Local Anesthetics ( Box 43.1 )

The most widely used drugs for neuraxial analgesia are local anesthetics, which produce a reversible conduction blockade of impulses in peripheral and central nerves. Local anesthetics have proven to be safe and reliable in interrupting nerve impulse conduction in both peripheral and central nerves. Local anesthetics are most commonly used to provide surgical anesthesia, postoperative pain relief, and relief of cancer pain. The discussion of local anesthetic agents will be brief.

Electrical impulses in the form of action potentials are conducted along nerve fibers. The propagation of electrical impulses along nerve fibers in the form of action potentials is responsible for transmitting sensory and motor information along the nerves. Action potentials are dependent on maintaining a resting membrane potential of approximately 60 to 70 mV, which requires various pumps and channels to sustain an electrochemical gradient. The most important channel is the voltage-gated sodium channel, which allows the influx of sodium ions during the depolarization phase of the action potential, thus enabling an impulse to travel down the nerve fiber. These sodium channels are complex three-dimensional structures integrated into the membrane lipid bilayer of the nerve cells. Local anesthetic agents act by interrupting the propagation of an impulse, thereby inhibiting the nerve conduction of a painful stimulus. Local anesthetics gain access to the sodium channel from either the plasma or cytoplasmic side of the channel protein, and bind within the pore of the channel. The binding of sodium channels by local anesthetics is state dependent. The local anesthetic can bind to the channel when it is open (active), closed (inactivated), or in a resting state. However, the receptor has the highest affinity when the channel is open or closed and the lowest affinity in the resting state. These processes are not sensory specific; thus, local anesthetics are capable of blocking the transmission of all nerve fibers, not just A-delta and C fibers. Therefore, the potential for motor blockade, as well as sensory blockade, limits the use of these agents.

Opioids ( Box 43.2 )

Opioid analgesics exert their actions through the inhibition of target cell activity. The existence of multiple types of opioid receptors was originally proposed by Martin and colleagues. Subsequent studies using both in vitro and in vivo pharmacologic methods, and employing alkaloid-derived and synthetic compounds, provide support for the existence of multiple opioid receptor subtypes, including mu, delta, and kappa receptors. Molecular cloning techniques have, thus far, identified three gene families that encode for these receptors. In addition, an “orphan” opioid receptor (ORL1), which shares substantial sequence homology but does not bind prototypic opioid receptor agonists with high affinity, has also been identified and structured. All of the opioid receptors belong to the GTP-binding protein superfamily of metabotropic receptors. Agonist binding to opioid receptors results in activation of inwardly rectifying potassium channels, the inhibition of N-type and L-type calcium channels, or the inhibition of adenylate cyclase activity, all processes by which neuronal excitability can be suppressed.

Whereas peripherally located opioid receptors have been identified, the predominant analgesic sites are believed to reside in the central nervous system (CNS). In the brain, these receptor sites include the brainstem, thalamus, forebrain, and mesencephalon. In the spinal cord, they include the postsynaptic receptors located on cells originating in the dorsal horn, as well as presynaptic receptors found on the spinal terminals of primary afferent fibers.

The effects of opioids are determined not only by their affinity for endogenous receptors but also by their ability to reach those receptors. The onset of analgesia is similar for intrathecal and epidural narcotics, suggesting that the penetration of neural tissue—and not the meninges—is the rate-limiting step. Intrathecal opioids exert their analgesic properties by presynaptically inhibiting the release of glutamate, substance P, and calcitonin gene-related peptide, molecules believed to be responsible for transmitting nociceptive signals across synapses. Epidurally administered opioids may operate via an additional mechanism. The systemic absorption of an epidural bolus of lipophilic opioids (e.g., fentanyl and sufentanil) is similar to that which follows an intramuscular injection and may play a role in the analgesic effects. The conflicting theories regarding how epidural lipophilic opioids exert their pain-relieving properties may be partially explained by the differing modes of administration. For example, a bolus of epidural fentanyl appears to produce analgesia mostly via spinal mechanisms, whereas uptake into the systemic circulation plays a major role in the analgesic effects produced by continuous epidural infusion. In contrast, hydrophilic opioids as morphine are more likely to diffuse across dural membranes where their primary analgesic effect is exerted through receptors in the dorsal horn.

Lipid solubility determines, in part, several other important characteristics of intrathecal opioids, including the spread of analgesia and side effects. Opioids that are highly water soluble, such as morphine, exhibit a greater degree of rostral spread when injected into the subarachnoid or epidural space than lipid-soluble compounds, so that in pain conditions requiring higher spinal levels or more extensive coverage, the degree of analgesia they confer may be superior. Conversely, because many of the adverse effects of spinal opioids, such as nausea, vomiting, and delayed respiratory depression, are the result of interaction with opioid receptors in the brain, the more water-soluble compounds are associated with a higher incidence of these problems.

