New Approaches to the Pharmacotherapy of Neuropathic Pain

Marie Besson, MD; Valérie Piguet, MD; Pierre Dayer, MD; Jules Desmeules, MD

Expert Rev Clin Pharmacol. 2008;1(5):683-693. 

Abstract and Introduction

Abstract

Pain is one of the most debilitating symptoms that presents with neuropathy. Neuropathic pain syndrome is a challenge to treat and, even with appropriate evidence-based treatment, only a 40% reduction of symptoms can be achieved in approximately half of patients. Furthermore, efficient doses are often difficult to obtain because of adverse effects. These observations underline that the treatment of neuropathic pain is still an unmet medical need. New approaches to the pharmacotherapy of neuropathy embrace different lines of work, including a fundamental mechanism-based approach, a clinical mechanism-based approach and an evidence-based approach. Moreover, interindividual variability in drug response, and genetic polymorphism in particular, is an emerging aspect to consider. Together with reviewing recent evidence-based guidelines as well as briefly discussing genetic polymorphisms that may influence the individual responses to treatments, this article will focus on what a mechanism-based approach is bringing to the clinical setting, on the perspective in fundamental research and on the difficulty of bridging the gap between fundamental notions and positive clinical outcomes.

Introduction

Pain is one of the most debilitating features of neuropathy. Neuropathic pain syndrome is actually a group of disorders, characterized by various etiologies affecting the nervous system, including infection, inflammation, metabolic disease, trauma or compression and chemical-induced nerve damage. The nerve lesion may be related to the CNS (brain or spinal cord), PNS, or both. Examples of peripheral neuropathic pain are diabetic polyneuropathy (DPN), postherpetic neuralgia (PHN), HIV sensory neuropathy or carpal-tunnel syndrome. Examples of central neuropathic pain are traumatic spinal cord injury, central poststroke pain or pain associated with a degenerative disease, such as multiple sclerosis.^[1,2]

Absolute incidence of neuropathic pain taken as a whole is unknown, but probably underestimated. A recent epidemiological study describing the incidence in the community of four chronic neuropathic pain conditions over a 10-year period in the UK reported an incidence, per 100,000 person-years, of 40 (95% confidence interval [CI]: 39–41) for PHN, 27 (95% CI: 6–27) for trigeminal neuralgia, 1 (95% CI: 1–2) for phantom limb pain and 15 (95% CI: 15–16) for DPN.^[3]

Irrespective of the cause, patients suffering from neuropathic pain often display a shared semiology, consisting of a burning constant pain with stabbing paroxysmal attacks. On clinical examination, signs of hypersensitivity, such as hyperalgesia (defined as an exaggerated sensation after a painful stimulus) or allodynia (a painful sensation elicited by a nonpainful stimulus) associated with several degrees of sensory loss, are typically found.^[4]

Neuropathic pain is a medical challenge as it is poorly responsive to classical peripheral antiinflammatory or powerful centrally acting analgesics, such as opioids. Moreover, the improvement after the usual recommended approaches, such as those provided by antidepressant or anticonvulsants, offer a partial reduction of symptoms in approximately only half of patients.^[2,5] Finally, the adverse-effect profile of these drugs and their social image may lead to poor levels of compliance.^[6]

Since chronic neuropathic pain can be very disabling, severe and intractable, it is associated with significant psychological and social consequences, which contribute to a reduction in quality of life. Treatment of neuropathic pain is still an unmet medical need and a better approach to treat this complex condition is required.

Together with describing what evidence-based medicine has told us regarding what to use and what not to use as first-line medication, this article will discuss what a mechanism-based approach is bringing to the clinic and how critical it is to bridge the gap between the fundamental notions regarding mechanisms generating and maintaining pain and positive clinical outcomes.

A broad Medline search with the keywords 'neuropathy', 'neuropathic pain', 'central sensitization', 'antidepressant', 'anticonvulsant', 'opioids' and 'cannabinoids' was conducted first. Systematic reviews, meta-analyses and randomized, controlled trials were initially chosen. Checking references in these studies then allowed for more information to be added.

