FormalPara Key Points

Opioid analgesic medications induce a paradoxical state of increased pain sensitivity, known as opioid-induced hyperalgesia, which likely interferes with patients’ overall treatment outcomes.

A unanimous method of diagnosing and treating opioid-induced hyperalgesia in clinical practice remains indeterminate.

There are several promising avenues of research that may lead to an effective pharmacological treatment paradigm for opioid-induced hyperalgesia.

1 Introduction

In 1870, English physician T. Clifford Albutt chronicled the first documented observation of increased pain sensitivity, or hyperalgesia, resulting from opioid exposure. He stated “At such time, I have certainly felt it a great responsibility to say that pain, which I know is an evil, is less injurious than morphia, which may be an evil. I have much reason to suspect that a reliance upon hypodermic morphia only ended in a curious state of perpetuated pain”. In his discourse, he questioned whether morphine “tends to encourage the very pain it pretends to relieve” [1]. Over 140 years later, we are only beginning to unravel the enigmatic mechanisms underlying this paradoxical phenomenon and are ultimately coming to the realization that Albutt’s observations were likely accurate.

Predicated from a sample of non-institutionalized adults in the US, an estimated 100 million people suffer from some form of chronic pain [2], with an approximate 25 million experiencing moderate to severe, debilitating pain [3]. Encompassing a range of chronic pain states, estimates of cost range from US$560 to US$635 billion each year [2]. It is notable that the estimated 33 % prevalence rate for chronic pain in the US appears to be higher when compared with other countries; for instance, sample estimates for India were approximately 13 % [4], while another study reported chronic pain in 22 % of women versus 17 % of men in a Danish sample [5]. Further, an Australian study reported chronic pain in 19 % of their primary care sample [6], while a Canadian study reported similar estimates in 2011 [7].

The assertion by the US Joint Commission on Accreditation of Health Care Organizations that pain be considered “the fifth vital sign” has resulted in an increased focus on providing treatment for chronic non-cancer pain. This assertion has been criticized by some, as pain is more accurately perceived as a subjective, multidimensional symptom rather than an objective sign [8]. Nonetheless, limited availability of alternative therapies often leaves opioid pharmacotherapy as the only viable option, the result of which has been a dramatic increase in prescription of high-dose opioids for chronic pain.

Opioids have become the most frequently prescribed drug class, surpassing 259 million prescriptions in 2012 in the US [9]; this increase in number of prescriptions has given rise to a drastic increase in the rate of opioid addiction, along with a striking increase in the number of deaths due to prescription opioids. Of particular concern is that the majority of fatal opioid overdoses (50–80 %) have occurred in patients with a history of chronic pain [10] on prescribed opioid regimens [11]. Particularly disconcerting is that opioid prescription is at its highest in spite of reports questioning the efficacy of long-term opioids for chronic pain, and supporting a dose-dependent risk in misuse, addiction, overdose, and death [3, 12]. In fact, one observational study reported that 90 % of patients presenting to a pain management center were already on an opioid regimen [13], suggestive of either inadequate pain relief, addiction, or both. A solid estimate of patients presenting with chronic pain and a comorbid opioid use disorder (OUD) is not available, is likely dependent on duration and dose of opioid prescription [14], and reportedly ranges between 3–40 % [15]. Furthermore, OUD has an estimated economic cost of US$72.5 billion in healthcare services in the US [16], and those who abuse this drug class generate an average annual healthcare cost that is 8.7 times higher than non-abusers [17].

Opioids are associated with multiple adverse effects that include respiratory depression, constipation, cardiovascular issues such as hypotension and bradycardia, and a high abuse liability [18]. Furthermore, opioids elicit a paradoxical state of increased pain sensitivity, known as opioid-induced hyperalgesia (OIH). OIH manifests as an increased pain sensitivity that leads to higher opioid consumption, and in turn putatively diminishes overall pain relief [19, 20]. Further, OIH may interfere with the patient’s overall chronic pain treatment outcome and has the potential to propagate addiction. Thus, it has been suggested that OIH be considered a treatment target in addition to analgesia to ultimately improve management of patients treated with opioids [2123].

The purpose of this review is to evaluate what is currently known about the neurobiological substrates of OIH, to encapsulate the human studies on pharmacological treatments for OIH, and to recommend future considerations for addressing OIH in clinical settings. To provide background for our review, we first provide definitions for OIH and similar phenomena such as tolerance and sensitization. We then summarize the recent advances on the neurobiological mechanisms of OIH, and address issues related to assessment of OIH in preclinical and clinical studies. Several relevant systematic reviews of OIH have been published recently [2022, 2426]; our review extends the previous work by providing an updated evaluation of the field with a focus on viable clinical treatment targets.

2 Methods

Between August 2014 and April 2015, a search of PubMed, EMBASE, and Google Scholar was conducted to retrieve articles related to OIH. Specifically, both original reports and past review articles pertaining to OIH’s definition, prevalence, measurement, relationship to other opioid-induced phenomena, underlying neurological substrates, and treatment paradigms were sought. Search terms included the following: opioid-induced hyperalgesia, pain, opioid use disorder, NMDA receptor, GABA receptor, glia, gabapentinoids, neurosteroids, and cannabinoids. All terms were filtered by either animal or human studies when relevant.

3 Definition and Theories of Opioid-Induced Hyperalgesia (OIH)

Following opioid exposure, OIH contrasts analgesia (relief from pain) and antinociception (decreased pain sensitivity), and manifests as subjectively perceived increased pain sensitivity to an already painful stimulus. OIH frequently includes allodynia, which is pain evoked by a stimulus that is not under normal circumstances considered painful [22, 27]. Perhaps most notable is that this increased pain sensitivity is diffuse and often manifests beyond the area of original pain, is independent of the condition that the opioids were initially prescribed for, and persists or even worsens after dose escalation of opioids [2830].