Whereas the earliest studies evaluating the chronic use of intrathecal opioids were conducted in patients suffering from cancer pain, other studies have found intrathecal and epidural narcotics to be effective in nonmalignant pain as well. These conditions include not only nociceptive pain but also neuropathic pain, a heterogeneous group of disorders originally believed to be resistant to narcotics. Certain aspects of neuropathic pain, such as tactile allodynia, may be less responsive to the effects of spinal opiates. Therefore, many neuropathic conditions require adding nonopioid adjuvants to spinal opioids for successful pain relief. The caveat with the treatment of nonmalignant pain via intrathecal opioids is that these patients have a higher likelihood of complications resulting from the opioid therapy because of their normal life expectancy compared to those with cancer. Aside from the common intrathecal opioid side effects such as tolerance, constipation, sweating, nausea, and urinary retention, these patients are at risk for the development of intrathecal granulomas. Factors that are positively associated with the development of catheter tip granuloma include drug dosage, drug concentration, and the duration of intrathecal opioid therapy. In 2008, Deer and associates published recommendations for the maximum daily dose and maximum concentration of intrathecal medications ( Table 43.1 ).

When opioids are administered directly into the cerebrospinal fluid, only a fraction of the systemic dose is required, as there are no anatomic barriers to be crossed, and vascular reuptake is slow. Not all side effects of intrathecal opioids are dose related, but in many instances, the drastic reduction in dosage translates into reduced side effects. In fact, one of the primary indications for a trial with intrathecal or epidural narcotics is a good analgesic response to systemic opioids coupled with intractable side effects. Among the adverse opioid effects reduced by switching from oral formulations to spinal administration are sedation and constipation. Effects that may be increased include pruritus, urinary retention, and edema. The mechanisms contributing to the various adverse effects of opioids are incompletely understood but probably multifactorial and include those that are mediated via interaction with specific opioid receptors and those that are not. Undesirable effects not mediated by opioid receptors, such as CNS excitation and hyperalgesia, cannot be reversed with naloxone. The incidence of various opioid-induced side effects depends on a number of different factors including the opioid infused, route of administration and dosage, extent of disease, concurrent drug (e.g., oral opioids, adjuvants), age, concomitant medical problems, and prior exposure to opioids. The most frequent side effects of intrathecal morphine are constipation, sweating, urinary retention, nausea and vomiting, and disturbances of the libido.

Calcium Channel Antagonists ( Box 43.3 )

Voltage-dependent calcium channel (VDCC) conduction plays an integral role in pain transmission. Diversity among these voltage-gated calcium channels was originally described through differences in the biophysical properties of calcium currents recorded from individual neurons. Multiple distinct voltage-gated calcium currents were observed, including the L-type responsive to dihydropyridines, the T-type responsive to ethosuximide, and the N-type and P/Q-type responsive to conotoxins. Molecular cloning has led to the identification of multiple genes encoding the calcium channels that correspond to the biophysical and pharmacologic profiles of the receptor subtypes. VDCCs are located in the plasma membrane of all excitable cells, including the neurons of the peripheral and central nervous system. These calcium channels are found in high concentrations in the dorsal horn of the spinal cord and dorsal root ganglia.

The neuraxial administration of some VDCC modulators has been shown to produce antinociception, antiallodynia, and antihyperalgesia in animals. High threshold VDCCs, including N- and P/Q channels, are primarily found at synaptic sites involved in the release of transmitters; L-type channels are mainly observed at cell bodies and dendrites. Several VDCCs have been found to co-localize in the same neuron; therefore, antagonists to a given channel type usually block only a fraction of the VDCCs present and hence may have additive effects when combined. Finally, the relative importance of the various channel types depends on the functional status of the neuron. For instance, in acute models of nociception in uninjured naïve animals, there is a large body of evidence implicating N types, limited evidence for L types, and no evidence for P/Q types in producing antinociception. However, under conditions of persistent nociception induced in animals by chemical, inflammatory, or neuropathic stimuli, all three of these subtypes have been found to have anti-allodynic or antihyperalgesic properties. Although the physiologic and pharmacologic properties of VDCCs are largely determined by the molecular identity of the α1 subunit, the auxiliary subunits, including a ₂ d ₁ , also play a substantial role in the receptors’ characteristics. Gabapentin and pregabalin are two medications originally designed as structural analogs of GABA; however, neither drug acts as an agonist at GABA _A or GABA _B receptors, and neither drug acutely alters GABA uptake. It is likely that their analgesic effects are mediated at the a ₂ d ₁ subunits of VDCCs, for which both have a substantial affinity. Intrathecal gabapentin has no effect on an animal’s nociceptive threshold in the uninjured state, but it does possess antihyperalgesic and anti-allodynic properties in animal models of chronic pain. A substantial portion of the effects exerted by gabapentin are mediated by the N-type VDCC.

