OUP user menu

Multimodal Analgesia for Chronic Pain: Rationale and Future Directions

Charles E. Argoff MD, Phillip Albrecht PhD, Gordon Irving MD, Frank Rice PhD
DOI: http://dx.doi.org/10.1111/j.1526-4637.2009.00669.x S53-S66 First published online: 1 July 2009


Chronic pain is a multifaceted disease requiring multimodal treatment. Clinicians routinely employ various combinations of pharmacologic, interventional, cognitive–behavioral, rehabilitative, and other nonmedical therapies despite the paucity of robust evidence in support of such an approach. Therapies are selected consistent with the biopsychosocial model of chronic pain, reflecting the subjective nature of the pain complaint, and the myriad stressors that shape it. Elucidating mechanisms that govern normal sensation in the periphery has provided insights into the biochemical, molecular, and neuroanatomic correlates of chronic pain, an understanding of which is leading increasingly to mechanism-specific multidrug therapies. Peripheral and central neuroplastic reorganization underlying the disease of chronic pain is influenced by patient-specific emotions, cognition, and memories, further impairing function and idiosyncratically defining the illness of chronic pain. Clinical perceptions of these and related subjective elements associated with the suffering of chronic pain drive psychosocial treatments, including, among other options, relaxation therapies, coping skills development, and cognitive–behavioral therapy. Treatment selection is thus guided by comprehensive assessment of the phenomenology and inferred pathophysiology of the pain syndrome; patient goals, preferences, and expectations; behavioral, cognitive, and physical function; and level of risk. Experiential, practice-based evidence may be necessary for improving patient care, but it is insufficient; certainly, well-designed studies are needed to support therapeutic decision making. This review will discuss the biochemical basis of pain, factors that govern its severity and chronicity, and foundational elements for current and emerging multimodal treatment strategies.

  • Chronic Pain
  • Multidrug Therapy
  • Multimodal Analgesia


Chronic pain is a complex disease requiring multidimensional assessment and multimodal treatment, while insights into the neurobiology of nociception continue to influence our understanding and management of numerous debilitating pain syndromes [1]. In chronic pain, dysregulation of one or many locations within the complex nociceptive and pain neural circuit leads to neuroplastic reorganization and increased spontaneous activity (spontaneous pain), as well as hyperresponsiveness to noxious (hyperalgesia) and non-noxious (allodynia) stimuli [2]. The neuroanatomy of the nociceptive loop is comprised grossly of nerve endings of peripheral primary sensory neurons, the spinal dorsal horn, cranial sensory ganglia, supraspinal nuclei, the thalamus, limbic and cerebral cortices, diencephalic processing, and supraspinal descending modulatory control of the initial peripheral input at the spinal dorsal horn level [3]. Together, these neural structures create a modulated afferent warning system receiving information from the peripheral nervous system (including interoception) that is processed at multiple levels of the neuraxis to the cortex, the seat of human perceptions of pain (Figure 1). The nature of nociception demands immediate neural (and, at times, cognitive) attention; at each level of processing along the nociceptive loop, therefore, efferent (outward from the central nervous system [CNS]) responses are generated, including, for instance, adaptive reflexive withdrawal and nociceptive/acute pain behaviors, but which also include the associated maladaptive sensations and behaviors of chronic pain. Despite advances in our scientific understanding of the salient events that govern nociceptive signaling and their aberrations in chronic pain, translational efforts in the development of multimodal therapeutic strategies are still lagging. Individualizing patient care in part depends on operationalizing the biopsychosocial phenomenology of chronic pain, which resides at the intersection of genetics, biochemistry, physiology, culture, cognition, behavior, and perception [1]. Therefore, multimodal therapeutic plans require a “multivisit” appraisal across the phenomenology and inferred pathophysiology of the pain complaint to assess functional impairment, the patient's goals and expectations, psychosocial needs and stressors, and the level of risk of nonmedical medication use. Careful integration of these assessments may inform appropriate treatment selection of both pharmacologic and nonpharmacologic modalities. Opioid and non-opioid analgesics, when used alone or in combination and when employed as components of a multimodal regimen, often may lead to effective pain control, including improved function in certain patients' outcomes. Structuring opioid-based multimodal strategies that are consistent with the level of risk of nonmedical use of medications remains centrally important to improving patient care. Methodologically sound clinical trials are required to define combination practices with medical and nonmedical therapies that will maximize analgesia while minimizing the risks to the patient, prescriber, and public. In the absence of rigorous evidence-based guidance, clinicians rely on their judgment, cumulative experience, and practice-based evidence.