There are several prominent theories underlying OIH. Perhaps the most intensely debated theory, though recently largely unsupported, is that OIH is a causative factor in analgesic tolerance (see Sect. 3.2). Specifically, the increased pain sensitivity that one experiences after opioid treatment causes the need for increasing doses of opioids, thus reflective of tolerance [3133]. A second theory has postulated a role for pronociceptive opioid metabolites, specifically resulting from morphine biotransformation. However, recent studies have suggested otherwise (see Sect. 4.4). A third prevalent theory is that OIH is an adaptive response that serves as a systems-level opponent process in response to opioid analgesia after prolonged opioid treatment. That is, as the body works to retain homeostasis, one’s adaptive response to exogenously elicited analgesia is to create an opposing state of hyperalgesia in an effort to regain physiological equilibrium [19, 3436]. Finally, a fourth theory that is gaining popularity, and is related to the aforementioned opponent process, is that OIH is the result of a glial-regulated immune response (see Sect. 4.2) [37, 38].

3.1 Prevalence of OIH in Humans

OIH is observed in patients who utilize opioids with both acute and chronic administration, though its overall prevalence remains indeterminate. The closest estimates indicate that a disconcerting proportion of patients present with what may be OIH; for instance, Mercadante et al. [39] reported nearly 14 % of 81 patients with cancer experienced diffuse pain not otherwise explained by patient history or disease progression, which was unresponsive to morphine titration and alleviated only by rotation to methadone. Ackerman [40] described a similar outcome in 28 % of a sample of 197 patients with chronic pain receiving opioids. While some studies suggest that OIH develops after chronic opioid use [4143], both rodent [44, 45] and human studies [4649] have documented OIH within hours of acute opioid administration. Studies using shorter-acting opioids such as remifentanil, sufentanil, fentanyl, and morphine provide the most suspicion of OIH [4956]; however, there is documentation of this phenomenon in patients receiving long-acting opioids (e.g., methadone or buprenorphine) as well [23, 43, 5760].

Though ample preclinical studies spanning over three decades provide evidence for OIH, clinical studies have several notable limitations which challenge accurate assessment and thus prevalence estimates of OIH. Of primary concern is the fact that studies investigating OIH often utilize different measures of pain (e.g., thermal or cold stimulation) that may not be related to clinical pain, so interpretation and comparison between studies becomes difficult. These limitations led some authors to question the validity of OIH in humans [20, 25, 61]. Secondly, the numerous populations under study likely influence the evaluation of OIH. Specifically, a majority of existing clinical studies of OIH focus on several cohorts: healthy volunteers receiving acute opioids while undergoing human experimental pain testing [6265], those with acute pain following various surgical procedures [5153, 66, 67], patients suffering from chronic pain [41, 6870] or chronic illnesses such as cancer [50, 7174], and populations of individuals with OUD [23, 43, 5760]. Third, OIH does not manifest in all patients receiving opioids; individual differences, including psychological factors, likely play a role in whether one develops OIH following opioid exposure, though research on this topic is needed. Finally, distinguishing OIH from tolerance and opioid withdrawal remains a challenge in clinical settings, as these phenomena share similar features, but are indeed distinct (Sects. 3.2, 3.3; Table 1). For instance, many studies assessing OIH, particularly under surgical contexts, are often criticized for not accurately measuring OIH but rather opioid tolerance and/or withdrawal. Specifically, a hallmark outcome of these studies is an increase in postoperative opioid consumption following abrupt termination of an acute, fast-acting opioid, indicative of tolerance and/or withdrawal rather than OIH [20, 25]. As outlined in Tables 1 and 2, OIH, opioid tolerance, opioid withdrawal, and OUD have distinct clinical features and thus can be separated.

Table 1 Distinguishing between opioid tolerance, withdrawal, opioid use disorder, and opioid-induced hyperalgesia: definitions, symptoms, and known solutions
Table 2 Criteria for diagnosing opioid-induced hyperalgesia in clinical populationsa

3.2 OIH versus Tolerance

OIH is often associated with development of analgesic tolerance, defined as decreased analgesic potency following prolonged opioid administration [18, 75]. While it has been commonly believed that analgesic tolerance and OIH are synonymous, both preclinical and clinical studies have indicated that OIH is unrelated to prior opioid analgesia or tolerance [42, 68, 7678]. In an attempt to elucidate the clinical ambiguity between tolerance and OIH, Chen et al. [30] distributed a survey to 1408 US practitioners consisting of targeted questions that assessed their clinical experiences with opioid treatment. The results identified a significant knowledge gap in the differentiation of tolerance versus OIH; the distinguishing factor was the finding that increasing opioid dose overcomes tolerance, but leaves OIH unresolved [30]. Most striking is the finding that a hyperalgesic patient subject to upward dose titration experiences a significant worsening of diffuse hyperalgesic symptoms as the opioid dose increases [18]. As such, effectively treating chronic pain while adjusting for tolerance and attempting to define and avoid hyperalgesia becomes a difficult clinical problem (see Table 2 for suggested criteria for diagnosing OIH).

3.3 OIH’s Relationship to Addiction

An additional complexity is OIH’s unexplored relationship with OUD. In the Diagnostic and Statistical Manual of Mental Disorders 4 (DSM-IV) [79], opioid use was diagnosed as either abuse or dependence, with the former being a less severe or early phase manifestation of the latter. However, the DSM-5 has combined abuse and dependence into a single diagnosis of OUD, measured on a continuum ranging from mild to severe. According to the DSM-5, OUD is defined as a pervasive pattern of opioid use despite significant consequences [80]. OUD includes criteria clusters for impaired control over drug use, social impairment, risky use, and pharmacological criteria that includes tolerance and withdrawal (Table 1) [8082]. It is important to note that for individuals who are taking opioids under proper medical supervision, tolerance and withdrawal are expected and thus should be weighted appropriately when considering a diagnosis of OUD.