Ziconotide is a synthetic form of the peptide, ω-conotoxin MVIIA, which is isolated from the venom of the marine cone snail, Conus magus. It potently blocks N-type VDCCs in vitro and inhibits wind-up. The process of wind-up involves the activation of spinal NMDA receptors after injury, which induces a state of facilitated processing from repetitive small afferent fiber stimulation. This in turn leads to an increased response to high and low threshold stimulation and enhanced receptor field size. Ziconotide is the first member in the new drug class of selective N-type voltage-sensitive calcium channel blockers. It has been approved by the U.S. Food and Drug Administration and the European Medicines Agency for intrathecal treatment of patients with severe chronic pain that is refractory to other treatment modalities. Ziconotide blocks N-type calcium channels in the spinal cord and inhibits release of pain neurotransmitters from the central terminals of primary afferent neurons. By way of this mechanism, ziconotide has been found to effectively reduce pain. In a multicenter, double-blind, placebo-controlled study evaluating intrathecal ziconotide for the treatment of refractory pain in 111 patients with cancer and acquired immunodeficiency syndrome (AIDS), Staats and colleagues found that the treatment group obtained significantly enhanced pain relief compared to the control group (53% versus 18% improvement). The observation that there was no loss of efficacy for ziconotide in the maintenance phase is consistent with animal studies showing the absence of tolerance with calcium channel blockers. The most common side effects noted in this study were confusion, somnolence, and urinary retention; however, 97% of patients receiving ziconotide experience side effects, and 30% had serious adverse events. All side effects were reversible, with their incidence decreasing after the initial dosing period. In a randomized, double-blind, placebo-controlled trial, Wallace and colleagues demonstrated that intrathecal (IT) ziconotide provides relief in patients suffering from severe chronic nonmalignant pain who are unresponsive to conventional therapy. Patients were treated over a 6-day period and were found to respond to dosing of 0.1 µg/hr to 2.4 µg/hr. The original starting dose of 0.4 µg/hr was decreased due to a high incidence of side effects. The mean reduction in pain score from baseline was 31.2% in the treatment group compared to 6.0% in the placebo group. Patients in the treatment group reported suffering more side effects, including abnormal gait, amblyopia, dizziness, nausea, nystagmus, pain, urinary retention, and vomiting. However, Rauck and associates demonstrated that slow titration can provide comparable pain relief while minimizing the side effects experienced by patients. Wallace and colleagues showed that ziconotide could be safely given for long-term treatment of refractory pain. More than 644 patients with severe chronic pain participated in this open-label, multicenter longer-term trial. Thirty-two percent of patients receiving ziconotide for over 360 days, with a dose range between 0.048 µg/day and 240 µg/day, experienced greater than or equal to 32.7% improvement in pain scores. Although only FDA approved as single therapy ( Table 43.2 ), emerging evidence suggests that ziconotide can be safely combined when administered intrathecally with baclofen, morphine, sufentanil, and bupivacaine. Webster and colleagues added morphine to stable IT ziconotide therapy in 25 patients, which resulted in a mean 26.3% reduction in pain and a 49.1% reduction in opioid consumption at week 4 of treatment. This study suggests a synergism between morphine and ziconotide. When IT ziconotide is combined with other drugs such as opioids and baclofen, the stability declines, which may necessitate more frequent refills.

More recently, ziconotide has been anecdotally reported to successfully treat neuropathic pain and trigeminal neuralgia. The main limitations regarding the use of IT ziconotide are its steep cost and the high incidence (> 80% in many studies) of side effects, which can include the neurologic, psychiatric, cardiovascular, gastrointestinal, and genitourinary systems. Although in some guidelines ziconotide is now considered to be a first-line treatment for chronic pain, this therapy has an extremely narrow therapeutic window with substantial side effects and should be reserved for patients for whom other agents and therapies have been exhausted.

The epidural administration of VDCC modulators has also been examined. In a double-blind study conducted in healthy cohorts, adding a low dose (5 mg) of the L-type calcium channel blocker, verapamil, to epidural bupivacaine both pre- and postsurgical incision was found to reduce postoperative analgesic requirements in patients undergoing abdominal surgery. In a case report, Filos and collaborators described a modest and short-lived analgesic benefit for epidural nimodipine in two terminal cancer patients. The authors reported significant discomfort upon administration, but with no increased incidence of sedation, mood disturbances, or hypotension.

Gamma-Aminobutyric Acid Agonists ( Box 43.4 )

Three types of GABA receptor are currently recognized: GABA _A , GABA _B , and GABA _C . Only GABA _A and GABA _B receptors are present in significant quantities within the CNS. The GABA _A receptor is a ligand-gated ion channel. Activation of this receptor by GABA results in an influx of chloride ions and stabilization of the membrane potential, which decreases neuronal excitability. The receptor possesses two binding sites for GABA, as well as sites at which barbiturates, inhalational anesthetics, neurosteroids, and benzodiazepines bind to modulate the action of GABA. The receptor itself is a pentameric arrangement of different subunits. In contrast, the GABA _B receptor is a metabotropic receptor. Activation of the GABA _B receptor by GABA results in activation of inwardly rectifying potassium channels, inhibition of calcium channels, or inhibition of adenylate cyclase activity, all of which suppress neuronal excitability.