Figure 1

The neural loop and site of action of various analgesics. AMPA = alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate; CR = controlled release; DRG = dorsal root ganglion; GABA = γ-aminobutyric acid; NMDA = N-methyl-D-aspartate; SNRI = serotonin-norepinephrine reuptake inhibitor; SSRI = selective serotonin reuptake inhibitor; TCA = tricyclic antidepressant.

Normal Sensation in Peripheral Tissue

Mechanisms regulating peripheral sensation play a critical role in the manifestation of chronic pain [4]. Within the periphery, several classes of primary sensory neurons with somata located among the dorsal root or cranial sensory ganglia have been described and respond to noxiousand non-noxious mechanical (mechanoreceptors), thermal (thermoreceptors), or chemical (chemoreceptors) stimuli [5–8]. Each primary sensory axon ending in body tissues is associated with a characteristic end structure, activation threshold, and duration/adaptation of response to a preferred stimulus (i.e., mechanical, chemical, thermal, and electromagnetic). The unique electrophysiologic response properties of primary sensory neurons are, in part, due to axon caliber, unique receptor-ending morphologies and biochemistries, and sites of termination within the periphery [5–8]. The four broadly defined primary sensory afferent types include the following:

  1. The largest caliber, heavily myelinated muscle spindle and Golgi tendon afferents responsible for signaling limb position and maintaining proprioceptive sense.

  2. Large-caliber, heavily myelinated Aβ axons responsible for signaling highly discriminative mechanical stimuli transduction.

  3. Small-caliber, lightly myelinated Aδ axons.

  4. Smaller, unmyelinated C axons that form morphologically simple free nerve endings (FNE) that terminate within the epidermis, around the vasculature, and all other skin appendages, as well as muscle, all visceral organs, and bone [5,8–12].

The Aδ axons and C neurons are the most abundant types of ending observed in peripheral tissue, a minority of which are characterized as responding to non-noxious thermal, chemical, and mechanical stimuli over a normal range of sensation. A majority of FNE are believed to respond specifically to various noxious stimuli and are collectively termed “nociceptors.”[13–16].

Segregation of Aβ axons from Aδ axons and C axons (together with coded sensory information) begins at the level of entry to the spinal cord or brain stem [6–8,13]. Here, Aβ axons establish different locations for synaptic termination from Aδ and C axons, and ultimately forge distinct connections and neural pathways in the thalamus as well as deep into the cerebral cortex. Sensory information from Aβ axons is thus transmitted along different anatomic routes from Aδ and C axons [6–8,13]. These anatomic pathway discriminators between large-caliber axon proprioception/fine-touch sensation and small-caliber axon thermal/nociception sensation have led to the identification of specific nociceptive pathways, where aberrations and interventions at any level can have clinical relevance.

The Aδ and C nociceptors are preferentially activated by various types of stimuli, perceived as being painful, and are categorized as being preferentially responsive to extreme forces (mechanical nociceptors), extreme temperatures (thermal nociceptors), noxious chemicals (chemo-nociceptors), or combinations of these stimuli (polymodal nociceptors) [7,13,15,16]. Characterized as thin-caliber fibers with different conduction velocities, nociceptors may be lightly myelinated Aδ axons (fast, sharp, stabbing pain) or unmyelinated C axons (slow, throbbing, burning pain) [6–8,13]. As many as half of the Aδ and C axons are normally unresponsive to test stimuli used to functionally characterize sensory innervation by electrophysiology. Many of these unresponsive axons have nociceptive characteristics and are believed to be driven only by severe tissue damage or prolonged experimentally induced pain, and are therefore referred to as “silent” nociceptors [9]. However, more recent studies have revealed that all nociceptors are molecularly complex and are all polymodal to some degree. Like other sensory systems where populations of similar, yet variably tuned, sensory receptors allow for graded responses and immense variation, nociceptor populations exhibit differences in expression of biochemical pathways that are required to transduce and transmit peripheral noxious sensations to the dorsal horn of the spinal cord [15,17–23].

