization is the decreased activation threshold and increased basal activity of itch-related receptors and nerve fibres (Potenzieri and Undem, 2012). Central sensitization results from the neuroplasticity that occurs in the spinal cord and brain such that non-pruritic stimuli are perceived as, or augment, itch. Together, these are the likely mechanisms of alloknesis and hyperknesis described earlier.
Parasitosis (including scabies, trichinosis, ascariasis, and hookworm)
See Table 2
Itch in the elderly
In some cases, such as uraemia or HIV infection, a specific disorder has multiple causes for itch, but the disorder has been placed in a single category for simplicity.
Itch and pain are distinct sensations that interact. The clearest differentiation is the effector limb of both—withdrawal for pain compared with scratch for itch. The interaction between itch and pain traditionally has been thought to be antagonistic, with pain dominant over itch (Ständer et al., 2003; Binder et al., 2008). This is most evident when noxious counter-stimuli on the skin suppress itch sensation, and when opioids inhibit pain but cause itch (Handwerker, 2010; Schmelz, 2010). The mechanism of this itch suppression is probably central, as supported by at least two experimental observations: (i) noxious counter-stimuli may be several centimetres or more away from the itch source, suggesting that different afferent fibres are activated and their signals integrated at the spinal cord level (Seo et al., 2009; Davidson and Giesler, 2010); and (ii) secondary hyperalgesia induced by various noxious stimuli in a skin patch of several centimetres abolishes cowhage or histamine-induced itch in that area, consistent with a central excitatory state that activates pain and blocks itch (Brull et al., 1999; Binder et al., 2008).
Unfortunately, a direct antagonistic relation is too simplistic when pathological lesions are considered. Depending on the location and pathophysiology of the lesion, there are varied manifestations. For instance, herpes zoster neuropathy can produce either chronic pain or itch, or pain and itch concomitantly (Oaklander et al., 2002; Ikoma et al., 2006). Further, many central causes of itch (e.g. spinal cavernous hemangioma) are associated with no reduction in pain despite very prominent itch (Potenzieri and Undem, 2012; Lanotte et al., 2013). Lastly, hyperknesis, in which itch is augmented by pain, also implicates a reversal of the traditional model. Although the exact mechanisms in these instances are unclear, some have suggested that pathological disruption of the antagonistic interaction unmasks normally silent neuronal pathways that allow a sensory stimulus to evoke a new sensation, such as a painful stimulus evoking itch (Ma, 2012).
Peripheral biology of itch
In typical circumstances, pruritogens stimulate skin receptors and activate the peripheral pathway of itch. This provokes a signalling cascade and action potentials in at least two types of C-fibres. These nerve fibres conduct the action potential to the dorsal horn of the spinal cord (Fig. 1).
Peripheral anatomy of itch. Two itch pathways (histamine in blue, cowhage in red) and the peripheral limb of the pain pathway (green) are shown. The itch pathways are stimulated by histamine and cowhage skin receptors in the epidermis and dermis, respectively. Impulses are transmitted primarily via mechanically insensitive C-fibres and polymodal C-fibres, respectively, to secondary neurons in the dorsal horn. One means of modulation by the pain pathway (depicted only partially) is through a Bhlbb5 interneuron. STT = spinothalamic tract; Bhlbb5 = transcription factor protein; CMi = mechano-insensitive C-fibres; C-polymodal = polymodal C-fibres.
A key debate is the encoding of itch in the afferent pathway. Similar to theories on pain, the opposing positions are specificity theory and pattern theory. Supporters of the specificity theory argue for modality-specific receptors and peripheral nerves that constitute a ‘labelled line’ from the skin to the brain, as originally articulated in Muller’s 1826 doctrine of ‘specific nerve energies’. Proponents of the pattern theory argue that somatic sensations including itch are generated by receptors and peripheral nerves that are not specific to the stimulus and that deliver a series of signals modulated and decoded centrally (Craig, 2003).