In laboratory animals, GABA _A agonists including muscimol, isoguvacine, and midazolam have anti-allodynic and antihyperalgesic effects in chronic pain models, whereas the GABA _A antagonists, bicuculline and picrotoxin, induce allodynia and hyperalgesia in naïve rats. However, GABA _A receptors are also closely linked to large-diameter afferents and, thus, are likely involved in modulating innocuous sensations as well. Similar to GABA _A receptors, GABA _B receptors are found in greatest abundance in the superficial dorsal horn of the spinal cord. Within the GABAergic system, nociceptive transmission is primarily regulated primarily by GABA _B receptor activity. The GABA _B receptor agonist, baclofen, blocks the activity of peripheral C and A-delta nociceptive fibers. Further, in the spinal cord, GABA _B receptors are found on interneurons as well as on terminals from primary afferent neurons. GABA _B receptor agonists administered via the intrathecal or epidural route produce pre- and postsynaptic inhibition, and therefore block the release of glutamate, substance P, and calcitonin gene-related peptide (CGRP) from primary afferents, and GABA from interneurons.

In naïve animals, the intrathecal administration of antagonists to either receptor increases pain behaviors, suggesting that tonic release of GABA in the spinal cord prevents innocuous stimuli from being perceived as noxious. In animal models of acute and persistent nociception, GABA _B agonists such as baclofen produce antinociception and anti-allodynia, respectively, at doses in which no motor impairment is observed. In contrast, the highly selective GABA _A agonist, muscimol, was found to be effective only in animal models of persistent nociception. Interestingly, in the same study, midazolam, the only GABA _A agonist presently available for human use, was ineffective in either acute or persistent pain models at doses that did not produce substantial motor impairment. In summary, the results from animal studies suggest that both GABA _A and GABA _B receptor agonists could be used to treat chronic pain.

In human studies, both GABA _A and GABA _B agonists have been shown to contain analgesic effects when injected into the intrathecal or epidural space. The literature shows that the neuraxial administration of GABA _A agonists in combination with local anesthetics increases the duration of motor and sensory block, increases the time to first analgesic request, and decreases postoperative analgesic requirements. The administration of the GABA _A agonist, midazolam, via either the intrathecal or the epidural route in combination with a mixture of other analgesics, including local anesthetics and opioids, was found to reduce opioid requirements and enhance postoperative analgesia for a variety of different surgical procedures. Ghai and associates found that adding 20 µg/kg/hr of midazolam to a continuous postoperative epidural of 0.125% bupivacaine reduced the requirement for rescue analgesia in children following upper abdominal and flank surgery. In a meta-analysis of 13 randomized controlled trials involving 672 patients, Ho and Ismail found that adding midazolam to other spinal medications delayed the time to request for rescue analgesia and reduced the incident of nausea and vomiting. A prospective, randomized, double-blind study by Shadangi and associates involving patients undergoing elective lower abdominal, lower limb, and gynecologic procedures found that adding midazolam to bupivacaine significantly improved analgesia. One hundred patients were randomized to receive a combination of either 0.4 mL (2 mg) midazolam with 3 mL of 0.5% bupivacaine or 0.4 mL normal saline with 3 mL of 0.5% bupivacaine. The duration of sensory blockade was prolonged in the group receiving midazolam by over 25 minutes without increasing motor blockade. In a double-blind study evaluating the effects of adding intrathecal midazolam to bupivacaine in patients undergoing hemorrhoidectomy, the addition of midazolam was found to expedite the onset of spinal analgesia in a dose-dependent manner. Other studies have found intrathecal midazolam to be effective in treating chronic mechanical low back pain, musculoskeletal pain, and neurogenic pain. In an open-label study, Prochazka and colleagues demonstrated the effectiveness of midazolam in the treatment of chronic low back pain and failed back surgery syndrome. Midazolam 2 to 5 mg was administered 500 times in 126 patients from 1995 to 2010. The analgesic effect lasted 9.7 weeks, with 65% of patients experiencing relief lasting 4 weeks or longer. The administration of subarachnoid midazolam was shown to be effective in treating pain associated with chronic, nonmalignant pain.