Fundamental Mechanism-based Approach

As neuropathic pain is refractory to classical analgesics, and as our understanding of the mechanisms generating pain has noticeably increased, the idea of targeting some key molecules has emerged and a mechanism-based approach to treatment has been advocated.^[5]

A very schematic view of pain transmission is shown in Figure 1. A tissue injury leads to the activation of peripheral nociceptors, primary afferents and their central targets. This amplified peripheral excitatory signal is then modulated by central inhibitory influences. As the tissue heals, these processes extinguish themselves. When primary afferent function is impaired by an injury, a disease or lesion in the CNS, these processes can persist without ongoing signals from the nociceptors. This abnormal state is named central sensitization and is a key feature of neuropathic pain.

An easy way to consider the altered pain processing is to think in terms of an increase of the nerve cell firing from the periphery and/or a decreased inhibition of neuronal activity in central structures (Figure 1).

At a molecular level, the lesion of a peripheral nerve causes several changes, including the release of chemical substances, the upregulation of ion channel receptors and the induction of new genes, resulting in decreased thresholds or spontaneous nociceptors activity, change of cell phenotype and recruitment of silent nociceptors. This hyperexcitability then spread centrally with the phosphorylation of N-methyl-D-aspartate (NMDA) receptors, the upregulation of the specific dorsal horn sodium channels and an altered gene expression. In addition, the nerve injury produces neuroimmunological change and immune and glial cells perpetuate the pain processing within the CNS.^[7]

This hyperexcitability from the periphery is then reinforced by a loss of central inhibition, which involve monoaminergic and, possibly, opioid pathways.

Fundamental Mechanism-based Approach

Various receptors implicated in the increase of nerve firing in the periphery or in the dorsal horn have been isolated and targeted in preclinical studies. Some of these targets and their antagonist ligand compounds have shown highly positive results in these trials; on the other hand, despite being alluring, only a few of these emerge in the clinical field. These targets are illustrated in Figure 2.

Sodium (Na) channel receptors have, for a long time, been known to play a key role in peripheral neuropathic pain processing^[8] and Na channel blockers, such as carbamazepine or class I antiarrythmics, have proved useful in the treatment of some neuropathic pain conditions.^[9,10]

Na channel subtypes and their various subunits have been cloned, leading to a better understanding of their function, and the current emerging trend is to identify molecules that interact differently with the different states (open, resting or inactivated) of the channel.^[2] An example of such a new investigated Na channel blocker is lacosamide, which enhances the slow inactivation of Na channels but without altering steady-state fast inactivation.^[11] Selectively blocking the electrical activity of neurons that are chronically depolarized compared with those at more-normal resting potentials,^[11] lacosamide may have the advantage of normalizing activation thresholds and decreasing abnormal neuronal activity, thus offering better control of neuronal hyperexcitability.^[12,13] Lacosamide is currently being tested in a Phase III clinical trial in patients with DPN.^[14]

Another main channel receptor implicated in pain transmission is the calcium (Ca) channel receptor. The role of the α₂δ subunit of the Ca channel receptor in interfering with the development of central sensitization has been underlined by the success of gabapentin and pregabalin targeting this subunit specifically.^[2,5]

A third class of ionic channel receptor that has gained considerable interest is potassium (K) channel receptors. These receptors are located in the periphery after a nerve lesion, as well as in the spinal cord and the brain. Retigabine, which is a selective K channel opener, has been shown to block ectopic discharges in axotomized sensory fibers and, thereby, may be of great interest for modulating some peripheral or central neuropathic pain symptoms.^[15]

Among other potential targets, dorsal horn NMDA receptors, for which the endogenous ligand is the amino acid glutamate, play a key role in the development and propagation of pain signaling in physiological, as well as pathological, conditions. Owing to its complex nature, a large number of regulatory subunits and binding sites, the NMDA receptor provides a high number of attractive molecular targets.^[2] Ketamine is a noncompetitive antagonist of NMDA receptors that has been used in the treatment of chronic neuropathic pain for almost 15 years. Until now, clinical results are rather disappointing with a moderate effect on pain and, occassionally, unbearable side effects with acute infusion, as well as in long-term use.^[16,17] Memantine and dextrometorphan are also NMDA receptor antagonists, having only a marginal effect in practice. However, a combination of dextromethorphan and quinidine (an enzyme inhibitor that acts to reinforce dextrometorphan CNS bioavailability) is currently being studied in Phase III trials and showing promising results.^[18,19]