The strongest evidence for a labelled line for itch came from studies using microneurography, spinal cord electrophysiology mapping, and genetic modification techniques. Schmelz et al. (1997) using microneurography in humans, described a histamine-stimulated itch pathway consisting of mechanically insensitive C-fibres that responded with a temporal profile consistent with itch as rated by human subjects. These fibres were characterized as slowly conducting thin axons with large innervation territories on the skin. Subsequently, a complementary histamine-stimulated central pathway for itch consisting of lamina I spinothalamic tract neurons was mapped in the spinal cord of cats (Andrew and Craig, 2001). These neurons had distinct conduction velocities and thalamic projections, differentiating them from pain or temperature projections. In a genetic mouse model, Sun and Chen (2007) identified a subset of dorsal root ganglion neurons releasing gastrin-releasing peptide (GRP) and their receptors (gastrin-releasing peptide receptor; GRPR) located in lamina I of the spinal cord. When GRPR was genetically knocked out or antagonized pharmacologically, scratching behaviour was inhibited. Most recently, Han et al. (2013) described a set of MrgprA3-positive neurons in the dorsal root ganglion of mice that innervate the epidermis and respond to multiple pruritogens. Ablation of these neurons reduced itch behaviour; conversely, their excitation, regardless of the stimulus (even when pain-producing, i.e., an algogen), evoked a scratch response.
The labelled-line theory does not explain two experimental observations: (i) the itch pathway may be activated by pain-producing stimuli; and (ii) different pruritogens may activate other peripheral pathways. The strongest experimental evidence for the first point was provided by Schmelz et al. (2003) who studied the response of previously identified itch fibres to capsaicin and bradykinin, which are algogenic agents. They showed that such algogens activated mechanically insensitive C-fibres initially thought to be specific for histaminergic itch.
The second point was fuelled by study of the pruritogen cowhage. Originating from the Hindi term ‘kiwach’ meaning ‘bad rubbing’, cowhage is the common name for barbed hairs of the tropical plant Mucuna pruriens (Namer et al., 2008). The spicules cause an intense itch without extended erythema, making it a useful agent to study non-histaminergic itch. Namer et al. (2008) used microneurography to study cowhage-induced itch in humans. They found that cowhage stimulated mechanically responsive or polymodal C-fibres and not the mechanically insensitive fibres related to histamine-induced itch. This finding accords with observations that cowhage and histamine activate different spinothalamic neurons in primate spinal cords (Davidson et al., 2007). Cowhage also activates a subpopulation of A-delta thinly myelinated fibres in addition to unmyelinated afferents (Ringkamp et al., 2011). Collectively, these findings suggest that a pruritogen may activate multiple pathways. Secondly, if one holds to the specificity theory, these findings support the contention that there are at least two separate itch pathways (histamine and cowhage) involved in a sensation experienced as relatively similar by human subjects. Histamine itch may be more ‘burning’ and cowhage itch more ‘stinging’ in quality, although this is difficult to distinguish by inexperienced subjects (Handwerker, 2010).
At the molecular level, the division and integration between itch pathways is even more complex. Each pruritogen activates G protein-coupled receptors (GPCRs) and downstream messengers that interact with other itch- and pain-signalling systems (Jeffry et al., 2011). For histamine, three GPCRs (HR1, HR3 and HR4) are implicated in acute itch responses in mice. These receptors activate phospholipase C, phospholipase A2 and transient receptor potential vanilloid 1 (TPRV1), resulting in elevated intracellular calcium in skin-specific dorsal root ganglion sensory neurons (Rossbach et al., 2011). Cowhage cleaves the extracellular domain of protease-activated receptor 2 (PAR2) that activates phospholipase C, TRPV1 and transient receptor potential ankyrin 1 (TRPA1), leading to membrane depolarization (Davidson and Giesler, 2010). PAR2 also has a role in peripheral sensitization; when it is stimulated, the primary sensory nerve endings become more responsive to multiple non-histamine pruritogens (Akiyama et al., 2012). Finally, chloroquine, another potent pruritogen, interacts with a GPCR termed MgrprA3 on cells that also express H1R, but likely with a separate intracellular signalling cascade (Potenzieri and Undem, 2012). Importantly, TRPV1 and TRPA1—the receptors originally found to be responsive to capsaicin and mustard oil, respectively—are not just pain sensors but also integrate various noxious stimuli, including those that elicit an itch sensation (Ross, 2011).