The safety of neuraxial administration of midazolam is controversial. Numerous animal studies have shown evidence of neurotoxicity with intrathecal midazolam. Svensson and associates found histologic evidence of neuronal death in the spinal cords of rats after 20 consecutive days of 100 µg intrathecal injections of midazolam. Unfortunately, the relative doses and concentrations of midazolam used in these animal studies were many times higher than those used in human studies. Malinovsky and colleagues reported that midazolam (1 mg/mL) produced histologic pathology that was greater than that of lidocaine or saline controls. Similar results were obtained by Erdine and colleagues, who found that rabbits infused with both preservative-containing and preservative-free intrathecal midazolam (300 µg daily, 1 mg/mL) over 5 days displayed vascular and other histologic spinal cord lesions on microscopic examination. Part of the controversy revolves around the pH of the medications administered. In an attempt to resolve this controversy, Bozkurt and colleagues administered normal saline, midazolam, and a saline vehicle control with the same acidic pH as midazolam epidurally in newborn rabbits. During electron microscopy examination on days 2 and 7, both the acidic midazolam and saline groups displayed significant pathologic spinal cord changes, such as degeneration of vacuoles, cytoplasm, and neurofilaments, disruption of myelin sheaths, lysis of cell membranes, perivascular edema, and pyknosis of nuclei. In contrast, the normal pH saline group displayed normal histology on spinal cord sectioning. More recent literature evaluating multiple doses of commercially available concentrations of midazolam in two species of animals found no histopathologic or behavior differences compared with saline controls. At present, the use of midazolam as a spinal analgesic is not FDA approved due to safety concerns. However, in clinical studies, there have been no reported adverse cardiovascular (hypotension or bradycardia), urologic (urinary hesitance or incontinence), or gastrointestinal (nausea or vomiting) side effects when compared to administering the local anesthetic alone. Prochazka and associates injected doses of 0.02 to 0.06 mg/kg into 14 patients with 10 administrations or more and reported that none of them displayed any clinical signs of neurotoxicity, such as bladder or bowel dysfunction, motor or sensory deficits, or new neuropathic pain. Canavero and colleagues reported successful intrathecal treatment of neuropathic pain with midazolam in a 60-year-old woman. They reported that intrathecal midazolam ranging from 2 to 4.6 mg successfully controlled neuropathic pain for over 6½ years. In one human study involving 1100 patients who were either administered intrathecal local anesthetic or local anesthetic and midazolam 2 mg for surgery, the authors reported no increased incidence of postoperative neurologic signs or symptoms or any other difference in complication rates between the two groups. Midazolam is currently considered a fifth-line neuraxial treatment for chronic pain.

Numerous studies support the use of neuraxial baclofen for spasticity in humans. There are multiple uncontrolled studies showing intrathecal baclofen to be effective for a variety of central pain conditions as well, including stroke, phantom limb pain, spinal cord injury, cerebral palsy, amyotrophic lateral sclerosis, and multiple sclerosis. Intrathecal baclofen has also been found to be an effective treatment for peripheral neuropathic pain conditions, such complex regional pain syndrome. In a small study by Lind and coworkers, the authors found that the addition of intrathecal baclofen in patients being treated with spinal cord stimulation for neuropathic pain improved pain scores to a greater degree than concomitant treatment with oral baclofen. The beneficial effect of intrathecal baclofen was dose dependent, peaking at 50 µg. Furthermore, at follow-up Lind and colleagues found that patients who were treated for a mean of 67 months enjoyed the same degree of pain relief, requiring only a modest (30%) dose increase. More recently, van Rijn and associates found that intrathecal baclofen provides substantial improvement in patients suffering from dystonia and in those with complex regional pain syndrome (CRPS). Thirty-six patients received a pump for continuous intrathecal baclofen for 12 months. Patients experienced a decrease in pain and disability as well as an improved quality of life. However, patients did experience a high complication rate, most often associated with pump or catheter system defects. Intrathecal baclofen administration as a therapeutic modality remains promising and would benefit from more research. Neuraxial baclofen has also been found to be beneficial in the treatment of musculoskeletal pain. Loubser and Akman found that intrathecal baclofen reduced musculoskeletal, but not neurogenic, pain in 12 patients with chronic spinal cord injury–related pain. Based on the results of this study and the temporal disparity regarding its analgesic effects on central pain and muscle spasm, it is likely that different pain-relieving mechanisms exist for these two conditions.

At therapeutic doses, baclofen is associated with numerous adverse effects, including drowsiness, flaccidity, headache, confusion, hypotension, weight gain, constipation, nausea, urinary frequency, and sexual dysfunction. Intrathecal baclofen overdose can lead to respiratory depression, seizures, obtundation, and, if not adequately treated, death. Withdrawal from abrupt cessation of intrathecal baclofen treatment can last days to weeks, or possibly longer, and can also be life threatening. There is growing evidence to suggest that replacement with oral baclofen may not always be adequate to control the symptoms, possibly because oral baclofen does not reach the central concentrations near the range achieved with intrathecal administration. Some studies suggest that tizanidine may represent a viable option for patients with spasticity and blood pressure lability, specifically with hypertension. Currently, there is animal model evidence to support intrathecal tizanidine for neuropathic pain, but it is not approved for use in humans. Baclofen is currently considered a fourth-line treatment for chronic pain ( Table 43.3 ).