Most cutaneous small-caliber axons terminate in the epidermis among the keratinocytes or within the wall of blood vessels as relatively simple morphologic structures, often straight and unbranched (epidermal), or within an elaborate mesh-like network (vessels) [14,24]. Differential responsiveness and function of primary sensory neuron FNE are partially explained by their unique locations. Thus, even FNE with comparable molecular expressions that terminate in different locations will be subject to different combinations of stimuli. The functional diversity of FNE is also explained by an equally diverse morphology and biochemistry. Structure dictates function, and sensory axon endings are, somewhat independently of size, studded with a unique repertoire of molecular receptors and ion channels that differentially respond to mechanical, chemical, and thermal stimuli [13,16,24]. Nociceptor FNE express a wide variety of receptors and ligands, with each ending likely having a different pattern of expression and functional activation, generated through an integration and summation of sensory stimuli [25]. Among all primary sensory neurons, temporal and spatial summation of stimuli with varying intensity and duration elicits graded (i.e., measured) responses in the form of action potential waves across the sheet of axon endings [5,6,8]. Depending on the type, intensity, and duration of the stimulus, different varieties of sensory receptor endings will be activated in different proportions and combinations, including non-nociceptive small- and large-caliber axons. The resultant pattern of activity among the entire population of peripheral sensory endings is coordinated by central mechanisms that govern sensory perception. Interestingly, recent studies have demonstrated that neurons, keratinocytes, and vascular endothelial cells express common signaling pathways, suggesting a bidirectional communication between the nervous system and cells of the epidermis and vasculature, the clinical relevance of which is under investigation [24,26,27].

Until only very recently, the epidermis and keratinocytes, in particular, were thought to have circumscribed roles in water barrier homeostasis, immune surveillance, and ultraviolet-stimulated endocrine activity. Results from clinical and animal studies now suggest that the epidermis transduces and integrates sensory stimuli as well [24,26–29]. Depending on the type, intensity, and duration of a stimulus, epidermal keratinocytes release different combinations and proportions of excitatory and/or inhibitory ligands, a differential release of signaling molecules that likely influences neuronal activity [24,30,31]. Interestingly, there is also considerable evidence suggesting neural interactions with melanocytes, Langerhans cells, and other epidermal cells, although their roles in sensory perception are unknown [16].

Sensory endings employ complex mechanisms that balance excitatory and inhibitory signals. Select primary sensory neurons, for example, may co-express calcitonin gene-related peptide, substance P, and opiate receptors (OR), the latter of which has three variants (δ, κ, µ), shown to differentially bind various opioid ligands. Although subsets of small-caliber axons express µOR, peripheral opiates exert little analgesic effect under nonstressed tissue states. Upon physical injury or disease, however, peripherally acting endogenous opiates demonstrate powerful analgesic effects [24,32–34]. Conceivably, this inhibitory (and ultimately analgesic) function of opiates on sensory axon activity is “unmasked” in stressed tissue. Although the biochemical basis of this unmasking is not well understood, studies have demonstrated a role for β-endorphin, an endogenous opiate peptide produced by superficial keratinocytes, which may represent a tonic inhibitory tone on µOR-expressing epidermal FNE [24,33]. This opiate-mediated tonic inhibition of epidermal FNE establishes the threshold for action potential generation and is therefore a critical determinant of neuronal excitability. Loss of this tonic inhibitory mechanism within the keratinocytes would likely lead to hyperexcitability of the µOR-expressing fibers, heightened nociceptive-like responses to varied stimuli, and potentially to chronic pain [24].

In addition to β-endorphin, keratinocytes release the purine nucleotide adenosine triphosphate (ATP) in response to air contact and various mechanical stimuli [27,35,36]. ATP likely contributes to the excitation of FNE axons through metabotropic G protein-coupled mechanisms (P2Y receptors) and through ionotropic direct ion-gated channels (P2X) that are present on small-caliber axons [24,37–39].

Hypothesized Peripheral Mechanisms of Pain

The term nociception is used to describe the sensory receptiveness of noxious or threatening sensory stimuli. All animals experience nociception, and nociception is essential to species viability. Nociceptive stimuli responses—so-called nocifensive behaviors—are immediate, without conscious thought, and directed at tissue preservation (i.e., they are adaptive).

Normal sensory function of primary epidermal or visceral FNE requires an exquisite balance between inhibitory and excitatory mechanisms. Alterations in the signaling pathways governing this balance influence the properties of neuronal firing and contribute to acute and chronic pain conditions. Classic definitions of pain perception center upon nociceptors, the specific class of primary sensory neurons dedicated to firing action potentials only after a noxious stimulus [7]. This dedicated nociceptor hypothesis proposes that a subset of primary sensory nerve endings (nociceptors) are anatomically positioned and biochemically primed to respond only to excessive types of stimuli that are perceived as painful in humans [7]. Within the framework of normal tactile perceptions, acute pain is a relatively rare event, and thermo-, mechano-, and chemosensitive small-caliber neurons are presumably active, and tactile perceptions are true (although under electrophysiologic recording, these axons often fail to exhibit action potentials without stimulation of their cutaneous receptive field). However, under acute tissue injury, quiescent nociceptive neurons may beactivated by direct stimuli—mechanical, thermal, or chemical, including such inflammatory cytokines as interleukin (IL)-1β and IL-6 released both immediately following injury and subsequently during tissue repair processes [40–42]. Inflammatory cytokines thus activate nociceptive neurons, leading to the generation of pain and nocifensive behaviors that accompany tissue injury and repair [41,42].

