Pharmacological Therapy for Neuropathic Pain

Introduction

Neuropathic pain (NP) has been a focus of intense research over the past 3–4 decades, which has resulted in the elucidation of many mechanisms; unfortunately, not many have translated into successful therapies. NP refers to a group of pain disorders, some of which are readily recognized, such as post-herpetic neuralgia (PHN), whereas a large number of others are less well defined. All of them have one fundamental characteristic: they are the result of injury, disease, or trauma involving the somatosensory nervous system ( ), and consequently, NP is in general a chronic pain disorder ( ). Similar to all chronic disorders, there is no cure for NP, so the primary focus of this review is on disease management and symptomatic pain relief. Optimal treatment of chronic NP requires comprehensive evaluation and a multimodal and multidisciplinary approach to therapy, which is addressed in this textbook. Although a large number of controlled clinical trials have been conducted, only a few were successful to the point of providing sufficient evidence for regulatory approval of a very limited number of new medications. A number of reasons were suspected for this failure to translate the advances from pain research into successful therapies, and this trend was considered unsustainable to the point that the Food and Drug Administration (FDA) in the United States supported the development of public–private partnerships whose goal is to investigate and improve the status of pain clinical trials ( ).

Currently available medications for symptomatic treatment of NP are traditionally grouped into four categories—anticonvulsants, antidepressants, opioids, and topical agents—and these categories are a reflection of either the conditions for which these treatments were first developed and then borrowed for NP therapy, such as anticonvulsants and antidepressants, or older therapies, such as opioids or topical agents, rather than therapy that was built from the ground up based on mechanisms of action. All these treatments have limitations, and thus far they have demonstrated similar results, which is that half the patients treated with any single drug obtain “moderate” or better pain relief ( , ). In the absence of comparator studies that would test and provide information for the best treatment approach, a number of treatment algorithms and guidelines have been developed internationally, and most of them state that they start with published evidence from controlled trials, although interpretation and recommendations are as a rule based on the consensus of a selected group of experts. The lack of validated tools for predicting the efficacy of specific treatments in individual patients is another difficulty in selecting the most appropriate therapy in the clinical setting.

In this chapter we start with the mechanisms underlying NP and its complexity, with consideration of the implications of these mechanisms on drug treatments; we then review assessment of NP, with consideration of the implications of such assessment on selection of drug therapy, as well as current therapies across traditional pharmacological categories on the basis of clinical trials; and we finish with a summary of recommendations and a review of future directions in the development of pharmacotherapy for NP.

Mechanisms Underlying Neuropathic Pain and their Implications for Treatment

The list of peripheral and central nervous system (PNS, CNS) diseases that can result in acute or chronic NP is lengthy ( Box 70-1 ), but multiple controlled clinical trials are available for only a minority of the conditions. For purposes of treatment, there may not be much difference among any of these syndromes except their PNS or CNS origin, although etiological diagnoses such as human immunodeficiency virus (HIV)-associated painful neuropathy and chemotherapy-induced neuropathy do not respond to some of the agents such as tricyclic antidepressants (TCAs) or α ₂ δ calcium channel–modulating agents ( , ). Traditionally, clinical trials have been directed toward a specific disease entity rather than specific signs or symptoms, with the majority of studies being performed on patients with PHN and painful diabetic neuropathy (PDN). Management of patients based on their disease state, however, is probably not the most effective way to address specific NP signs and symptoms ( ).

The neurobiological and clinical concept has been termed mechanism-based and symptom-based classification and clearly differs from the current classifications of pain based on disease, duration, and location ( , ). By identifying in individual patients which neural mechanisms are responsible for their pain, individualized treatment specifically targeting these mechanisms might be used. This shift in paradigm will require that we have ways of ascertaining which mechanisms are responsible for an individual patient’s pain ( , ). It is more than likely that any individual patient has NP as a result of many underlying mechanisms. However, there may be a dominant mechanism that when treated, reduces pain to minimal levels. For example, if it can be demonstrated that ectopic impulse generators, caused by abnormal sodium channel activity, located in injured or abnormally functioning primary afferent fibers are generating increased traffic entering CNS pain pathways, treatment with a sodium channel–blocking agent that reduces ectopic firing may dramatically reduce the pain.

Caveats come from clinical studies addressing the predictive value of clinical symptoms to guide treatment. It was hypothesized that topically applied lidocaine, which is believed to act on ectopic discharges in nociceptive fibers, would in particular be beneficial for patients with sensitized peripheral nociceptors as compared with patients with loss of dermal nociceptors. In contrast to this hypothesis, however, skin biopsy, quantitative sensory testing (QST), the histamine test, and nerve conduction studies could not be used to identify lidocaine responders with painful neuropathies ( ) and PHN ( ). Alternatively, it might be possible that the surviving A fibers in C-nociceptor–deprived skin may express sodium channels, develop ectopic firing, and therefore potentially be the target for lidocaine.