Adrenergic Agonists ( Box 43.5 )

Adrenergic agonists have analgesic effects when applied for the management of chronic pain. Alpha-adrenergic receptors are widely distributed throughout the body and consist of two clinically significant classes: alpha ₁ and alpha ₂ . Alpha ₁ receptors are found in the smooth muscle cells of the peripheral vasculature and play an essential role in the regulation of systemic vascular resistance; they have no known significant role in analgesia. Alpha ₂ receptors however, are present throughout the peripheral and central nervous system and play a substantial role in modulating pain signals. There are several known subunits of the alpha-receptor: 2a, 2b, and 2c. Clonidine binds to pre- and postsynaptic α2 receptors in the dorsal horn. Activation of these receptors depresses presynaptive C-fiber transmitter release and hyperpolarizes the postsynaptic membrane through the Gi-coupled potassium channel. A different physiologic effect transpires, depending on the particular alpha subunit involved. The neuronal responses can be either inhibitory or excitatory. For example, the 2b subtype produces hemodynamic responses (primarily hypotension), whereas the 2a receptor is responsible for analgesia. The mechanism of action of neuraxial alpha ₂ agonists is similar to that for opioids in that they can exert effects on presynaptic and postsynaptic neurons. On the presynaptic neuron, they bind to alpha ₂ receptors of primary afferent neurons, resulting in hyperpolarization and depressed release of neurotransmitters involved in pain transmission. On postsynaptic neurons, alpha ₂ agonists hyperpolarize the cell by increasing the transmission of potassium through G _i -coupled potassium channels. Among the alpha-adrenergic agonists used for analgesia, clonidine remains the prototypical nonselective alpha ₂ agonist, partly because it has been the most studied ( Table 43.4 ). Clonidine is a nonselective agonist that produces antinociception via interaction with the alpha 2a receptor. However, it also produces substantial hemodynamic side effects due to its interaction with the 2b receptor. A newer agent, dexmedetomidine, is a selective alpha _2a receptor agonist that contains analgesic and sedative properties, with less respiratory depression and fewer cardiovascular effects. Dexmedetomidine demonstrates increased selectivity of the alpha 2a subunit than previous alpha-adrenergic agonists. Alpha-adrenergic agonists have also been shown to activate spinal cholinergic neurons, which may contribute to their analgesic effects. In addition to their antinociceptive properties, alpha ₂ agonists can produce dose-dependent sedation, presumably by inhibitory mechanisms involving the brainstem.

Alpha ₂ agonists have been administered intrathecally in humans since 1985. The antihypertensive medication, clonidine, is the most studied alpha ₂ agonist for neuraxial use. Although it is FDA approved for epidural use only in cancer pain, clinical reports have shown it to be effective intrathecally and epidurally for nonmalignant pain as well. Intrathecal clonidine has been reported to provide significant analgesia in combination with opioids for neuropathic pain and cancer pain, and it has been used in the pediatric population for these reasons. Clonidine has not consistently been shown to be effective as a single agent, but studies have shown it to prolong and enhance the effects of spinal and epidural anesthesia when coadministered with local anesthetics. When added to opioids, clonidine can extend the duration of pain relief for labor analgesia and postoperative pain. However, for acute pain, the evidence that adding clonidine to an epidural or intrathecal opioid is more effective than either analgesic alone is weak and inconsistent. Neuraxial clonidine has shown efficacy in treating central pain and spasticity after spinal cord injury. Studies show that dexmedetomidine prolongs local anesthetic motor and sensory blockade when compared to opioids. Alpha ₂ receptor agonists may also be suitable for patients suffering from neuropathic pain. In a randomized, placebo-controlled trial evaluating epidural clonidine in refractory reflex sympathetic dystrophy, Rauck and colleagues found that 300 µg of clonidine was equally effective but associated with less side effects than 700 µg.

Wu and collaborators compared preoperative epidural clonidine followed by patient-controlled epidural analgesia (PCEA) with clonidine, morphine, and ropivacaine to a control group who received preoperative epidural saline followed by PCEA with morphine and ropivacaine in 40 patients scheduled for elective colorectal surgery. Patients in the clonidine group exhibited longer PCEA trigger times, lower pain scores at rest and while coughing, less morphine consumption, and a faster return of bowel function throughout the 72-hour postoperative period compared with patients in the control group. Interestingly, the concentration of certain proinflammatory cytokines was also decreased in the clonidine group 12 and 24 hours following surgery. The most common side effects of neuraxial clonidine are sedation, hypotension, nausea and vomiting, and bradycardia. Hypotension and bradycardia are likely the result of alpha ₂ effects on preganglionic fibers in the thoracic spinal cord. Sedation results from their action exerted at supraspinal sites.