It is possible that OIH may increase the likelihood of developing OUD. Specifically, OIH manifests as increased pain sensitivity, which could lead a patient to believe that they must consume more opioid than truly required to treat the original underlying pain condition. Furthermore, it is not clear if the hyperalgesia seen during opioid withdrawal is synonymous with OIH seen during continuous use, as preclinical studies have demonstrated a mechanistic difference [78, 8386]. In the case of opioid withdrawal syndrome, hyperalgesia appears after abrupt termination of opioid use. For instance, in preclinical paradigms using discontinuous morphine delivery (i.e., multiple injections), hyperalgesia may be intensified by ‘mini-withdrawal’ episodes when compared with chronic administration paradigms [35, 8789]. However, in the case of OIH, increased nociception persists or even worsens following continuous treatment with any one opioid [29, 42]. Further, case reports suggest that OIH typically dissipates after opioid cessation [50, 90]; however, only a handful of studies have investigated this directly, the results of which suggest that more than 1 month of abstinence is necessary for pain responses of former opioid users to approach that of healthy control participants. Specifically, Compton [91] reported the resolution of OIH in former opioid users abstinent for an average of 1 year, although Pud et al. [60] reported no change in reduced cold pressor pain tolerance after 1 month of opioid cessation in a similar population.

Opioid overuse pain syndrome (OOPS) has recently been proposed by Mehendale et al. [8] to classify a subset of opioid users with chronic non-cancer pain who do not adequately respond to high doses of opioids, show signs of inappropriate use/OUD, have decreasing function despite opioid use, favor pharmacological treatments, and demonstrate OIH. Of particular importance, OOPS takes into consideration the biopsychosocial aspects of pain, noting that opioids also alleviate depression and anxiety related to pain catastrophizing, while also propagating euphoria. As such, the authors postulate that opioids may actually be treating the psychological response to pain more so than the physical pain itself. Such a diagnosis may be warranted in patients with chronic non-cancer pain who demonstrate the need for multidimensional treatment beyond the sole use of opioids, the latter of which should only be managed by a highly trained pain management physician as a last resort. In fact, it is recommended here that patients are tapered off opioids completely, and referred to other pharmacological and psychological interventions [8].

One should also note the distinction between OIH and pseudoaddiction, the latter representing scenarios in which a patient presents with drug-seeking behaviors/reports of increased pain as a result of insufficient treatment. In such a case, these behaviors cease once an effective pain management regimen is established [92].

4 The Neurobiology of OIH

Potential mechanisms underlying OIH have emerged from the escalating amount of attention this phenomenon has received over the past several decades. Perhaps most intriguing is the finding that OIH occurs independently of opioid receptors [42, 77, 9396], and is not a consequence of either prior or concurrent opioid receptor activity [42]. As such, research spanning nearly two decades has focused on understanding how OIH manifests without direct involvement of the opioid receptor system, the main findings of which are summarized in the following sections.

4.1 Central Sensitization in OIH

Prolonged opioid exposure leads to positive system adaptation, or central sensitization, that includes allodynia, hyperalgesia, and an expansion of the receptive field, which results in secondary hyperalgesia. In nociceptive pathways, sensitization specifically references increased synaptic transmission/efficacy in somatosensory neurons of the spinal dorsal horn, and thus pain hypersensitivity, following peripheral noxious as well as innocuous stimuli [97, 98]. Importantly, as central sensitization results in reduced nociceptive thresholds and the consequential phenotypic expression of pain, it is likely an important mechanism for the development of OIH [98, 99]. Specifically, OIH may result from sensitization of primary afferent neurons as well as microglia [98], resulting in neuroplastic changes of receptor systems in the peripheral and central nervous system that lead to enhanced production, release, and spread of excitatory neurotransmission. This, in turn, results in the suppressed reuptake or increased release of neurotransmitters such as glutamate and substance P, subsequent sensitization of pain pathways, and ultimately manifests as amplified subjective nociceptive responses [32, 33, 89, 97, 100].

For instance, cells in the dorsal horn that express G-protein coupled neurokinin-1 receptors (NK-1) have been shown to project to supraspinal areas that facilitate pain processing, and an upregulation of this receptor has been implicated in OIH [101]. The endogenous ligand for NK-1, substance P is an excitatory neurotransmitter synthesized by primary afferent nociceptors [102], and is released in the dorsal horn after noxious stimulation [103]. Importantly, substance P has a documented role in central sensitization-induced hyperalgesia accompanying inflammatory pain [100]. For instance, in pathological pain states, there is increased substance P release in primary afferents following noxious stimulation, resulting in the internalization of NK-1 receptors in both superficial regions and deep layers of the dorsal horn. Although this is the theorized neuronal substrate of inflammatory pain, this mechanism has been hypothesized to play a role in the neuronal plasticity associated with OIH [104]. For example, preclinical studies have shown that substance P administration evokes chemical, thermal, and mechanical hyperalgesia following either intermittent or chronic exposure to morphine. Further, OIH evoked by both systemic and intrathecal morphine elicits substance P activity and increases spinal NK-1 receptor expression, suggesting a strong role for central sensitization in this paradoxical pain state [33, 89, 100].

In addition to the sensitization of afferent pathways, descending facilitation of nociception from the rostroventral medulla (RVM) has also been implicated in OIH. Specifically, this descending feedback has been found to increase spinal levels of the endogenous opioid peptide dynorphin [105107]. Though originally proposed to be antinociceptive, dynorphin has more recently been found to be a paradoxically pronociceptive kappa-opioid receptor agonist, modulating synaptic transmission via non-opioid receptor mechanisms [107]. That is, dynorphin potentiates pronociception that is naloxone-insensitive by increasing neuronal receptive field size and sensitizing N-methyl-D-aspartate (NMDA) receptors, resulting in the increased release and subsequent binding of excitatory neurotransmitters (i.e., glutamate) as well as cytokines [108], and consequent release of intracellular calcium [107, 109]. For instance, preclinical studies indicate that intrathecal administration of dynorphin produces mechanical, tactile, and cold allodynia, and mechanical hyperalgesia that is blocked by NMDA receptor antagonism (MK-801 or LY235959), but is unresponsive to opioid receptor blockade. Additionally, dynorphin antiserum obfuscates morphine-induced hyperalgesia [107, 110, 111]. The results of these studies suggest a role for central sensitization involving endogenous opioids in the descending, facilitative control of OIH.