However, many more attempts are being made to develop treatments that specifically target the presumed underlying mechanisms of NP. These include reducing release of transmitters in pro-nociceptive neurons by opiates or α ₂ δ calcium channel–binding drugs, inhibiting post-synaptic excitatory receptors such as the N -methyl- d -aspartate (NMDA) or α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/kainate receptors, and potentiating inhibitory transmitters by reducing transmitter uptake or administering an agonist or use-dependent sodium channel blocker. Many analgesic targets are re-regulated after nerve injury, including sodium and potassium channels, the α ₂ δ calcium channel accessory protein, and cannabinoid and opiate receptors, although how this contributes to NP and interferes with analgesic therapy in humans has not been established ( ).

Assessment of Neuropathic Pain and its Implications for Treatment

Assessment of the various components of NP and associated symptoms is needed for diagnosis and to guide therapy. No single symptom or sign is pathognomonic for NP. An etiological diagnosis has to be considered and evidence of neurological injury has to be established. NP is the result of disease or injury to the nervous system, and consequently, the clinical manifestations typically include both negative and positive sensory symptoms and signs ( ). Careful physical and neurological examination can help answer the basic neurological question of “where the lesion is” and assess non-neuropathic factors—musculoskeletal, inflammatory, myofascial, and psychological—that contribute to the particular characteristics of the patient’s pain ( ).

A sophisticated technique to test both the nociceptive and non-nociceptive systems in the periphery and the CNS is QST, which uses standardized mechanical and thermal stimuli (graded von Frey hairs, several pinprick stimuli, pressure algometers, quantitative thermal testing). Recently, QST data from a large cohort of patients with a variety of peripheral and central NP states showed that every single positive or negative sensory abnormality occurred in each neurological pain syndrome, but with different frequencies; for example, thermal and mechanical hyperalgesia was most prominent in patients with complex regional pain syndrome (CRPS) and peripheral nerve injury and allodynia in those with PHN. To display relevant combinations of sensory abnormalities, a coding system was applied that was essentially based on the development of “loss” of function and “gain” of function (so-called LOGA classification). For this purpose, signs of hypoesthesia after thermal stimuli (loss of detection in the cold pain threshold [CPT] or heat pain threshold [HPT]) were coded as L1 and signs of hypoesthesia to mechanical stimuli (loss of detection in the mechanical or vibration detection threshold) as L2. Signs of hyperalgesia to thermal stimuli (gain of function in HPT or CPT) were coded as G1 and signs of hyperalgesia to mechanical stimuli as G2. When both thermal and mechanical abnormalities were present, L3 or G3 was defined, respectively. This resulted in a range of variation from L0G0 to L3G3. The most frequent combinations of gain and loss were mixed thermal/mechanical loss without hyperalgesia (central pain and polyneuropathy), mixed loss with mechanical hyperalgesia in peripheral neuropathies, and mechanical hyperalgesia without any loss in trigeminal neuralgia (TN) ( ). Currently, some of the newly initiated clinical trials are making use of this classification to investigate the value of subgrouping patients into those with and without loss and gain of function in their somatosensory system.

Clinical analysis with QST is one approach to find a rationale for stratification of patients to individual treatment, but use of NP questionnaires is another, even simpler way to classify patients based on individual perception of their pain components. Pain is an individual-specific phenomenon that is described with patient-specific symptoms and expressed with certain intensity. Therefore, there is a strong rationale to examine the value of verbal descriptors and pain qualities as a basis for distinguishing NP from other types of chronic pain. In the past 5 years, much research has been undertaken to develop screening tools for this purpose. These tools have been developed with the goal of assisting in the initial diagnosis and assessment of NP ( , , ) and are based on verbal pain description with or without limited bedside testing ( ).

Cluster analysis of patients’ pain descriptions with PainDETECT revealed five typical features of description within groups of patients with PDN, PHN ( ), fibromyalgia syndrome ( ), and radiculopathy ( ). The data show that there is considerable overlap and mismatch in the verbal description of pain independent of the existing disease condition. This suggests that besides QST phenotyping, verbal description with NP questionnaires also allows subgrouping of patients according to their perceived symptoms and that this could be a strategy to stratify patients to different treatment regimens ( , ). Although this, as well as clustering with QST, has not been proved to increase outcome or responder rates in clinical trials, at the moment this looks to be the most favorable way to delineate more specific treatments. A first step in this direction has been undertaken in patients with low back pain and associated leg pain ( ). In this study the scores on a pain questionnaire were useful as a prognostic factor but were not predictive of response to treatment that was not targeted at NP.