The next step regarding glutamate receptors is a shift to metabotropic glutamate receptor (mGlu1), whose role in analgesic activity has been demonstrated in animal models.^[20]

Modulation of Neuronal Inhibition

Currently, most research has been focused on decreasing excitability from the periphery or in the dorsal horn. How to reinforce neuronal self-inhibition is more complex to explore, for the obvious reason that the CNS is difficult to access and is, therefore, researched mainly by indirect measures. The wide use and the efficacy of antidepressants is one example of this indirect evidence, as these molecules act on the monoaminergic system, a main neurotransmitter involved in bulbospinal inhibitory pathways (Figure 2).

Another mechanism underlying central desinhibition is a loss of dorsal horn GABAergic interneurons.^[7] However, the role of the GABAergic pathways has, until recently, been poorly explored. Some results suggest that GABA agonist compounds may play a role in modifying pain transmission in animals.^[21]

The emergent question is the role of the glial cells. It is well recognized that the development of neuropathic pain also involves Schwann cells, satellite cells in the dorsal root ganglia, spinal microglia and astrocytes (Figure 2).^[22] Therefore, the reciprocal signaling pathways between neuronal and non-neuronal cells offer new opportunities for disease-modifying agents and a better management of pain.^[23]

Modulation of Neuronal Inhibition

Clinical Mechanism-based Approach

Despite major advances in the pathophysiology of neuropathic pain and the discovery of these promising targets, until now, clinical development has been slightly disappointing. A clue to this difference between preclinical observations and results in clinical trials is certainly brought by the multiplicity and the complexity of the pathophysiological mechanisms underlying neuropathic pain and the difficulty of separately testing specific hypotheses in humans. Moreover, the clinical science of translating pain complaints and sensory findings into specific clinical mechanisms that have treatment implications is in its infancy.^[5,24,25]

Indeed, one mechanism may produce various symptoms and signs; for example, upregulation of the Na channel in c-fibers leads to an increased activity and may result in constant burning pain, paroxystic pain or tactile allodynia. On the other hand, one symptom, cold allodynia for example, can be the result of various mechanisms in peripheral or central neuropathic pain.^[25]

Typically, clinical mechanism-oriented studies include fewer than 40 patients, are extremely labor intensive and require specialized equipment for testing.^[5] The use of ever-more sophisticated devices, such as imaging, to target clinical mechanisms may lead to major advances, but suffers with the disadvantage of being time, money and technology consuming. Moreover, there is an unmet need for strategies that reliably translate the results of technologically complex tools into bedside tests that link symptoms and pain mechanisms. However, even something as simple as mapping the area of allodynia in a patient with PHN requires training and practice and adds significant time to clinical trial study visits.^[5,26]

Most of the clinical trials performed have been in patients suffering from PHN and DPN. These two conditions are frequently chosen, mainly because they are prevalent conditions. In these patients, confirming diagnosis is relatively easy and most sufferers are otherwise in reasonably good health. Along with other peripheral neuropathic pain syndromes, both conditions share common clinical signs, leading to the assumption that important mechanistic similarities connect the many different types of neuropathic pain. At present, the concordance between results for DPN, PHN and other types of neuropathic pain is somewhat limited; the way forward requires a better understanding of how to clinically distinguish the manifestations of different neuropathic pain mechanisms and their importance in response to treatment.^[5,25]

Another point that is important to discuss is the interindividual variability in response to treatment. First, the psychological dimension and/or comorbidities, as well as the social repercussions of pain, are, obviously, major aspects that are not taken into account in these mechanistic trials and that play an important role in this response. Another emergent topic is the role of genetic polymorphisms in the interindividual variability in response to analgesic treatment. The interindividual variability in drug response occurs as a result of molecular difference at the level of drug-transport proteins or metabolism enzymes (pharmacokinetics) and drug targets (pharmacodynamics). At present, the impact of the pharmacogenetics on the response to treatment in neuropathic pain is poorly known, but several polymorphisms, among them the cytochrome P450 (CYP)2D6, may intervene.