Peripheral Pathologies Likely Contribute to Chronic Pain Mechanisms

Preferential loss of epidermal endings is a common feature of chronic pain patients, and epidermal FNE quantification has become a standard method for aiding in peripheral neuropathy diagnosis [24,43,44]. A loss of FNE may result in the following:

  1. Disproportionate activity among remaining endings, leading to an abnormal reorganization of central connectivity or loss of centrally mediated inhibition.

  2. Increased spontaneous activity in select neurons in the dorsal horn of the spinal cord.

  3. Dysregulation of epidermal chemistry and inappropriate release of various ligands from keratinocytes [43,44].

Pathologic loss of FNE is often accompanied by morphologic alterations in the remaining FNE, including excess branching, which would likely increase the depolarizable (excitable) surface area, and possibly contribute to increased spontaneous activity (i.e., spontaneous pain), allodynia, or hyperalgesia [43,44]. Loss of FNE may also disturb signaling pathways within sensory endings, promoting sensory hyperactivity and peripheral sensitization, and further compound spontaneous pain, allodynia, or hyperalgesia [43,44]. Moreover, distribution of sensory endings to inappropriate targets (e.g., vascular sensory endings wrongly distributed to the epidermis) may cause axon populations to become hypersensitive to otherwise non-noxious cutaneous stimuli [43,44]. Deviant patterns of innervation and loss of proper tissue sensory feedback may relate to the etiology, progression, and/or maintenance of chronic pain and other chronic diseases, and to post-trauma sequelae, where the loss of tissue regulation and general homeostasis are observed.

Importantly, dysregulation of excitatory mechanisms and endogenous opiate inhibitory mechanisms in keratinocytes likely contributes to increased excitability of FNE [3]. Indeed, under conditions where any of these alterations have occurred—including, for example, peripheral tissue injury and subsequent upregulation of pro-inflammatory cytokines—the normal range of sensations and adaptive responses to sensory stimuli become skewed, leading to overactivation of sensory endings. With an acute nociceptive response from tissue injury or inflammation, removal of the initiating stimuli returns sensation to normal. However, under chronic pain conditions, neuronal and non-neuronal peripheral cell (keratinocytes, vessel endothelium, fibroblasts, etc.) alterations likely remain [24].

Rationale for Combination Therapies Aimed at Peripheral Mechanisms

Currently, only a few epidermal (in particular, keratinocyte-mediated) or vascular mechanisms governing sensory functions have been identified, but many more are likely present [27,30]. For example, keratinocytes express numerous “neural-related” signaling molecules, although most have been described as functioning in keratinization and/or wound healing, and recent vessel adventitia experiments demonstrate that the surrounding tissue of vessels, as well as the endothelial cells of the vessel wall, functions in vascular tone [24,26,45]. The recent recognition that keratinocyte expression of these mediators modulates sensory stimuli transduction and integration provides for an even greater dynamic range of response of cutaneous FNE, as well as other neural innervated skin appendages [24]. The clinical relevance of these findings is just now beginning to emerge, with important implications for therapeutic intervention. For example, a recent investigation of these sensory mechanisms in patients with complex regional pain syndrome and postherpetic neuralgia has revealed an increased expression of voltage-gated sodium channels (NaV) in keratinocytes [27]. Keratinocytes release the purine nucleotide ATP upon depolarization and other stimulations, and nociceptive sensory neurons sense ATP, which likely activates FNE-expressing purinergic receptors [24,27,31,35,46].

Certainly, the precise delineation of peripheral sensory mechanisms has implications for the discovery and development of novel analgesics, and for the development of new combination treatment strategies for chronic pain. Peripheral mechanisms are amenable to therapeutic intervention from systemically delivered compounds, and cutaneous mechanisms can be directly targeted by topical application of common pharmacologic agents (such as lidocaine), physical stimulation, nonphysical stimuli such as ultraviolet light, or genetic manipulation [47–49].

Of note is a study demonstrating upregulation of α2δ receptors, the cognate receptor for the gabapentinoids, in an animal model of neuropathic pain [50]. Increased expression occurred in response to peripheral nerve injury only; central nerve lesions had no effect, underscoring the importance of peripheral signaling in neuropathic pain. Whether the expression of other nociceptive-relevant receptors is altered in response to nerve lesions is an area of active research.