Imaging and electromyographic (EMG)/nerve conduction studies are certainly an integral part of diagnosing and assessing spinal pain, although imaging studies may be negative or even misleading. It should be noted that EMG/nerve conduction studies are insensitive to abnormalities in small-diameter sensory fibers. Quantitative thermal sensory testing relies on the patient’s psychophysical ability to discriminate fine changes in thermal stimuli but is not widely used because it requires specialized equipment and training. Functional magnetic resonance imaging can assess pain-related brain structures, but its role in clinical practice will remain limited in the near future. A comprehensive approach based on the history, physical and neurological examination, and diagnostic studies can establish the pain diagnosis in most cases.

Clinical Trial Evidence and Development of Treament Recommendations for Neuropathic Pain

Pharmacotherapy for NP was one of the pioneering areas in clinical trials of pain in general, and it was the basis for establishing the efficacy of TCAs in relieving NP in PHN ( ) and in PDN ( ) starting 3 decades ago. Many of these trials were of crossover design. It was approximately a decade and a half ago that larger clinical trials using a parallel design with novel antiepileptic drugs built on the concept that controlling hyperexcitability with this class of drugs would be the shared mechanisms for efficacy in controlling epilepsy and NP, and the first oral medication, gabapentin, was approved for PHN based on these studies. Other controlled trials continued to be conducted, and the majority still involved patients with PHN and PDN, which led to FDA approval of pregabalin for PHN ( ) and PDN ( , ), duloxetine for PDN ( , ), topical lidocaine for PHN ( ), and high-concentration capsaicin for PHN and HIV-associated distal sensory polyneuropathy ( , ). Approval of any drug is highly variable across the globe; for example, approval of a drug in the United States does not have equivalent approval in Europe or other countries, and vice versa. The majority of the evidence base of controlled trials comes from examination of only two chronic NP syndromes, PDN and PHN. The applicability of clinical trial results from one chronic neuropathic syndrome to others cannot be determined because no data exist to support or dispute such applicability. Even more basic challenges stemming from diagnostic difficulty are illustrated by the example of CRPS type 1, in which controlled trials of analgesic pharmacotherapy are lacking and the chronic pain is believed to be due to nervous system dysfunction without permanent injury to a nerve trunk ( , ). Similar issues face chronic neuropathic back pain and cervical and lumbar radiculopathic pain, probably the most prevalent chronic pain syndrome with a neuropathic component. There are no accepted diagnostic criteria for separating out the neuropathic component, but a combination of neuropathic, skeletal, and myofascial mechanisms probably account for most of the pain.

The ongoing difficulty in objectively assessing the efficacy of pharmacological agents is that studies conducted over the past decade and a half have been reported only as posters or platform presentations at meetings and never published as a peer-reviewed article. Therefore, the lack of access to the full extent of studies conducted with specific agents does not allow rigorous meta-analysis even though similar outcome measures and data analysis schemes have been used. For example, topiramate, a marketed anticonvulsant that acts on sodium channels and reduces release of excitatory amino acid neurotransmitters, has been tested in four placebo-controlled trials for PDN totaling more than 1200 subjects. The first three studies did not demonstrate any analgesic efficacy of topiramate for NP ( ), but the fourth showed modest efficacy ( ). None of the studies have been published in full format in the peer-reviewed literature. A few of the smaller studies with experimental compounds that failed to show analgesia have been published in full format, such as trials of a sodium channel blocker and an NMDA–glycine site antagonist ( ). “Negative” trials of other experimental compounds no longer in development may never be published.

A number of individuals, societies, and academic institutions have conducted systematic reviews, and meta-analyses have been published over the past decade with the goal of estimating the relative efficacy of different drugs for NP by assessing results from controlled trials ( ; ). The concept of number needed to treat (NNT) has been introduced as a tool for comparison of different agents in the absence of comparator studies. In the field of NP a significant issue is that nearly all studies published before the mid-1990s are small single-center trials. NNTs calculated from these trials need to be interpreted with caution.

Recently, a systematic review of the effects of treatment of PDN ( ) found that treatment with carbamazepine, lamotrigine, or sodium valproate was favorable, with a total odds ratio (OR) of 5.33 and a withdrawal-related adverse event OR of 1.51, and that treatment with gabapentin or pregabalin was also favorable, with a total OR of 3.25 and a withdrawal-related OR of 2.98. Comparison of TCA with duloxetine revealed almost the same relationship. Although older, single-center small studies benefit from higher rates of response to treatment, little or no response to placebo, and very little withdrawal, recently performed clinical studies show deficits in all these criteria. Consequently, the authors found that anticonvulsants and antidepressants are still the most commonly used options for managing PDN, and based on their systematic review they conclude that oral TCAs and traditional anticonvulsants are better than the newer-generation anticonvulsants for short-term pain relief. Evidence of long-term efficacy of oral antidepressants and anticonvulsants is still lacking ( ). The latter statement holds true for older drugs, whereas more long-term data exist for modern drugs.