Cytochrome P450 2D6

Cytochrome P450 isoenzymes metabolize the majority of xenobiotics in humans. The CYP2D6 family is involved in the metabolism of a number of drugs, including analgesics and coanalgesics (such as 'weak' opioids, codeine or tramadol), antidepressants (such as amitriptyline, clomipramine or venlafaxine) or NMDA antagonists (such as dextrometorphan). The activity of CYP2D6 is modulated by a genetic polymorphism, and four phenotypic groups have been identified: poor metabolizers (PMs; that is, a complete deletion of the gene and a potential absence of activity); intermediate metabolizers (IMs; that is, slow activity); extensive metabolizers (EMs; that is, normal activity) and ultrarapid metabolizers (UMs; that is, fast activity).^[27] The role of CYP2D6 phenotype in codeine metabolism has been well illustrated by a case report of life-threatening codeine intoxication with modest doses in a patient who was an UM and who, therefore, had a higher blood level of morphine (coming from codeine, the inactive parent molecule) than expected.^[28]

Genetic polymorphism of CYP2D6 may also play a role in the efficacy of tramadol. Tramadol (the parent molecule) has monoaminergic (noradrenergic and serotoninergic) activity but low opioid activity. However, the opioid activity of O-demethyl-tramadol (its metabolite), catalyzed by CYP2D6, is approximately 200-fold higher than the parent molecule. Therefore, monoaminergic and opioid activity of tramadol is closely correlated with the phenotypic status.^[27,29]

Regarding the response to antidepressants in patients treated for depression, the polymorphism of CYP2D6 influences the occurrence of adverse effects (PM people tended to develop more adverse effects than UM people), as well as the failure of treatment (UM people were more refractory to treatment) in a pilot study.^[30] One may, therefore, assume that the CYP2D6 polymorphism could also play a role when these molecules are administered for pain relief.

Evidence-based Approach

In addition to the mechanism-oriented trials, randomized controlled studies and meta-analyses have allowed the drawing up of evidence-based recommendations for the treatment of neuropathic pain. According to the recent guidelines on pharmacological treatment of neuropathic pain, antidepressants, anticonvulsants and opioids have been found to be efficient with a high level of evidence.^[31,32] This section will discuss the new recommendations within these groups, as well as briefly comment on the place of cannabinoids.

summarizes the main values for efficacy and the main adverse effects of the drugs discussed in the session^[33] (for a detailed review on efficacy and safety according to different etiologies, see elsewhere^[34]).

Antidepressants

Among pharmacological agents, antidepressants are the most studied and the most used in the management of neuropathic pain. Multiple mechanisms of action have been advocated, but they mainly act by reinforcing descending inhibitory pathways, although they have Na channel blocker properties at higher doses than theose usually recommended.

The effectiveness of tricyclic antidepressants (TCAs) has been demonstrated repeatedly in several randomized controlled trials, as has been recently reviewed in a meta-analysis.^[35] The number needed to treat to obtain one patient with more than 50% of pain relief (NNT) is 3.1 (95% CI: 2.7–3.7).^[33] The limiting factors to their use are, however, their tolerability and safety, particularly in elderly people. Dizziness, sedation, orthostatic hypotension, dry mouth and constipation are common side effects of TCAs that may cause withdrawal. Moreover, they are contraindicated in patients with glaucoma, prostatic hypertrophy or some cardiac conduction disturbances. Hence, there is growing interest in the serotonin- and noradrenalin-reuptake inhibitors (SNRIs), such as venlafaxine and, more recently, duloxetine. Venlafaxine has an NNT of 3.1 (95% CI: 2.2–5.1) for global improvement, as defined in a Cochrane systematic review.^[35] The efficiency of duloxetine has been demonstrated in three randomized controlled trials in patients with DPN.^[36-38] Duloxetine 60 mg once daily improved the 24-h average pain severity score on an 11-point visual analogue scale (VAS) by approximately 3 points in a 12-week randomized controlled trial. The onset of action was rapid, with a separation from placebo beginning at week 1.^[39] The administration of 60 mg twice daily did not add further improvement.^[38] The main adverse effects were nausea (28%), dizziness (15%), fatigue (14%) and diarrhea (11%).^[39] In a post hoc analysis that pooled the data of the three randomized, controlled trials, the NNT of having 50% or more reduction from baseline in a mean 24-h VAS was 5.2 (95% CI: 3.8–8.3) for 60 mg once daily and 4.9 (95% CI: 3.6–7.6) for 60 mg twice daily. The number needed to harm (NNH) based on discontinuation for an adverse effect was 17.5 (95% CI: 10.2–58.8) for 60 mg once daily and 8.8 (95% CI: 6.3–14.7) for 60 mg twice daily.^[40] In the same population of patients, an open-label 52-week extension study showed a significant difference in favor of duloxetine compared with a routine-care group in the SF-36 physical-component summary score and subscale scores of physical functioning, bodily pain, mental health and vitality. There were no significant therapy group differences observed for patients with at least one serious adverse event.^[39]