The interest in dexmedetomidine has grown in recent years because of its decreased cardiovascular and respiratory effects. Kanazi and colleagues showed, in a prospective, double-blind study, that 60 patients undergoing transurethral resection of prostate or bladder tumor under spinal anesthesia experienced a prolonged duration of motor and sensory blockade, with no increase in adverse hemodynamic effects or level of sedation. Patients were randomized to three groups receiving either dexmedetomidine (3 µg) or clonidine (30 µg) added to bupivacaine or bupivacaine alone. These results suggest that low-dose dexmedetomidine has similar analgesic efficacy to clonidine.

Clonidine has been studied extensively in animals and has produced no evidence of neurotoxicity. The number of studies involving human exposure continues to grow, and these studies have not revealed any clinical evidence of neurotoxicity. Clonidine, therefore, appears to be a drug that can be administered safely in humans via the spinal route.

Glutamatergic Receptor Antagonists ( Box 43.6 )

Similar to GABAergic and cholinergic receptors, glutamatergic receptors are divided into the G-protein-coupled (metabotropic) receptors (mGluR) and ion channel (ionotropic) receptors, which include NMDA, AMPA, and kainite receptors. The NMDA receptors contain ion channels permeable to calcium, sodium, and potassium. The NMDA ion channel is somewhat unique in that ambient concentrations of magnesium block NMDA responses in a use- and voltage-dependent manner. In addition to the glutamate binding site, the NMDA receptor has binding sites for glycine, which functions as an obligatory co-agonist, phencyclidine-like compounds, and endogenous protons and polyamines. Endogenous protons inhibit NMDA receptors via their interactions with an extracellular proton sensor on one of the receptor’s subunits. Endogenous polyamines such as spermine and spermidine bind at a separate site but shield this proton receptor, thereby potentiating NMDA receptor activity. The AMPA and kainate receptors function as ion channels that are permeable to sodium and potassium. Activation of the ionotropic receptors results in depolarization and neuronal excitation. Metabotropic glutamate receptors are divided into three groups. Group I receptors are positively coupled to phosphatidylinositol hydrolysis, and activation of these receptors ultimately increases intracellular calcium levels. The second and third groups of mGluRs are similar to opioid receptors in that they are negatively coupled to adenylate cyclase; hence, activation of these receptors results in neuronal inhibition.

In animal studies, intrathecally applied, selective agonists at the NMDA, AMPA, kainite, and group I mGlu receptors produce spontaneous pain behavior in naïve animals, and allodynia and hyperalgesia in neuropathic and inflammatory models of persistent pain. Antagonists to these receptors can reverse these nociceptive responses.

For humans, there are no AMPA, kainate, or mGluR agonists or antagonists available for clinical use. As such, the most studied glutamate modulators are the NMDA receptor antagonists. Perhaps the most studied NMDA antagonist for neuraxial use is ketamine, a noncompetitive NMDA antagonist that has been administered both epidurally and intrathecally in humans for acute and chronic pain relief. Following tissue injury, the activation of spinal NMDA receptors induces a state of facilitated processing from repetitive small afferent fiber stimulation, leading to an increased response to high and low threshold stimulation and enhanced receptor field size. This process, known as “wind-up,” is thought to be responsible for phenomena such as allodynia and hyperalgesia.

In a case report by Kristensen and colleagues, the spinal administration of CPP (3-[2-carboxypoperazin-4-yl]propyl-1-phosphonic acid), a competitive NMDA antagonist, was noted to suppress wind-up, but not spontaneous pain or allodynia in a patient with a peripheral nerve injury. Four hours after the last injection of CPP, psychomimetic side effects developed that were attributed to the rostral spread of medication. In combination with intrathecal morphine or other analgesic agents in patients suffering from cancer pain, the addition of intrathecal ketamine was shown to enhance the analgesic effects of opioids and other drugs while reducing the development of tolerance. In two case studies, Selda and colleagues found that adding low doses of epidural ketamine to morphine and bupivacaine improved analgesia while minimizing the side effects in patient suffering from terminal cancer with neuropathic components of pain. Bion reported that hyperbaric intrathecal ketamine mixed with epinephrine provided adequate short-term anesthesia in young soldiers undergoing field surgery. More recently, Murali Krishna and coworkers demonstrated that a combination of low doses of intrathecal ketamine and midazolam with bupivacaine improved postoperative analgesia in patients undergoing orthopaedic surgery. However, an open-label study by Hawksworth and Serpell conducted in 10 male patients undergoing prostate surgery found that the high frequency of psychomimetic disturbances, the short duration of action, and the high incidence of incomplete anesthesia precluded its use as a sole anesthetic agent. Similar findings were reported by Kathirvel and colleagues in a prospective study involving 30 healthy women undergoing a brachytherapy application for cervical cancer. The authors found that although the addition of 25 mg of ketamine had local anesthetic-sparing effects, it neither extended postoperative analgesia nor reduced the postoperative analgesic requirements. Compared to the patients who received bupivacaine alone, those who received bupivacaine and ketamine experienced an increased incidence of nausea, vomiting, sedation, dizziness, and “strange feelings.” In an interesting case report, the long-term intrathecal administration of the S(+)-ketamine enantiomer was found to be effective in a patient with severe neuropathic cancer pain refractory to conventional therapy. This study reported no adverse side effects and found low plasma concentrations of ketamine after the third week of treatment.