There have been attempts to reconcile the labelled line theory with the findings of non-specificity described above, but the issue remains unsettled. Many advocate for a mechanism that includes both labelled lines and pattern decoding. The ‘population-coding theory’ suggests that pain and itch labelled lines interconnect through excitatory and inhibitory interneurons that modulate the activity of each other, usually in the spinal cord. If pain and itch fibres are activated together, the sensation of pain alone may then emerge because inhibition from interneurons and central descending pathways masks itch sensation. However, in patients with chronic itch, failed crosstalk of these lines creates ‘pro-pain’ and ‘pro-itch’ pathways irrespective of the stimulus (Ma, 2012).
Central biology of itch
The central pathway that mediates itch begins with a synapse in the dorsal horn from a primary to secondary afferent neuron. The axons of the secondary cells travel in the contralateral spinothalamic tract and synapse onto third-order neurons in the thalamus, from which axons project diffusely to cortical and subcortical regions (Fig. 2).
Central anatomy of itch. Histamine (red) and cowhage (blue) itch fibres travel in the spinothalamic tract to the thalamic nuclei (listed in the left box) usually contralateral to the stimulus. Cowhage neurons project to a broader set of nuclei than histamine neurons [data derived from Davidson et al. (2012) and Papoiu et al. (2012)]. Thalamic neurons transmit to multiple cortical and subcortical structures (listed in the right box) with a more diffuse pattern in cowhage itch. These structures are usually activated in a bilateral and symmetric pattern, except that the insular cortex, claustrum, basal ganglia, and putamen have a minor emphasis contralateral to the stimulus [data derived from Leknes et al. (2007); Papoiu et al. (2012)]. STT = spinothalamic tract.
At the dorsal horn, neurotransmitters such as calcitonin gene-related peptide, gastrin-releasing peptide (GRP), substance P, and glutamate have been implicated in the first itch synapse (Davidson and Giesler, 2010). In mice, these are theorized to bind to GRPR in the spinal cord to mediate itch (Sun and Chen, 2007; Sun et al., 2009). Koga et al. (2011) found that glutamate, not GRP, is the primary neurotransmitter binding to GRPR. Histamine and cowhage primary afferents do not activate or converge on the same secondary neuron, suggesting mutually exclusive dual projections in the spinothalamic tract that mediate itch (Davidson et al., 2007).
The dorsal horn synapse likely represents the site of modulation by parallel pain-processing neurons and descending pathways from the brain. In mice, a population of inhibitory interneurons (Bhlhb5 neurons) inhibit itch within the dorsal horn, and loss of these neurons result in persistent itch (Ross et al., 2010). Such interneurons may represent the mediators between noxious and pruritic pathways, and serve as the cellular basis of how scratch inhibits itch (Ross, 2011). This concept is supported by the observation that scratching inhibits spinothalamic neurons during histamine-evoked activity but not spontaneous activity (Davidson et al., 2009). There is also biochemical evidence that scratching excites inhibitory interneurons to release glycine and gamma-aminobutyric acid and thereby inhibit itch-responsive neurons (Akiyama et al., 2011).
The evidence for central sensitization in itch is accumulating. In one study, such sensitization was demonstrated in patients with atopic dermatitis, who were more likely than healthy subjects to perceive noxious stimuli as itchy, even though such stimuli usually reduce itch. Further, when conditioned with histamine, the healthy participants also reported that noxious stimuli caused itchiness. It was suggested that the chronic pruriceptive input caused central sensitization for itch and—in consequence—pain stimuli caused rather than inhibited itch (Ikoma et al., 2004). Unlike pain, study of the molecular mechanisms of central sensitization in itch pathways is still in its early stages. Nevertheless, it is known that toll-like receptor 3 (TLR3) induces spinal cord long-term potentiation, resulting in central sensitization and synaptic plasticity. In mice lacking the Tlr3 gene, there is decreased scratching in response to both histaminergic and non-histamingeric pruritogens (Liu et al., 2012).