4.2 Neuroimmune Mechanisms

Microglia are activated in response to drugs of abuse, including opioids, through activation of the toll-like receptor 4 (TLR4). As suggested by multiple studies, microglial activation results in reduction of the analgesic efficacy of opioids by development of tolerance and hyperalgesia. Specifically, microglia activation results in the release of pro-inflammatory cytokines [tumor necrosis factor (TNF)-α, interleukin (IL)-1β, and IL-6 and others], chemokines, lipid mediators of inflammation, matrix metalloproteases, excitatory amino acids, and nitric oxide [36, 38, 112118], all of which enhance neuronal activation. The release of pro-inflammatory cytokines results in increased number and conductance of alpha-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate (AMPA) and NMDA receptors, and down-regulation of gamma-aminobutyric acid (GABA) receptors [36, 38]. Thus, the consequence is neuronal hyperexcitability and pain enhancement [36, 116, 117, 119, 120]. This pronociceptive microglial activity suggests that OIH may be a neuroinflammatory response initiated in opposition to analgesia, resulting from prolonged exposure to exogenous opioids [34, 38, 117, 118]. Thus, it is plausible that these proinflammatory mediators (cytokines and chemokines) may be clinically assessed as a biomarker of microglial activation and thus potentially OIH [37, 112, 115, 121] (Fig. 1).

Fig. 1
figure 1

Neuroimmune mechanisms of opioid-induced hyperalgesia. Opioids activate microglia, which in turn causes an increase in excitatory transcription factors, excitatory amino acids, nitric oxide, cytokines, and chemokines. The net result of this neuroimmune activity is inflammation, an increase in pain sensitivity, and thus the putative modulation of OIH. Pharmacological agents that inhibit microglial activation may attenuate OIH. AMPA alpha-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate, GABA gamma-aminobutyric acid, OIH opioid-induced hyperalgesia, NMDA N-methyl-d-aspartate. Figure is created by C. A. Arout using ChemBioDraw Ultra 14.0 (Cambridge Software, USA)

Opioids have been shown to have direct actions on brain and spinal microglia [38, 115, 117, 119, 122125]. Specifically, microglia express opioid receptors, though the behavioral significance of this remains unclear [113]. Of particular interest to microglia’s role in neuroexcitatory opioid effects is TLR4, where morphine, buprenorphine, oxycodone, meperidine, methadone, fentanyl, and remifentanil are thought to exert significant agonist activity [36, 38, 114116, 119, 120, 122, 126]. TLRs are innate immune pattern recognition receptors found primarily on microglia, and following inflammation, this receptor’s activity is increased throughout the CNS. TLR4 stimulation is fundamental in the downstream activation of inflammatory transcription factors and release of proinflammatory factors [36, 117, 118, 127, 128] (Fig. 1).

TLR4 agonists reportedly enhance pain, whereas antagonizing TLR4 activity by blockade of either TLR4 or downstream signaling, or the use of TLR4 knockout mice, show potentiated morphine analgesia and attenuated morphine tolerance [117]. Furthermore, TLR4 binds (−)- and opioid receptor-inactive (+)- ligands nonstereoselectively; thus, studies have found that (+)- and (−)- naloxone and (+)- naltrexone stereoisomers block TLR4 signaling to suppress glial activation and neuropathic pain [36, 38, 116, 119, 120, 122, 129, 130]. While these studies indicate a potential role for TLR4 activation in OIH, other studies report conflicting evidence. For instance, a mouse model of OIH employed by the Kest lab utilizes concomitantly administered naltrexone in both acute and chronic models of OIH induced by either morphine [42] or fentanyl [44]. That is, while morphine analgesia is prevented in this paradigm, OIH persists in the presence of NTX, conceivably narrowing the possibility of morphine working directly via TLR4 to mediate hyperalgesia. However, further research is needed to determine the precise role, if any, that TLR4 has in OIH.

4.3 Neurotransmitter Mechanisms

Both the glutamate and GABA neurotransmitter systems have been implicated in the development and maintenance of OIH. Their proposed role in modulating opioid effects is summarized in the following sections.

4.3.1 Glutamate

The majority of research regarding the systems modulating OIH implicates a role for the NMDA receptor, for which glutamate is the endogenous ligand. The NMDA receptor system has a well-documented role in neural and behavioral plasticity, long-term potentiation, learning, and memory. In 1991, Trujillo and Akil [131] indicated a likely role for the NMDA receptor in the development of opioid tolerance and dependence, resulting from neural adaptations after repeated morphine exposure. As such, it is an excellent candidate receptor system for the study of the mechanism underlying opioid-induced effects, including OIH [131, 132].

NMDA receptors play a primary role in central sensitization and consequent inflammatory pain, although a knowledge base for the role it plays in peripheral sensitization and OIH is not comprehensively established [46, 133135]. As opioids do not have any binding affinity for NMDA receptors and NMDA receptor antagonists do not negatively affect acute opioid analgesia, it is implausible that opioids directly interact with this receptor system [31]. However, NMDA receptors are present in particularly dense concentrations in areas of the CNS that modulate pain signaling such as dorsal root ganglia, superficial layers of the dorsal horn laminae, the thalamus, and the hippocampus [93]. NMDA receptors in the spinal dorsal horn [31, 136, 137] and RVM [32, 33] appear to be particularly associated with OIH. Thus, several NMDA receptor antagonist compounds have been found to reduce OIH in both animal and human research, further illustrating a substantial role for this system in the modulation and potential treatment of this nociceptive state [39, 42, 46, 64, 88, 93, 138, 139].