Considering the indisputable place of antidepressants in the pharmacotherapy of neuropathic pain, these positive results in terms of efficacy, as well as in terms of tolerability and safety, are encouraging, and there is a growing place for SNRIs in the treatment of neuropathic pain.

Anticonvulsants

Anticonvulsant drugs are the second group of well-studied pharmacological agents for the treatment of neuropathic pain and have been used since the 1960s. Among them, gabapentin and pregabalin, which preferentially act on the α₂δ subunit of the presynaptic Ca subunit channels, are currently the most widely used and have been studied in large randomized controlled trials in peripheral and central neuropathic pain conditions. They are similar to older anticonvulsants, such as carabamazepine, which has a NNT in trigeminal neuralgia of 2.5 (95% CI: 2.0–3.4)^[9] or phenytoine, which has a NNT in diabetic neuropathy of 2.1 (95% CI: 1.5–3.6).^[41] Gabapentin and pregabalin have shown efficacy in neuropathic pain. Moreover, they have a good tolerability and a safe long-term-use profile. However, as with older anticonvulsants, their adverse effects may include impaired mental function, which may limit their clinical use, particularly in the elderly.^[42] The efficacy of gabapentin has been reviewed recently in a meta-analysis.^[42] In chronic neuropathic pain, the NNT for a significant improvement in all trials with evaluable data was 4.3 (95% CI: 3.5–5.7). More specifically, the NNT for effective pain relief in diabetic neuropathy was 2.9 (95% CI 2.2–4.3) and was 3.9 for postherpetic neuralgia (95% CI: 3–5.7).^[42] Based on large randomized controlled trials in a broad range of conditions, the magnitude of effect of pregabalin is similar.^{[1,31,32,43-45]}.

The main differences between these two molecules and the main advantage brought by the new pregabalin lie with the pharmacokinetic profile of these two compounds.^[1] Gabapentin has a nonlinear and dose-dependent profile of absorption, which brings a high intersubject variability to the absolute bioavailability.^[46] In practice, gabapentin needs slow individual titration; some patients may have effects with low doses, whereas others need higher doses, which increases the time needed to establish whether a patient is a responder or not. By contrast, pregabalin has a linear pharmacokinetic profile, with a bioavailability of 90%, which makes its effect far more predictable. Furthermore, the onset of the pain-relieving action of pregabalin is quicker; it has significant differences compared with placebo when taken for less than 1 week.^[47]

Of note, pregabalin, as mentioned with gabapentin before, has been proved efficient in controlling central neuropathic pain in a recent randomized controlled trial in patients suffering from spinal cord injury.^[48] As in peripheral neuropathic pain, pregabalin improved pain and sleep with a good profile of tolerability.^[48]

Regarding other new anticonvulsants, such as lamotrigine or topiramate, results are not convincing. The efficacy of lamotrigine has been reviewed in a recent meta-analysis and the conclusion is that 'the limited evidence currently available suggests that lamotrigine is unlikely to be of benefit for the treatment of neuropathic pain'.^[49] Given the risk of potentially serious cutaneous reaction and the availability of more effective treatments, lamotrigine does not have a significant place in therapy at present.^[49] Despite its interesting action on a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid/kainate receptors, which are glutamate receptors, topiramate also failed to demonstrate a significant effect in DPN.^[50,51]