A number of thalamic nuclei are implicated in the itch pathway. Cowhage-sensitive itch neurons have a broader thalamic representation than histamine-responsive itch neurons. Both sets of neurons have axons terminating in the contralateral ventral posterior lateral, ventral posterior inferior, and posterior nuclei, but cowhage neurons have additional projections to the contralateral suprageniculate and medial geniculate nuclei (Davidson et al., 2012) (Fig. 2). Neuroimaging data suggest an increased activation of thalamic nuclei including the pulvinar in cowhage itch (Papoiu et al., 2012), but this was not confirmed in the above electrophysiology studies.
Functional neuroimaging studies (PET and functional MRI) have been used in humans to study brain regions implicated in itch. Multifocal regions related to perception, evaluative processes, motivation, attention, emotion, and motor function are involved. This parallels regions involved in pain processing, though not identically. Specifically, most studies reveal activation of the thalamus, somatosensory cortex, parietal cortex, motor areas (primary motor cortex, supplementary motor area, premotor cortex), prefrontal cortex, anterior cingulate gyrus, insula, and midbrain (Hsieh et al., 1994; Drzezga et al., 2001; Mochizuki et al., 2003; Leknes et al., 2007; Papoiu et al., 2012). In contrast to pain, most studies, but not all (Papoiu et al., 2012), have shown a lack of activation of the secondary somatosensory cortex. When cold pain and itch are experienced simultaneously, there is activation of the periaqueductal grey, which may serve as a modulator of the sensory input, with accompanying deactivation of the cortico-subcortical network (Mochizuki et al., 2003). Again, there appears to be a distinction between histamine and cowhage itch as the claustrum, insula, basal ganglia, putamen and thalamus are more activated by cowhage than histamine (Papoiu et al., 2012) (Fig. 2).
Neuropathic itch syndromes
Neuropathic itch results from a lesion in the afferent sensory pathway (Twycross et al., 2003), rather than from a cutaneous lesion or because of a peripheral stimulus. The site of involvement may be (i) peripheral nerve and root; (ii) spinal cord; or (iii) brain (Table 2). Most conditions are chronic itch states accompanied by other sensory features (paraesthesias, hyperaesthesia, hypoaesthesia) in a distribution that may suggest the site of the lesion (e.g., dermatomal for radicular or spinal cord lesions). The impact of itch on patients’ quality of life is significant; one study indicated that chronic pruritus has an impact comparable to that of chronic pain (Kini et al., 2011).
Cavernous hemangiomas; other vascular malformations
Post-traumatic Brown-Séquard syndrome
Ischaemic brainstem and subcortical strokes
Demyelinating inflammatory disorders
When a neuropathic itch syndrome is suspected, neurophysiological and pathological studies may help to support the diagnosis or suggest the underlying pathology. Quantitative sensory testing may demonstrate abnormal thermal and pain thresholds and itch intensity experience (Bin Saif et al., 2012). A recent study showed C-nerve fibre sensitivity to itch and temperature varies in different body areas (Bin Saif et al., 2012). Skin biopsies, with study of the intraepidermal nerve fibre density, may also be useful in showing characteristic neuropathic changes, as is discussed below with regard to prurigo nodularis, herpes zoster, and notalgia paraesthetica.
Polyneuropathies and itch
Itch is reported to occur in ∼30% of patients with peripheral neuropathies, and in one-third of these patients the itch is rated as moderate to severe in intensity by the response to the item on itch in the Neuropathy Pain Scale (Binder et al., 2008). The main peripheral syndromes with prominent itch are summarized in Table 2.
Although there are multiple causes of itch with HIV infection, prurigo nodularis (itchy skin nodules) is one cause that probably relates to the polyneuropathy that occurs in these patients (Singh and Rudikoff, 2003). In prurigo nodularis, there is a stronger immunoreactivity of dermal nerves to a neurotrophic agent, p75 low affinity nerve growth factor receptor, than in normal tissue. This leads to hyperplasia and arborization of unmyelinated nerve fibres and to increased production of neuropeptides that can result in an inflammatory reaction (Liang et al., 1999). Pruritus is also a common side effect of certain chemotherapy agents, especially platinum compounds (Shepherd, 2003). Paclitaxel-carboplatin treatment, for example, may cause disabling pruritus, sometimes accompanied by painful paraesthesias and hyperalgesias, which is unresponsive to antipruritic agents, including steroids (Dunphy et al., 1997). Intriguingly, some polyneuropathies are accompanied by itch suppression. For example, although most patients with Fabry disease report neuropathic pain due to a small-fibre neuropathy (Weidemann et al., 2008), they may have a lack of itch sensation after histamine injection, and indeed note an absence of itch after mosquito bites (Cable et al., 1982).