These and related molecular insights provide a biochemical basis for novel multidrug analgesic regimens. For example, combining peripherally acting opioid agonists with purinergic blockers, which independently suppress activity of primary sensory neurons, may have potentially additive or supra-additive analgesic effects [24,27]. One multidrug approach designed to quell primary axon activity would include, for example, a topical NaV blocker to limit keratinocyte-derived ATP release, in combination with a purinergic blocker to block ATP receptors on the FNE, additional dermal nerves, keratinocytes, and other purinergic-expressing cells [24,27].

Additional classes of compounds—vanilloids, serotonergic blockers, cannabinoids, adrenoceptor agonists, nonsteroidal anti-inflammatory drugs, corticosteroids, and blockers of tachykinins and neurotrophins—have been used with variable success to achieve analgesia via peripheral mechanisms [51]. These compounds elicit numerous and varied effects on peripheral mechanisms, allowing potential combinations to achieve additive and/or synergistic efficacy.

Central Mechanisms of Chronic Pain

Mechanisms uncovered thus far that have an essential, supportive, or modulatory function in central mechanisms of both acute and chronic nociceptive and pain signaling are manifold, and include spinal, supraspinal, brain stem, thalamic, basal ganglia, and cortical-level neural processing [3,52–55]. A complex neural loop processes nociceptive sensory information, leading to nociceptive responses (nocifensive behaviors) and the negative affect associated with the pain experience. Peripheral nociceptive sensory input is initially processed within the spinal dorsal horn and involves the complex interaction of classic neurotransmitters, neuropeptides, lipid-derived signaling intermediates, amino acid transmitters, steroids, hormones, and extracellular matrix ligands [56–61]. At each subsequent level of the central neuraxis, the afferent nociceptive signaling is processed as it progresses to cortical regions and conscious attention. Importantly, autonomic- and motor-efferent responses are generated to nociceptive inputs, likely at each level of processing, including a modulatory brain stem descending output activity with projections to the spinal dorsal horn that can inhibit or facilitate neurotransmission events at the terminals of initiating primary sensory afferents [56,59,62,63]. Thus, spinal dorsal horn (and likely cranial sensory nuclei) processing allows for the immediate and directed control of all nociceptive peripheral inputs by the CNS via synaptic interactions among the primary sensory neurons, spinal interneurons, spinal relay neurons, and glia.

Nociceptive primary afferent fibers use glutamate and neuropeptide transmitters to convey the status of the peripheral receptive field. Modulation of glutamate and its cognate receptor family are critical to nociceptive processing in the dorsal horn, and likely have implications in neuropathic and chronic pain conditions [64,65]. Studies have partially elucidated biochemical cascades that orchestrate pain processing within the dorsal horn and that modulate glutamatergic synaptic plasticity, strength, and neuronal excitability [64,65]. Additionally, primary peripheral nociceptive input is under tonic control by descending brain stem-level regulatory systems, using serotonin, norepinephrine, and opioids as transmitters/modulators, which can inhibit or facilitate nociceptive transmission in the spinal dorsal horn [53,56,62]. Many drugs—including, for example, opiates, antidepressants, and anticonvulsants, among others—can influence the signaling of these descending systems and thereby achieve effective analgesia for patients with chronic pain. Clinical studies evaluating combinations of these agents have had encouraging results (see below). Additional alterations in signaling pathways involved in spinal-level nociceptive processing are not only limited to neural descending presynaptic control or to postsynaptic N-methyl-D-aspartic acid (NMDA) receptor modulation, but also include glia. Astrocytes can modulate synaptic function and release cytokines that sensitize the system, and have been shown to discriminatively respond to active neurons, implying a bidirectional communication not previously appreciated [42,57,66–70].

Central sensitization is a well-defined phenomenon characterized by facilitation of pain transmission after peripheral injury and can occur at any level of the CNS. In spinal-level sensitization, for example, persistent small-caliber fiber stimulation generates long-term potentiation—a memory of the injury—in select neurons of the spinal dorsal horn. Spinal dorsal horn sensitization and “wind-up” have been extensively studied, and numerous effectors of sensitization have been identified, including postsynaptic NMDA glutamate receptors, glial-derived cytokines, and receptor tyrosine kinases of the Eph family, among others [61,64,68,71–74]. These dorsal horn signaling pathways represent potential targets for analgesic intervention. For example, an intrathecal block of dorsal horn EphB tyrosine kinase receptor inhibits the induction and maintenance of nerve injury-induced thermal hyperalgesia and mechanical allodynia.