Opioids

From a clinical point of view, neuropathic pain states were considered resistant to opioid analgesia. This clinical observation may be correlated with pathophysiological mechanisms after nerve injuries, such as the sprouting of large calibre A fibers in the superficial layer of the dorsal horn,^[52] decrease in expression of µ-receptors within the spinal cord and within the brain^[53] or modulation of the opioid's analgesic activity by cholecystokinin, an endogenous peptide with pronociceptive activity.^[54] Moreover, a longitudinal study looking at the dose–response relationship of buprenorphine in nociceptive (soon after thoracic surgery) and neuropathic (1-month post-thoracotomy) pain found a shift to the right of the dose–response curve, meaning that the analgesic dose needed to reduce neuropathic pain was significantly higher than the one needed to reduce nociceptive pain.^[55]

In contrast with this notion that drugs acting on the opioid µ-receptor are poorly effective, and according to US and European guidelines,^[31,32] weak and strong opioids are, nowadays, among the range of treatments of neuropathic pain.

The NNT of opioids is 2.5 (95% CI: 1.9–3.4).^[33] However, recent meta-analysis showed that, in general, opioids used for a mean time of 28 days decreased the mean VAS score of 13 points (on a scale from 0 to 100) compared with baseline (95% CI: -16 to -9; p < 0.00001).^[56-58] Usual adverse effects, such as nausea (NNH: 4.2), constipation (NNH: 4.2), drowsiness (NNH: 6.2) and dizziness (NNH: 7.1), were threefold more frequent in the treatment group than in the placebo group. In total, 11% of the patients withdrew because of an adverse event in the active group compared with 4% in the placebo group.^[56-58] However, these positive results should be moderated by some questions that remain to be answered. Is this improvement clinically meaningful and sufficient to balance the burden of the adverse effects? It is interesting to note that in the trials that had positive results regarding pain and that, surprisingly, produced so few side effects, the results on the overall quality of life were weak or not significant.^[1,56] Finally, long-term risks, such as development of tolerance or even addiction, have not been sufficiently assessed in neuropathic pain patients.

Among opioids, special attention should be paid to a few molecules that may have added properties that are specifically beneficial in neuropathic pain. The first is tramadol, a compound with a weak µ-opioid effect but also weak monoaminergic effect, which could account for its efficacy. Compared with placebo, tramadol showed a significant benefit in neuropathic pain, with a NNT of 3.9 (95% CI: 2.7–6.7).^[33] However, data were insufficient to draw conclusions about the effectiveness of tramadol compared with either clomipramine or morphine.^[59]

Oxycodone is a second compound whose specific ability to relieve neuropathic pain is currently being discussed. Two randomized controlled trials in patients suffering from DPN reported a significant effect on total pain, mean daily pain and disability, compared with placebo.^[60,61] In one randomized controlled trial, the combined NNT was 2.6 (95% CI: 1.9–3.4) and the quality of life was also improved^[61]. The apparent superiority of oxycodone compared with morphine in neuropathic pain, as well as whether and how oxycodone is different from morphine, is still largely debated.^[62,63] A series of studies on rodents argued that oxycodone is a κ-opioid receptor agonist and proposed that this property can explain the discrepancy between the weak µ-opioid receptor affinity of oxycodone and its good clinical efficacy compared with morphine. Another explanation could be found in the pharmacokinetic properties of oxycodone. Oxycodone and morphine do not share the same metabolic pathways and the contribution to analgesia of oxycodone and oxymorphone, its metabolite, is not clearly identified, complicating comparisons with morphine and its metabolite, m-6-glucuronide. Another major postulated difference between oxycodone and morphine can be found in their passage through the blood–brain barrier (BBB). Both drugs are equally hydrophilic but oxycodone can be actively transported through the BBB by a transporter that has not yet been identified.^[62] Nevertheless, whether oxycodone is better than morphine in relieving neuropathic pain needs to be studied in head-to-head comparisons in controlled clinical trials.^[62]

Methadone and its NMDA receptor-antagonism properties also has a good theoretical potential in the treatment of neuropathic pain. It has been increasingly used to manage neuropathic pain but the data on its efficacy are scarce.^[64,65] Moreover, some of its properties, such as the increase of the QT interval, raise concerns about its safety and, according to the US National Center for Health Statistics, methadone was involved in more overdose deaths than any other prescription drug.^[66,67] Therefore, methadone has been quoted as a fourth-line treatment in recent guidelines.^[1,65]