This disorder, analogous to postherpetic neuralgia, is a chronic itch state in which refractory itch occurs in a region previously affected by shingles. In one remarkable case, a 39-year-old woman with postherpetic itch scratched through the skull into her frontal lobe after ophthalmic zoster. Quantitative study showed loss in the skin patch of all sensory modalities except itch, which was enhanced by histamine. Immunolabelled skin biopsies showed near-complete loss of epidermal innervation by afferent nerve processes. The authors suggest that analgesic skin led to a lack of inhibitory pain signals and hyperactivity of central itch neurons (Oaklander et al., 2002).
Brachioradial pruritus is itch—commonly combined with tingling, burning, and stinging—in the dorsolateral aspect of the upper arm and is often bilateral. Two predominant mechanisms have been suggested: (i) sun exposure causing symptoms with seasonal variation; and (ii) intervertebral foraminal narrowing with compression of nerve roots causing non-seasonal symptoms. The first mechanism was more common in a study of 95 patients (Veien and Laurberg, 2011).
A peripheral neuropathy of the posterior rami of the thoracic nerves in the area of the second and sixth vertebrae probably accounts for this syndrome. Typically, pruritus occurs with or without sensory loss in an area just medial to the scapula border in the associated dermatomes. Some have attributed this disorder to spinal nerve root impingement by degenerative changes (Savk and Savk, 2005). We favour damage to the peripheral nerves by musculoskeletal compression; such an aetiology is supported by the finding of a reduced intraepidermal nerve fibre density in the pruritic region compared with an ‘anatomically identical’ non-involved site (Huesmann et al., 2012).
Trigeminal trophic syndrome
This syndrome is characterized by unilateral itch-induced nostril erosion and ulceration with lower face anaesthesia and paraesthesia after trigeminal nerve surgery, alcohol injection or postherpetic neuropathy (Nagel and Gilden, 2011). There are also reports of this condition after stroke, described below. Often exacerbated by dementia, patients sometimes require complex maxillofacial reconstructions (Bhatti et al., 2008) or protective facemask (Swan et al., 2009). The mechanism, similar to postherpetic pruritus, is thought to be absent pain inhibition leading to enhanced itch (Fitzek et al., 2006).
Keloids and burn scars
Keloids and burn scars commonly lead to itch and pain. For keloids, one study of 28 patients showed that 86% experienced itch mostly at the edge of the keloid, which has been speculatively attributed to small nerve fibre dysfunction (Lee et al., 2004).
In the brain, itch syndromes usually have a subcortical or brainstem localization, and are most commonly caused by ischaemic strokes. Unilateral pruritus has been described in brainstem, thalamic, subcortical and, rarely, cortical ischaemic strokes (Kimyai-Asadi et al., 1999; Seo et al., 2009; Curtis et al., 2012). Similar to central neuropathic pain, symptoms may develop several days to weeks after stroke (Kimyai-Asadi et al., 1999). In lateral medullary ischaemic strokes, in particular, there is a predisposition to cause facial neuropathic itch that may lead to trigeminal trophic syndrome (Seo et al., 2009). The latter develops when patients not only have facial itch, but also hypoaesthesia that results in scratching to the point of self-injury (Fitzek et al., 2006). The mechanism of post-stroke pruritus is unclear, but disrupted central modulatory structures may be involved to produce central sensitization.