When functioning properly, the integration of the nociceptive afferent-receptive and efferent-responsive systems allows for the detection of harm and potential harm before overt tissue damage occurs. However, interruption and/or dysregulation at any point within the afferent–efferent loop can result in immediate or delayed sensations of pain and discomfort, which can be acute or intermittent, of any description (i.e., sharp, dull, stabbing, burning, etc.), and which may act to skew any other point along the loop. Importantly, an injury to the periphery can irreversibly alter central signaling mechanisms, and central perturbations of these mechanisms can likewise alter peripheral signaling mechanisms [24]. Reciprocal interactions within this interdependent loop may have important implications for the onset and progression of chronic pain.

Emerging but Limited Evidence Supporting Multidrug Therapy

Complex diseases are often characterized by a collapse of homeostatic mechanisms involving multiple regulatory loci, each of which constitutes a target for therapeutic intervention. Elaborate biochemical pathways thus provide a rationale for combination therapies that modulate multiple disease-relevant targets. In hypertensive patients in whom mechanisms that otherwise tightly control blood pressure unravel, multidrug therapies (MDTs) employing diuretics, β-blockers, angiotensin-converting enzyme inhibitors, and other agents have proven effective. MDT is widely accepted and consistently utilized for many other complex chronic medical disorders, including diabetes, HIV, and cancer; in fact, in many instances, it is the accepted standard of care [75,76]. Similarly, insights into the neurobiologic basis of chronic pain are providing an increasingly cogent rationale for MDT.

Further highlighting the complexity of chronic pain, patient care and outcomes have been shown to significantly improve when treated with spinal cord stimulation, selective nerve blocks, epidural steroid injections, and a wide array of cognitive–behavioral therapies. Clinical evaluation of rational combinations of therapies is in its early stages and may over time demonstrate potentially additive or even supra-additive (i.e., synergistic) benefits of therapies comprising pharmacologic, interventional, and psychosocial therapies.

Two seminal studies on MDT strategies for neuropathic pain have been published. In one study, patients with either diabetic neuropathy or postherpetic neuralgia were evaluated in a randomized double-blind, active placebo-controlled, four-period crossover trial that assessed the analgesic effect of morphine, gabapentin, or their combination. The results demonstrated that, when combined, gabapentin and morphine achieved better analgesia, at lower doses, than either medication alone. Constipation, sedation, and dry mouth were the most frequent adverse effects of this combination [77]. Additionally, Hanna and coworkers recently reported results from a 12-week study of 338 patients with moderate to severe painful diabetic neuropathy, all of whom were already taking a maximally tolerated dose of gabapentin; patients were randomized to receive gabapentin with extended-release (ER) oxycodone or placebo. Doses of ER oxycodone were allowed to be titrated during the entire study. Compared with placebo–gabapentin treatment, the combination of oxycodone and gabapentin resulted in statistically greater pain relief and fewer discontinuations of therapy due to suboptimal response [78]. In these MDT regimens, two medications with distinct and complementary mechanisms of action proved to have additive effects for patients with chronic pain. Proposed principles for MDT have been previously published and are summarized in Table 1[79].