4.3.2 GABA

The GABA receptor system and the opioid receptor system are closely intertwined, both neuroanatomically and functionally. For instance, GABA neurons express opioid receptors, particularly μ-opioid receptors, in multiple regions of the central nervous system critical to nociception [140, 141]. It has been suggested that chronic opioid administration leads to a downregulation of GABA receptors, resulting in decreased inhibition of pain signaling [36, 142].

Opioids exert their analgesic effects by inhibition of GABAergic neurons in the periaqueductal gray (PAG), a region critical to pain processing. Furthermore, opioid inhibition of GABA indirectly results in increased excitatory activity, and likely contributes excitatory opioid effects, including OIH [140, 143, 144]. For instance, a recent preclinical study demonstrated that morphine induces hyperalgesia by downregulating the K+/Cl- co-transporter KCC2, resulting in altered Cl- homeostasis in the dorsal horn. The authors suggest the clinical application to be the restoration of GABAA inhibition to restore Cl- homeostasis in the dorsal horn, as this was effective in preventing OIH induced by morphine [78].

Another promising avenue of research is revealed in neurosteroids acting via the GABA receptor system. Neurosteroids bind to GABAA receptors to modulate a number of therapeutic effects (reviewed in Sect. 6.3.), including neuropathic pain [145147]. As such, the GABA receptor system serves as a promising target of treatment in patients with OIH and warrants further investigation.

4.4 Opioid Metabolites

Subsequent to the finding that morphine metabolites possessed pronociceptive properties, it was initially thought that opioid metabolites mediate OIH. However, more recent preclinical studies have suggested otherwise. For instance, mice possess a variant UDP-glucuronosyltransferase (UGT) isoform involved in glucuronidation, which allows only for the formation of M3G and no M6G following morphine administration [148]. Thus, M6G is likely uninvolved in OIH, as mice do not produce this metabolite but nonetheless exhibit hyperalgesia [149]. Further, Swartjes et al. [93] demonstrated an indirect relationship between OIH and blood plasma concentrations of M3G in mice. Specifically, the group found that trough plasma levels of M3G corresponded to peak levels of morphine, at which point hyperalgesia worsened. This finding was detrimental in that it suggests morphine itself as the mediator of OIH; however, human studies have reported conflicting results regarding subjective pain relief and the ratio of morphine to its metabolites [150152]. Swartjes et al. [93] also reported that mice lacking the ability to export M3G from the liver into systemic circulation (MRP3−/− mice) also develop OIH after an acute morphine administration. Further, Arout et al. [45] reported that while morphine in combination with the opioid receptor antagonist naltrexone resulted in significantly less activation in the PAG during a hyperalgesic state, naltrexone combined with M3G resulted in no such reduction of activity. Finally, other opioids shown to produce hyperalgesia, such as fentanyl, do not yield any known pronociceptive metabolites [44, 153]. As such, a rationale for opioid metabolites in the modulation of hyperalgesia induced by administration of the parent opioid is implausible.

5 Prospective Treatments for OIH

It is possible that OIH manifests as an adverse reaction to one particular opioid in an individual; as such, clinical findings indicate that opioid rotation [154] is sometimes sufficient to eliminate hyperalgesia and allodynia [39, 50, 71, 155]. Other studies suggest that downward dose titration [156159] or complete discontinuation following detoxification [72, 91] is the most efficacious way to relieve this opioid-induced increased pain sensitivity. However, opioid discontinuation is not always a viable option; therefore, a solution that attenuates OIH such that the patient may continue benefitting from chronic opioids is needed. If a way to treat OIH can be elucidated, this will allow clinicians to continue prescribing opioids without concern over paradoxically increasing the very symptom they set out to treat: pain. Thus, adding adjuvant medications to opioids that target the receptor systems thought to underlie their hyperalgesic effects appears most promising (Table 3).

Table 3 Clinical pharmacological studies investigating opioid-induced hyperalgesia

As reviewed above, there are a multitude of receptor pathways implicated in the modulation of OIH. Thus far, a host of animal studies have found that OIH is reversible by the NMDA receptor antagonists MK-801 [31, 42, 160165], ketamine, or traxoprodil [93]. Human studies have found similar results by adding adjuvant ketamine [46, 54, 88, 166], as well as gabapentinoids such as pregabalin [49, 55] or gabapentin [23], or the cyclooxygenase-2 inhibitor parecoxib [48, 167], as reviewed in the subsequent sections. However, more recent research has shown that manipulating these substrates does not unequivocally resolve OIH. As such, other more recent studies have begun targeting other receptor systems in attempt to attenuate OIH in clinical treatment.

5.1 NMDA Receptor Antagonists

5.1.1 Ketamine

Ketamine is a potent noncompetitive NMDA receptor antagonist, first used for surgical anesthesia, and later found to produce analgesia through a possible opioid receptor-mediated mechanism [134, 168, 169]. There is an abundance of preclinical and clinical literature describing the ability of ketamine to reverse various pain states, ranging from chronic neuropathic pain contexts to OIH [170172]. A wealth of literature details ketamine’s use as an adjuvant to several opioids in a perioperative context; for example, it has been found to significantly reduce the amount of postoperative opioid consumption, attenuate incisional hyperalgesia (Table 3), and in some cases enhance analgesia in both animal and clinical investigations [66, 134, 138, 173180].

How precisely ketamine induces analgesia is somewhat nebulous; it demonstrates activity at several different receptor systems in addition to the NMDA and opioid receptor systems, and affects multiple neurotransmitter pathways including both the cholinergic and monoamine systems. Additionally, it is not entirely clear if the pain relieving properties are due to analgesia, or antihyperalgesia [134, 169]. Further, it has been suggested that ketamine’s analgesic properties may be partially mediated by its active metabolite, norketamine. Norketamine has a longer half-life than ketamine itself, is yielded in higher concentrations than ketamine following biotransformation, and is proposed to significantly contribute to ketamine’s analgesic effects [181183]. Nonetheless, ketamine’s use in clinical practice for pain is limited, despite promise in studies of chronic pain [184], as well as clinically applied and experimental pain models of OIH (Table 3), in part due to its cardiovascular and psychotropic side effects [172, 185]. Further, ketamine is administered intravenously in clinical practice, as oral formulations are not commercially available; thus, use of this drug for chronic pain management is currently challenging and warrants further research.