Similar to strokes, neoplasms, especially in the posterior fossa, may manifest with localized persistent itch of the face or upper body (Darken et al., 2009). One patient who developed intense itch of the left face and trunk was found to have a brainstem paraneoplastic disorder associated with prostate cancer (Berger et al., 2009). Itch is relatively common in familial Creutzfeldt-Jakob disease carrying the E200K mutation. Imaging data in these cases show damage to the midbrain periaqueductal grey matter suggesting alteration of central itch modulation (Cohen et al., 2011). In demyelinating disorders, paroxysmal itch may particularly affect patients with brainstem or spinal cord plaques. In a study of neuromyelitis optica, 27.3% of patients reported itch within a week of other symptoms of transverse myelitis that usually involved the central cord. The high frequency, especially compared to itch in multiple sclerosis (4.5%), may be due to more direct involvement of dorsal horn neurons and peri-aqueductal pathways (Elsone et al., 2012).
Treatment of neuropathic itch
Management of neuropathic itch begins with non-pharmacological measures used for itch in general, followed by other therapies tried in a stepwise approach. Non-pharmacological treatments are the mainstay of initial management and should continue even if drug therapies are required. They include behavioural interventions (e.g., educating the patient about itch effects, nail-cutting and wearing protective garments) (Oaklander, 2011), use of moisturizers, wearing loose clothing, and avoidance of warm temperatures (Class IV evidence) (Twycross et al., 2003). Phototherapy with narrow-band ultraviolet B has been shown to help in five cases of notalgia paraesthetica, but its use has not been explored in other neuropathic itch conditions (Class IV evidence) (Pérez-Pérez et al., 2010).
As neuropathic itch is often induced without the involvement of histamine, it is not surprising that antihistamines are generally unhelpful in treatment. Any benefit appears related more to the somnolent side effects of H1-antihistamines than to a more specific antipruritic effect (Summey and Yosipovitch, 2005).
There are several medications reported for use in neuropathic itch. Capsaicin cream and topical anaesthetics such as a lidocaine patch (Sandroni, 2002) are commonly used initially (Class IV evidence). The mechanism of capsaicin is TRVP1 activation that leads to receptor desensitization and depletion of substance P from sensory nerve terminals in the skin (Ikoma et al., 2006). In a meta-analysis of six randomized controlled studies testing capsaicin for diverse itch conditions, there was no convincing evidence in favour of the treatment. Methodological and statistical problems limited the validity of these controlled studies. Specifically, it is difficult to design experiments because the burning sensation induced by capsaicin cannot be masked, making it challenging to find a suitable placebo to maintain the blinding of patients and investigators (Gooding et al., 2010). Other topical agents that have been used with anecdotal claims of benefit include lidocaine 5% gel (Feramisco et al., 2010), cortisone and tacrolimus (Nakamizo et al., 2010). The latter activates C-fibres with TRPV1 receptors through an increase of intracellular calcium ions (Senba et al., 2004).
As understanding has increased about the aetiologies and pathophysiology of itch, new therapeutic strategies have emerged, including anti-seizure medications and antidepressants. Among the anti-seizure medications, gabapentin—widely used for neuropathic pain—is best studied. As indicated previously, scratching excites inhibitory interneurons to release glycine and gamma-aminobutyric acid and thereby inhibit itch-responsive neurons (Akiyama et al., 2011). Gabapentin is a structural analogue of gamma-aminobutyric acid that is thought to block the alpha2delta subunit of voltage-dependent calcium channels in dorsal horn postsynaptic cells. In a randomized control trial of 60 patients with itch following burns, gabapentin was more effective and faster in action than the antihistamine cetirizine. There was no difference between gabapentin alone and the combination of gabapentin and cetirizine. All patients receiving gabapentin had become itch-free by 28 days, whereas only 3 of 20 patients receiving cetirizine alone became itch-free. Furthermore, gabapentin had no reported side effects, whereas cetirizine caused sedation (Class II evidence) (Ahuja et al., 2011). Other anecdotal reports describe the efficacy of gabapentin—either alone or in conjunction with topical agents such as capsaicin or tacrolimus—in treating single or a few patients with itch syndromes such as trigeminal trophic syndrome (Nakamizo et al., 2010) and brachioradial pruritus (Bueller et al. 1999; Winhoven et al., 2004) (Class IV evidence). Other anti-seizure and antidepressant agents were tried for itch by analogy with their use for treating neuropathic pain. There are anecdotal reports of antipruritic benefit with pregabalin (Thielen et al., 2008), lamotrigine (Crevits, 2006), carbamazepine, doxepin, amitriptyline, nortriptyline and paroxetine (Class IV evidence). These medications merit more rigorous trials to determine their efficacy.