View this table:
Table 1

Proposed principles for multidrug therapy

1. Perform an initial assessment and establish, if possible, an appropriate diagnosis, addressing all medical and psychological comorbidities, inferred underlying pathophysiologic mechanisms, laboratory findings (e.g., imaging studies, electrodiagnostic studies), and level of risk for adverse events and nonmedical use of medications.
2. Review past drug treatments: their doses, effectiveness, and adverse effects. Frequently, past experience with medications helps guide future choices. Over-the-counter preparations or complementary therapies should be included in this review. In particular, drug combinations should be identified.
3. Consider first a drug associated with a rigorous evidence-based and defined mechanism of action. Other considerations include:
  • Ease of use.
  • Minimal adverse effects and minimal or no end-organ toxicity.
  • Minimal drug–drug interactions.
  • Cost and reimbursement.
4. Initiate therapy with a dose at the low end of the recommended range and titrate to therapeutic effect slowly, particularly in elderly patients.
5. Assess continually for analgesic and functional improvement—only therapy that provides clinically meaningful relief (e.g., more then 30% improvement, or 2 out of 10 or more on the 0–10 pain rating scale) with documented improvement in function is to be continued. Discontinuation of agents that fail to demonstrate this functional improvement should be performed with a transparent, mutually respectful dialogue structured on patient goals and expectations. Alternate therapies should be considered.
6. Monitor for adverse effects. Slow titration allows for tolerance to adverse effects to develop, but if patient is still having significant side effects with minimal efficacy, drug should be discontinued.
7. Consider changing the medication if significant side effects overshadow efficacy; preferably select another medication within the same drug class and with a similar mechanism of action (sometimes referred to as “rotation”). If a similar drug is not available, then consider adding a medication to control side effects; these additional steps require experience with all of the drugs alone and in combination. If pain relief remains inadequate despite rotation to drugs within the same class, then consider addition of drug from another class with a different mechanism.
8. Select and use one drug at a time, in general, and adjust the dose as follows: start low, go slow, and monitor effects and adverse effects until maximum benefit has occurred or usual therapeutic dose level has been reached. Patients may not demonstrate dose proportionality: efficacy may be seen at lower doses and additional benefit may not be seen with higher doses. In this case, the lower dose should be continued.
9. Combine medications with differing modes of action, based on patient response and functional goals, pain syndrome, and clinical experience. There also may be instances in which it is appropriate to use two medications from the same class, such as two opioids, e.g., one with a long duration of action and one with a short duration of action, or two anticonvulsants with different mechanisms of action, e.g., pregabalin and carbamazepine.
10. When adding one more drug, weigh these and related issues; in particular, consider potential adverse interactions with drugs the patient is already taking.

Integrating Biopsychosocial Insights into Multimodal, Multidisciplinary Care

Pain is a subjective, multifaceted experience, shaped not only by biological mechanisms, but also by psychologic and cognitive variables, and by the social milieu in which the patient lives. Sensory input transduced and transmitted by neural pathways from peripheral FNE to the spinal cord and modulated by complex facilitatory and inhibitory neurotransmitter systems is perceived, interpreted, and ascribed meaning by discrete limbic and cortical regions of the brain [80]. Brain imaging studies have demonstrated extensive involvement of cortical areas in the perception of the pain experience [81]. Importantly, higher level and cortical reorganizations of nociceptive pathways likely play critical roles in the persistence and maintenance of chronic pain conditions. Emotional responses associated with chronic pain are governed by midbrain activity and shaped by cortical activity and cognition that may modulate and potentially amplify the emotions [1]. This emotive–cortical loop may lead to a vicious cycle, further increasing the distress and impairments associated with chronic pain. Patient-specific responses to aberrant nociceptive signaling are explained, in part, by cognitive and affective dimensions comprising fears, hopes, expectations, and memories, all of which are integrated into the pain experience through poorly understood mechanisms.

Poor adjustment to chronic pain is manifested by learned helplessness and hopelessness about pain, depression, stress, maladaptive beliefs and pain catastrophizing, and coping skill deficits [1]. All of these factors critically influence the chronicity and severity of pain and its impact on function. In one cross-sectional study of 211 patients with a variety of chronic pain syndromes, for example, catastrophizing was found to be significantly associated with pain severity and pain-related disability and distress [82].

Pathophysiologic mechanisms underlying the relationship between chronic pain and affective disorders are not well understood. Certainly, chronic pain imposes a stressful load on the patient, akin to the stress of extreme hunger, thirst, and other primal needs. The stress response generated and maintained by chronic pain is mediated through the autonomic nervous system and hypothalamic–pituitary–adrenocortical axis. Stress and pain thus participate in a vicious cycle, each mutually reinforcing the other, imposing anallostatic load thought to impair immune activity, as well as cognitive, behavioral, and physical function [1,83]. Anxiety, emotional distress, and negative emotions increase autonomic and CNS activity as well.

Social factors, too, are important determinants of pain symptomatology and related disability. Variables likely contributing to the multidimensional experience of pain, and its response to multimodal treatments, include poor diet, smoking, social support, job security, low socioeconomic status and limited access to preventative care, past history of sexual or physical abuse, cultural background and related beliefs, external locus of control, bereavement over lost family members, and marital discord.

Neuropsychologic and social dimensions of the chronic pain experience are thus increasingly regarded as an organizing principle for the individualization of patient care. Physician understanding of these dimensions informs psychosocial interventions, many of which are described in Table 2. Careful assessment of pain-associated sensory abnormalities, and cognitive, behavioral, and affective distress may help guide selection of these and related strategies. Of particular importance is addressing the emotional state of the patient, and not limiting care to objective tissue pathology, laboratory findings, and even the inferred pathophysiologic basis of the pain syndrome [1]. Assessment hinges on a number of elements; especially determinant are the patient's understanding and attitudes toward chronic pain and the chronic adjustments that must be made; coping style, strategies, and substance; and overall quality of life [84]. Personal and family history of psychopathology and/or substance use is often suggestive of an increased risk for inappropriate use of certain medications, particularly opioids [85]. A comprehensive survey of these and other approaches is beyond the scope of this review. In short, psychosocial modalities are designed to reduce the suffering of the pain experience; to educate patients about their condition; to reduce anxiety, fear, depression, anger, and stress; and to teach coping skills and improve self-efficacy [1].