5.1.2 Dextromethorphan

An antitussive drug, dextromethorphan is typically used as an over-the-counter cough suppressant, but has shown promise in pain suppression. Dextromethorphan antagonizes the spinal NMDA receptor system, and may manipulate opioid effects via this mechanism [186, 187]. Several studies have shown mixed results with dextromethorphan as an adjuvant medication in treating opioid tolerance [188191] and withdrawal symptoms [192197]. Numerous studies have cited dextromethorphan’s ability to reduce intra- and post-operative morphine consumption [198200], while others have not replicated these findings [201, 202]. Accordingly, Compton et al. [203] sought to determine if dextromethorphan would alleviate OIH in methadone maintenance patients. Following a 5-week trial, this group found no effect of dextromethorphan on OIH, as assessed by the cold pressor test and electrical stimulation pain measures [203] (Table 3). To date, no other clinical trials have investigated dextromethorphan in the context of OIH during concurrent opioid use.

5.1.3 Other NMDA Receptor Antagonists

While Swartjes and colleagues [93] attenuated hyperalgesia in rodents by administration of ketamine, they found additional evidence for the novel NR2B selective NMDA receptor antagonist traxoprodil in reversing hyperalgesia. Others have illustrated analgesic actions of this compound in rodents [204]. Such findings suggest a promising role for the NR2B subunit in particular in the development of OIH; this subunit densely populates both spinal and supraspinal sites such as the dorsal root ganglia, superficial layers of the dorsal horn laminae, the thalamus, hippocampus, and cortex [93]. Perhaps most appealing is that, in comparison with ketamine and other potent NMDA receptor antagonists, due to its selectivity for the NR2B subunit, traxoprodil appears to have limited psychotropic side effects. Thus, traxoprodil may be a plausible option for preventing or alleviating OIH. However, it is unexplored in this context, as only four animal studies [172, 204206] have examined its potential use for the treatment of acute and chronic hyperalgesic states. The results of these studies suggest that although less potent when compared with ketamine, traxoprodil may be clinically efficacious for pain relief; further, it possesses few or no side effects when administered at analgesic doses.

Several studies have demonstrated OIH in patients enrolled in methadone maintenance treatment [23, 43, 5760]. However, other studies have documented a reduction in OIH after patients were rotated onto methadone treatment. For instance, Axelrod and Reville [71] reported the success of methadone in reducing OIH in the case of a young girl suffering from cancer pain, who had previously been treated unsuccessfully with a multitude of other opioids. Additionally, Mercadante and colleagues [39] reported an individual case in which adjuvant methadone effectively reduced fentanyl-induced hyperalgesia. More recently, Mercadante et al. [39] reported the successful rotation of 12 patients to methadone, whose pain was originally unaffected by morphine. Methadone exhibits weak NMDA receptor-antagonistic properties; as such, it is plausible that a subset of patients that develop OIH resulting from other opioids may benefit from rotation to methadone. However, methadone’s potential efficacy for OIH remains to be tested in randomized, controlled studies.

5.2 Gabapentinoids

There are two gabapentinoids that are used clinically: pregabalin and gabapentin. Pregabalin was designed as an antiepileptic drug and manipulates a multitude of neurotransmitters including glutamate, norepinephrine, and substance P [207]. A host of clinical research has elucidated a role for pregabalin in alleviating both postoperative pain and neuropathic pain, and in 2004, pregabalin was approved by the FDA for the treatment of the latter [208]. Subsequently, one study has directly investigated the potential effect of pregabalin in patients receiving opioids. Specifically, Lee et al. [55] reported the attenuation of putative OIH in patients receiving intraoperative high-dose remifentanil by a single preoperative dose of pregabalin. A second study by Martinez et al. [49] combined pregabalin with ketamine, and reported that while patients who received only ketamine or pregabalin both demonstrated a significant decrease in postoperative morphine requirement, the combination of the two resulted in a significantly larger decrease in postoperative morphine. Thus, the combination of ketamine and pregabalin putatively offers a larger clinical benefit than the two in isolation (Table 3). As the impact of pregabalin on OIH is unclear, future studies should further investigate any potential benefit.

The anticonvulsant drug gabapentin was originally designed as an analog of GABA for the treatment of spasticity. Since, it has shown promise in chronic pain states, although its mechanism for this indication is not clear. It appears as though gabapentin may inhibit NMDA receptors, calcium channels, and peripheral nerves through GABA-mediated pathways. While a host of literature exists describing the ability of gabapentin to effectively treat neuropathic pain, clinical pharmacological studies regarding its impact on OIH are nearly nonexistent [209]. In fact, there is only one direct investigation of gabapentin this context. Specifically, Compton et al. [23] found that gabapentin effectively reduced OIH over the course of 1 week, as measured by cold pressor test latencies, in former opioid addicts currently maintained on methadone replacement (Table 3). These promising findings need to be replicated in different clinical samples that display OIH.

5.3 Propranolol

A β-adrenergic antagonist, propranolol was first administered to treat hypertension, and later found to attenuate anxiety. Following the results of a genetic study linking the β-adrenergic receptor to OIH [210], speculation arose regarding the clinical therapeutic effect of propranolol in treating OIH. Accordingly, Chu et al. [47] conducted a clinical trial in which propranolol effectively eliminated secondary hyperalgesia in healthy volunteers receiving remifentanil infusion and undergoing experimental pain testing (Table 3). As such, propranolol may prove useful in precluding OIH; however, further studies are needed.