Cannabinoids

The potential role of cannabinoid agents in the management of neuropathic pain has attracted considerable interest. The discovery of cannabinoid receptors 1 and 2 and the development of specific cannabinoid receptor agonist and antagonist ligands, as well as encouraging results from preclinical studies, point to a role of cannabinoids as a therapeutic modality. However, although animal work continues to suggest that cannabinoids may be useful for neuropathic pain, results in clinical studies have always been controversial. A few years ago, a meta-analysis examining cannabinoids failed to find convincing evidence of analgesic activity beyond that of weak opioids.^[68] By contrast, the following randomized controlled trials published on the subject, showed that cannabinoid δ-9-tetrahydrocarnabinol/cannabidiol oromucosal spray^[69,70] or dronabinol taken by mouth,^[71] produced a significant decrease in the mean intensity of pain and in sleep disturbance in patients with central neuropathic pain due to multiple sclerosis, as well as in patients with chronic neuropathic pain of mixed origin. These results were replicated in an uncontrolled open-label extension study, which confirmed the benefit at 2-year follow-up of the oromucosal spray.^[72] However, the important issue with the use of cannabinoids is adverse effects. In the two studies using oromucosal spray, the proportion of patients experiencing adverse events were 89^[69] and 91%,^[70] and 18% of patients withdrew owing to an adverse effect.^[70] The proportion of adverse events was 96% for dronabinol use and 17% of patients had their doses reduced for intolerable adverse events.^[71] In the three studies, the main adverse events were gastrointestinal and neuropsychological. Furthermore, driving regulation and legislation might limit cannabinoid utility in some European countries.

Finally, a recent randomized controlled trial compared the effect of nabilone, an oral synthetic cannabidoid, with dihydrocodeine in patients with chronic neuropathic pain of mixed origin and found that dihydrocodeine provided even better pain relief than nabilone and had slightly fewer side effects, although no major adverse events occurred for either drug.^[73]

In summary, despite a putative role in central neuropathic pain associated with multiple sclerosis,^[32] cannabinoids should not be considered as a first option in the treatment of neuropathic pain because of a debatable benefit–risk profile. The association between long-term use and precipitation of psychosis or schizophrenia is of particular concern.^[32]

Combination Medications

The available drugs to treat neuropathic pain have incomplete efficacy and dose-limiting adverse effects. As a matter of fact, the maximal tolerated doses of drugs that are considered as a first-line treatment when administered as single agents reduce pain by only 30%^[1] owing to incomplete efficacy, dose-limiting adverse effects, or both. The combination of mechanistically distinct analgesic agents may result in additive effects or synergism and may improve efficacy at lower doses, with fewer expected side effects than with the use of one agent alone. This strategy has been advocated in cases of partial treatment response, but rigorous supportive evidence is limited.

The first major trial in this field was a randomized, controlled trial studying the combination of morphine and gabapentin in patients suffering from DPN or PHN.^[74] The combination achieved better analgesia at lower doses of each drug than using either as a single agent. Mean daily pain VAS at the maximal-tolerated dose of each option was 5.72 at baseline, 4.49 with placebo, 4.15 with gabapentin, 3.70 with morphine and 3.06 with the gabapentin–morphine combination (p < 0.05) for the combination versus placebo, gabapentin and morphine. Constipation, sedation and dry mouth were the most frequent adverse effects. The lower tolerated dose of morphine and gabapentine in combination suggested an additive interaction. This study was not designed to solve the question of the mechanism of this interaction. However, an increase in gabapentine absorption may be a clue, as suggested by the difference in the frequency of constipation in the gapapentin alone versus the combination group, as well as the result of a pharmacokinetic study in healthy volunteers showing an increase in gabapentine plasma concentration when associated with morphine.^[75]

These positive results with a combination of gabapentine and an opioid have been replicated more recently in a greater number of patients suffering from DPN and using oxycodone.^[76] The overall treatment effect was greater with oxycodone–gabapentin than with placebo–gabapentine, the combination was associated with less rescue-medication use and fewer nights of disturbed sleep. Discontinuations owing to lack of therapeutic effect were lower with oxycodone–gabapentin. The commonly seen opiate-induced adverse events were not exacerbated by the combination of oxycodone and gabapentin.^[76]