View this table:
Table 2

Behavioral interventions for chronic pain

Assertiveness skills trainingDistressful emotions are reduced when communication skills are enhanced, reducing physiologic arousal [91].
BiofeedbackPatients are trained to influence physiologic processes—such as blood pressure, skin temperature—with the aid of visual or auditory devices that amplify these processes. Examples include electromyographic biofeedback, which measures tension in the frontalis muscle [92,93], and EEG biofeedback, which has been found effective in treating some chronic pain [94].
CBTCognitive strategies and skills are taught so that maladaptive processes and irrational thinking, which directly affect perceptions and experiences, can be overcome [95,96].
Habit reversalSkills are taught to identify and then overcome poor functional habits or distressing, recurring thoughts that precipitate and perpetuate painful sensations [97].
Hypnosis and guided imageryWith visualization and one's imagination, patients can obtain a hypnotic state that is essentially aroused yet has little or no peripheral awareness [98]. In guided imagery, patients focus on something (e.g., their chronic pain) that they would like to alter or eliminate [99]. Suggestibility is an important element of both hypnosis and guided imagery.
MeditationIn the health care setting, the forms of meditation that have been best researched include transcendental meditation [100,101] and mindfulness meditation [102]. In the former, the patient repeats a silent word or mantra to reduce and eventually transcend one's internal dialogue. In the latter, the patient maintains a nonjudgmental state of awareness in which emotions, judgments, beliefs, and so on are addressed.
Patient educationTeaching patients about common symptoms, possible adverse effects, appropriate treatments, self-care strategies, and the likely course of the discomfort has been found to reduce the anxiety that may be heightened in the uninformed, which in turn may prolong symptoms [103]. Psychoeducational approaches broaden the scope of the training to include adaptive psychological strategies such as CBT [104].
Relaxation techniquesHypoarousal may be obtained from several relaxation procedures, the most studied of which is progressive muscle relaxation, which aims to reduce muscular tension by alternately tensing and relaxing muscles [105]. Hypometabolic states in which sympathetic arousal is reduced—as in Benson's relaxation response—also may be achieved [106]. These techniques can ameliorate symptoms that are associated with chronic pain, such as anxiety, fatigue, and sleep disturbance.
  • CBT = cognitive–behavioral therapy; EEG = electroencephalogram.

Significant analgesia constitutes an untestable and at times elusive treatment goal for patients with chronic pain. Increasingly important is an emphasis on measurable patient outcomes, in particular, improvements in cognitive and physical function. Increasingly clear scientific support for multimodal therapy for chronic pain raises the expectation of evidence-based clinical guidance [86]. Studies have demonstrated that multidisciplinary, multimodal, and biopsychosocial rehabilitation can reduce pain, improve function, and build self-efficacy [87]. In his meta-analysis of 25 psychosocial modalities, Devine demonstrated the utility of psychoeducational interventions—in particular, cognitive–behavioral therapies emphasizing relaxation, support groups, and education—for patients with cancer-related pain [88]. In another study, patients with chronic low-back pain received either “usual” care or a more defined treatment that included working with a psychologist, physical therapist, and rehabilitation specialist. Other studies have focused on the use of multidisciplinary care to help reduce or stabilize medication use [89].

Although pain practitioners and researchers have emphasized the importance of integrating combination strategies, implementation may be difficult in clinical practice [90]. Methodologically sound studies that evaluate the effects of combining behavioral and/or interventional therapy with MDTs are needed. Until data from such clinical studies are available, clinical judgment and experience will continue to drive the individualization of care for patients with chronic pain.


Dr. Argoff has no disclosures to report.

Dr. Albrecht has no disclosures to report.

Dr. Rice has no disclosures to report.

Dr. Irving serves as a consultant/speaker for Eli Lilly.

This supplement has been sponsored by an unrestricted grant from King Pharmaceuticals®, Inc. Editorial support was provided by Megan Fink, Ariel Buda-Levin MS, John Lapolla MS, Maggie Van Doren PhD, Jim Kappler PhD, as well as Innovex Medical Communications.


View Abstract