5.4 Cyclooxygenase Inhibitors

Cyclooxygenase (COX) is an enzyme that is important for the formation of prostaglandins, which are hormones with a large role in sustaining homeostasis and mediating physiological effects in the body [211]. One proposed function of prostaglandins is to regulate inflammation; accordingly, several studies have sought to investigate whether inhibition of this system would affect OIH by utilizing NSAIDs that directly target COX. Specifically, two experimental pain investigations in healthy volunteers found that administration of one such drug, parecoxib, differentially affected experimentally induced OIH following remifentanil infusion [48, 167] (Table 3). Further, Tröster et al. [167] reported that parecoxib augmented remifentanil-induced analgesia. Thus, COX inhibitors such as parecoxib may have therapeutic potential in alleviating OIH.

6 Future Considerations in OIH

6.1 Glial Targets

To date, all approved medications for pain work by manipulating neuronal activity; however, more recent preclinical studies have suggested that other mechanisms be considered. Specifically, Hathway et al. [212] demonstrated hyperalgesia resulting from microglial central sensitization following brief, low-frequency stimulation of C-fibers in the dorsal horn, preventable by minocycline pretreatment. Importantly, these findings may have important implications for OIH; that pathological as well as physiological stimuli, the latter of which does not result in tissue damage, can lead to a microglial-regulated central sensitization state that is akin to that postulated in aberrant chronic pain [98, 212]. Further, co-administration of glial inhibitors has been found to potentiate opioid analgesia [117, 213, 214]; while their effects on OIH in clinical populations remain to be extensively studied, opioid administration does result in increased proinflammatory cytokine release and subsequent analgesic opposition [115]. Thus, it is feasible that glial inhibitors may attenuate OIH.

In light of the growing recognition that microglia play an important role in the development of nociception, this system may serve as a novel treatment target for OIH. Accordingly, numerous groups are studying the potential of inhibiting the glial system to enhance treatment of pain. For instance, the glial inhibitor ibudilast (MN166) has shown promise in the treatment of allodynia in neuropathic pain in an animal model [215, 216]. Considering that neuropathic pain and OIH share similar neurobiology, it is quite possible that targeting this system in treatment would reduce OIH and, in turn, increase the efficacy of opioid treatment. Further, another study reported an ability of this drug to induce a >3-fold increase in morphine analgesic potency [214]. Though the majority of research on ibudilast is in the preclinical stage, it has completed safety testing in humans [217, 218] and is thus of interest for the treatment of OIH.

A host of other glial-modifying agents have undergone animal testing in models of pain, the results of which are reviewed in Hameed et al. [127]. Minocycline has also shown promise in studies of pain, particularly in an animal model of morphine effects. Though minocycline has nonselective actions at both neuronal and glial cell types within the nervous system [219], its anti-inflammatory actions are theorized to be a result of its inhibitory influence on the latter [220, 221]. Studies have shown this tetracycline antibiotic microglial inhibitor to reduce negative aspects of morphine treatment (respiratory depression, tolerance, and reward) while augmenting analgesia [213, 222]. This is of particular interest for treatment of OIH in populations with opioid addiction; it is possible the minocycline could alleviate OIH while also potentially dampening the risk of addiction resulting from long-term opioid treatment. Appropriately, there is currently one ongoing clinical trial investigating the effects of this drug on OIH in opioid-maintained patients (NCT02359006; ClinicalTrials.gov). Interestingly, minocycline is also reported to have neuroprotective effects [219, 222, 223]. Studies should be undertaken to examine the role of these drugs in alleviating OIH, as a host of literature highlights a potential role of glia in both pain and opioid-induced effects [38, 113, 115, 117, 212].

6.2 Cannabinoid Targets

Since the successful cloning of the cannabinoid receptors in the early 1990s [224, 225], many studies have examined the therapeutic potential of the cannabinoids in the treatment of chronic pain. Several studies have found no pain relieving benefits of delta-9-tetrahydrocannabinol (THC). For instance, Attal et al. [226] reported negative findings of oral THC in patients with chronic neuropathic pain. Interestingly, Wallace et al. [227] reported a dose-dependent analgesic effect of inhaled cannabis on experimental capsaicin-induced pain in a study of healthy volunteers. Specifically, they found no effect with a low dose, and analgesia resulting from a medium dose. Other studies have found negative results in postoperative acute pain patients, where oral or smoked cannabis yielded no benefit and in some cases enhanced pain [227230]. Redmond et al. [231] also reported no effect of nabilone in healthy volunteers undergoing experimental pain testing. These sporadic negative findings are perhaps due to differences in formulation, administration, and potentially a lack of effectiveness for said acute pain conditions. Overall, it seems that studies of cannabis under acute pain paradigms and in healthy volunteers do not yield reliable results [227231].

In the context of chronic pain, preliminary results indicate that agonists of this receptor system effectively reduce pain resulting from various conditions such as cancer, neuropathic pain, rheumatoid arthritis, multiple sclerosis, fibromyalgia, and HIV, putatively via anti-inflammatory properties [232242]. For example, Eisenberg et al. [243] reported successful use of inhaled cannabis for the treatment of chronic neuropathic pain; a single dose was found to reduce pain by 45 % within 20 min, lasting for approximately 90 min with minimal adverse reactions. Unfortunately, generalizability of such findings is complicated by variations in cannabis products employed, ranging from crude herbal cannabis to several pharmaceutically developed products (described in [244]). Additionally, there are a limited number of large-scale randomized clinical trials examining efficacy and determining risk–benefit [244]. With that said, a function for cannabinoids in pain is mechanistically logical, as this receptor system densely populates regions intricately involved in pain processing such as the PAG, RVM, the dorsal spinal cord, and microglia [244, 245]. As such, the cannabinoid system may yield interesting results when combined with opioid therapy to reduce pain sensitivity, thus potentially eliminating OIH. In fact, Narang et al. [233] found that the combination of dronabinol and opioids resulted in increased pain relief when compared with participants utilizing only opioids; this synergy between THC and opioids has been supported by some investigations [246], but not others [247]. However, no studies have evaluated the effect of cannabis on OIH. Future studies should investigate a potential relationship between opioids, the cannabinoid system, and OIH, such that cannabinoids may be used as adjuvants to long-term opioid treatment to attenuate OIH.