In refractory cases, thalidomide or injectable treatments may be used, depending on the underlying disorder. Thalidomide is an effective treatment for prurigo nodularis, perhaps due to its suppression of tumour necrosis factor-alpha, which is elevated in many skin disorders with itch (Summey and Yosipovitch, 2005) (Class IV evidence). The limiting factors are teratogenicity and thalidomide-induced peripheral neuropathy. Injectable treatments include intracutaneous botulinum A toxin (Wallengren and Bartosik, 2010; Kavanagh and Tidman, 2012), epidural injections of clonidine and bupivacaine (Elkersh et al., 2003), and stellate ganglion blocks (Class IV evidence). In one study, botulinum A toxin was injected into pruritic skin patches at small intervals (1.5 cm apart) in a dose range of 0.8–1.4 units per point. (Wallengren and Bartosik, 2010). It showed moderate and mostly transient improvement in patients with localized itch, including notalgia paraesthetica. The proposed mechanism is modulation of TRPV1 receptors in a similar fashion to capsaicin to reduce the release of C-fibre neuropeptides (Tugnoli et al., 2007; Gazerani et al., 2009).
Additional treatments for refractory itch include peripheral and central surface stimulators. Cutaneous field stimulation electrically stimulates thin afferent fibres. When applied to the affected regions in patients with localized itch, including brachioradial pruritus and notalgia paraesthetica, there was a 49% reduction of itch compared to baseline and 40% reduction in itch-related epidermal nerve fibres (as determined by immunoreactivity) on skin biopsies (Class IV evidence) (Wallengren and Sundler, 2001). Transcranial direct current stimulation involves a non-invasive neuromodulator applied to the scalp to influence neuroplasticity. In a case report, it was found to improve one patient’s recalcitrant itch, but not pain, for 3 months (Class IV evidence) (Knotkova et al., 2013). Further study of stimulation techniques to relieve intractable itch seem warranted.
Except for a case report of microsurgical decompression in notalgia paraesthetica (Class IV evidence) (Williams et al., 2010), little has been published on neurosurgical interventions in the management of chronic itch syndromes, and the role of such interventions is unclear at this time.
Certain other therapies, which have been used to treat non-neuropathic itch, may also be tried in patients with otherwise refractory neuropathic syndromes (Feramisco et al., 2010), although no evidence of efficacy in neuropathic itch has been published. Mu-opioid receptor antagonists and kappa-opioid receptor agonists may alleviate pruritus occurring in the context of systemic or skin diseases. Naloxone has antipruritic effects due to modulation of mu-opioid sensitive interneurons or wide dynamic range neurons in the spinal cord (Schmelz, 2001). It is an effective treatment for opioid-induced itch, cholestatic itch, chronic urticaria and atopic dermatitis (Phan et al., 2010). Nalfurafine and butorphanol activate cutaneous kappa-opioid receptors without central side effects, and decrease itch especially in patients with uraemia (Reich and Szepietowski, 2012).
Future therapeutic approaches
In the future, multiple therapies may be derived from an appreciation of the itch pathways and mediators described here. As discussed by Tey and Yosipovitch (2011), novel histamine receptor antagonists target one of the peripheral histamine receptors (HR4) and show efficacy in histamine-based pruritus models. Modulators of the protease pathway such as nafamostat mesilate and tetracyclines target PAR2, which is involved in the cowhage itch pathway. Cannabinoids that affect the vanilloids, particularly TRPV1 involved in both histamine and cowhage itch, show promise as treatments for acute and chronic itch. Eventually, dorsal root ganglion neuron molecules such as GRP, GRPR, and Mrgprs may become important central targets to inhibit itch.
We thank Jon Levine, MD, PhD, for reviewing an early version of the manuscript, and Ms. Kathleen Jee for developing the figures.