Another combination that is currently being tested in Phase III studies in DPN patients is dextromethophan and the enzymatic inhibitor quinidine. This pharmacokinetic interaction increases dextromethophan bioavailability and, therefore, its potential efficacy. Preliminary results presented in two posters showed that the combination was superior to placebo in relieving pain, with a decrease from baseline VAS of approximately 60% after 2 months of treatment. Moreover, the quality of life improved in the combination group, being twofold better than in the placebo group. Adverse effects occurred in 90% of patient in the active group and were mainly dizziness, nausea, diarrhoea, fatigue and somnolence. There were no statistically significant differences in serious adverse event rates and in cardiovascular adverse effects between the active treatment and placebo groups.^[18,19]

Expert Commentary

New approaches to the pharmacotherapy of neuropathic pain embrace different lines of work, including a fundamental mechanism-based approach, a clinical mechanism-based approach and an evidence-based approach. These three visions are indisputably complementary to draw a more comprehensive picture of neuropathic pain that will lead to better care of the patients suffering from this syndrome.

The main challenge is to link these fundamental mechanisms to specific clinical symptoms and signs, as well as to sensory findings, which is a necessary step to understand their importance in response to treatment. Another decisive element is the role of the interindividual variability, which could be related to environmental as well as genetic reasons. As for genetic variability, screening for the main polymorphisms playing a role in the action of analgesic drugs may be a simple and reliable tool to discriminate responders from nonresponders. Moreover, this approach may allow for a better detection of patients who are at risk of developing major side effects.

On the other hand, the evidence-based approach is essential as it implicates the testing of drugs in large patient populations. This aspect is of great importance as it diminishes hazard bias and allows for some degree of generalization. However, as the patients participating in such trials are selected on clinical criteria, this generalization is also limited and, instead of establishing broad indications to the whole neuropathic syndrome from results obtained in one disorder (DPN or PHN, for example), efforts should also be made to test the different molecules in different diseases.

Antidepressants and anticonvulsants continue to be the first line treatment for the prevention of neuropathic pain. In these groups of substances, the trend is to seek compounds with a high specificity for a type of receptor or, even, for a particular subtype of a type of receptor, which leads to greater efficacy and fewer side effects. In contrast to old notions, opioids now have a place in the range of treatments, but as second- rather than first-line agents. Other options, such as NMDA receptor antagonists, should be tried as third-line treatments when other treatment options have failed. This should also be the case for cannabinoids, except, perhaps, in the treatment of central pain due to multiple sclerosis.

A link between mechanism-based and evidence-based methods is illustrated by the empiric use of several substances with a different mechanism of action in the same patient. This approach is currently being tested formally in randomized controlled trials and results are encouraging in terms of efficacy, as well as in terms of tolerability and safety.

Five-year View

From a mechanism point of view, at a molecular level, together with pursuing the work on potentially new molecules acting on the already identified targets, the role of the glial cells in pain processing will be characterized more precisely. At a clinical level, the use of sophisticated technology, such as imaging, to determine whether fundamental mechanisms are operative in patients, will increase. Ideally, progress will be made to find strategies for reliably translating these results into bedside tests that link symptoms and pain mechanisms.

To define drug response more precisely, our knowledge of genetic polymorphisms with potential importance for pain therapy will improve, and screening for them before initiating therapy may become a widely used approach. Finally, designs of randomized controlled trials should involve a larger range of diseases than only DPN and PHN and may include genetic screening, as well as several mechanistic aspects, tested with specific tools or indirectly provided by the use of combinations of substances.

Reprint Address

Jules Desmeules, Division of Clinical Pharmacology and Toxicology, Geneva University Hospital, Rue Micheli du Crest 24, 1211 Geneva 14, Switzerland; E-mail: jules.desmeules@hcuge.ch

Expert Rev Clin Pharmacol. 2008;1(5):683-693. © 2008  Expert Reviews Ltd.

Cite this: New Approaches to the Pharmacotherapy of Neuropathic Pain - Medscape - Sep 01, 2008.