6.3 Neurosteroids

Neurosteroids are steroidal hormones that are synthesized in the brain, and exert inhibitory effects mainly via GABAA receptors, glycine, and L- and T-type calcium channels [145147, 248251]. Among the most common neurosteroids are progesterone, pregnanolone, and allopregnanolone. Studies have revealed neurosteroids to have a therapeutic role in a number of conditions such as epilepsy, drug withdrawal-induced seizures, anxiety, premenstrual syndrome, stress, depression, and alcohol withdrawal. Furthermore, it is possible that some neurosteroids are produced in the nervous system in response to injury, exerting anti-inflammatory effects at the site of insult [145, 251]. Of particular relevance to pain and opioid research is neurosteroids’ role in modulating neuronal plasticity and related activities of the nervous system, as well as their role in the sexual dimorphism of pain processing [145].

Subsequent to the finding that neurosteroids produce anesthesia in rat models and thus suppress pain sensation [252254], neurosteroids have been found to promote analgesia [251, 255], and have documented efficacy in the treatment of neuropathic pain [145]. Whereas lower levels of testosterone or progesterone have been associated with low nociceptive thresholds, injecting these steroidal hormones has been found to elicit analgesia [251, 256260]. Several studies have also found that higher levels of progesterone decrease anesthetic [261] and postoperative analgesic requirements [262]. For instance, Lee et al. [262] found that parturient women with higher than average serum progesterone levels required less anesthesia and postoperative analgesic during cesarean delivery.

Due to their neuroprotective and antinociceptive effects [146], neurosteroids are a possible treatment target for alleviating OIH. Allopregnanolone in particular has shown promise in treating neuropathic pain (reviewed in [145]), and thus may be a potential candidate to attenuate OIH. While clinical use of neurosteroids is complicated by low bioavailability following rapid inactivation during metabolism, they have a favorable adverse effect profile and are markedly safer than currently available analgesics [145]. An interesting speculation lies in the potential connection between the aforementioned putatively dysregulated Cl- homeostasis in OIH [78] and the use of neurosteroids to treat OIH; while the authors suggested the use of benzodiazepines to restore the GABA inhibition induced by opioids [78], this combination is hazardous in clinical practice and is increasingly advised against [263]. Thus, neurosteroids could potentially achieve the same restoration of GABA inhibition while minimizing hazardous contraindications.

7 Concluding Remarks

This review highlights the importance of OIH in hindering the analgesic effects of opioids, especially with long-term opioid use. Additionally valuable would be considering OIH, in addition to analgesia, as a treatment target in the drug development process. Presently, clinical studies of OIH focus primarily on identifying this phenomenon in opioid-receiving populations, and determining its clinical relevance in regard to tolerance. To a lesser extent, there are studies investigating prevalence and methods to treat it; however, several methodological issues are present. Of primary concern is that while numerous investigations study OIH, it is not unequivocally recognized by clinicians due to a lack of a valid assessment tool to reliably separate OIH from tolerance or withdrawal, and aid in their diagnosis of suspected OIH. As a result, there are only a few studies examining the prevalence of OIH in clinical samples. Similarly, there is no consensus among researchers or practitioners as to the most appropriate method to detect OIH, as studies comparing the sensitivity and specificity of paradigms for detecting OIH are nearly nonexistent [264]. Furthermore, it has been reported that manipulations producing profound OIH on one pain assay may show no hyperalgesic liabilities on other assays, suggesting that OIH is both drug and modality specific. Animal studies also suggest that it is influenced by varying genetic backgrounds [265268] and epigenetic factors [269], which needs to be further investigated in future human studies [270].

Secondly, most reports on successful treatment of OIH are case studies; there are few controlled experiments, and the few that exist are performed mostly in populations undergoing surgical procedures and subsequent short-term opioid treatment. In these scenarios, most patients are anesthetized with the short-acting opioid remifentanil, and subsequently given morphine in response to what is occasionally identified as remifentanil-induced hyperalgesia. While acute treatment of surgical pain is the scenario in which opioids were intended for, it is no longer the primary use, is not the sole context contributing to the rise in addiction, and thus does not accurately represent the modern day opioid-using population. Further, it is controversial as to whether these scenarios truly represent OIH, or are a manifestation of hyperalgesia resulting from opioid withdrawal. Future research should assess clinical manifestations of OIH in patients maintained on long-acting opioids for chronic pain conditions. Comprehensive analyses evaluating OIH in the entirety of opioid-consuming populations, using validated measures of pain with clinical relevance, are desperately needed, such that pain medicine can be fully informed with regard to this paradoxical, complicated state. OIH, in addition to analgesia, must be considered an aspect of chronic pain treatment, such that the efficacy of opioid therapy is increased by reducing the incidence of OIH in clinical practice. As highlighted by this review, there are several avenues of research that may prove fruitful in treating OIH via adjuvant pharmacotherapies. Perhaps the most promising route concerns glial mechanisms of inflammatory pain. As this area of study continues to bustle with new findings in the relationship between opioids and glia, we in the scientific community must address potential microglial inhibitors that may serve as effective adjuvants to opioids. Secondly, as the medical use of cannabinoids expands, we will likely see an increase in the number of patients reporting cannabis use for pain relief. As such, research is needed to assess the efficacy of cannabinoids in reducing OIH.

Finally, OIH has been shown to manifest differently in females as compared with males, yet preclinical and clinical studies including both sexes are strikingly limited. Most studies utilize only males to eliminate hormonal influences, a variable that is proving to be of increasing importance in the effective treatment of chronic pain. The sexual dimorphism of OIH and pain in general is a fundamental characteristic that must not be ignored; future studies must be diligent in including both male and female participants, such that we may obtain the most accurate understanding of this paradoxical state of nociception.