Epidurally administered ketamine has been found to be a clinically more viable treatment than intrathecal delivery, with a lower incidence of dysphoric and other adverse effects. In a randomized clinical study performed by Sethi and coworkers, low-dose ketamine was added to the epidural mixture patients undergoing major upper abdominal surgery. Ketamine added to bupivacaine in postoperative PCEA was found to provide better pain control with less opioid use. In addition, patients experienced significantly less nausea, vomiting, and pruritus. More recently, Amr demonstrated in a large randomized, double-blind, controlled trial that ketamine administered epidurally with steroids and a local anesthetic provided superior pain relief in 200 patients suffering from lumbar radiculopathy from disk herniation than steroids and a local anesthetic alone. Patients receiving ketamine by injections reported significantly decreased pain scores lasting up to 12 months postinjection, though six patients in the ketamine group experienced short-lasting delusions immediately after injection. In a study involving patients undergoing hepatic resection, the combination of epidural ketamine and morphine was found to provide superior pain relief compared to morphine alone. In elderly patients, the epidural ketamine dose was reduced by 33% (20 mg versus 30 mg). There were no reports of psychomimetic effects, neurologic findings, or any other complications in this study. In a study by Himmelseher and colleagues assessing the impact of adding S(+) ketamine to epidural anesthesia with ropivacaine in patients undergoing knee arthroplasty, the combination group experienced significantly longer pain relief than patients receiving the local anesthetic alone. A review of 13 randomized controlled studies involving 584 children reported that patients receiving 0.25 to 0.5 mg/kg ketamine via caudal administration extended the time needed for rescue medication by about 5 hours. Furthermore, these findings held true irrespective of the local anesthetic dose. One of the theoretic advantages of ketamine for chronic pain is that it is not associated with tolerance.

Not all studies have found epidural ketamine to be beneficial. Lauretti and colleagues found no benefit to adding epidural ketamine to clonidine in 56 patients undergoing orthopedic surgical procedures. The 24-hour postoperative pain scores, time to first rescue medication, and quality of analgesia were similar in the clonidine, ketamine, and ketamine-clonidine combination groups. In a double-blind study by Weir and associates, the authors found that adding ketamine to epidural bupivacaine for knee replacement surgery failed to prolong postoperative analgesia or reduce analgesic requirements, but it did result in significantly more side effects. The majority of the existing literature demonstrates that epidural ketamine in doses ranging from 0.5 to 1 mg/kg is well tolerated in patients of all age groups and is most effective when combined with opioids or local anesthetics.

The potential neurotoxicity of intrathecal ketamine remains a subject of controversy. In most countries, racemic formulations (50% S(+) and 50% R(−)-ketamine) are available either preservative-free or with preservatives, such as benzethonium chloride and chlorobutanol. Compared to the racemic mixture, the S(+) enantiomer has a four-fold greater affinity for NMDA receptor, and consequently possesses 2 to 3 times the analgesic potency. Karpinski and colleagues reported that a terminal cancer patient who received a 3-week intrathecal infusion of racemic ketamine was found to have subpial vacuolar myelopathy on autopsy; a similar finding was reported by Stotz and coworkers. On postmortem examination of a terminal cancer patient who received a 7-day trial of intrathecal ketamine, focal lymphocytic vasculitis close to the catheter injection site was noted. Subarachnoid S(+)-ketamine is a matter of much debate, as the results regarding its toxicity are contradictory. Although no human studies of preservative-free racemic or pure S(+)-ketamine have examined the histopathologic effects in humans, preservative-free racemic ketamine has been shown to be without apparent neurotoxic effects after repeated administration in pigs. However, S(+) ketamine without preservative injected into dogs resulted in significant histologic abnormalities, including gliosis, axonal edema, central chromatolysis, lymphocyte infiltration, and fibrous thickening of the dura mater. Similar findings have been reported in rat models and rabbits. At present, ketamine is not FDA approved for neuraxial use in the United States and is considered experimental or only for use in palliative care.

Cyclooxygenase Inhibitors ( Box 43.7 )

Research was conducted with animal models that implicate the cyclooxygenase isoenzymes, COX-1 and COX-2, as playing a role in the development and maintenance of spinal neuropathic pain. In the spinal cord, prostaglandin E2 (PGE ₂ ) acts presynaptically to increase the release of glutamate from primary afferent C fibers and postsynaptically to directly excite dorsal horn neurons by activation of nonselective cation currents. Both effects promote the development and maintenance of central sensitization and enhanced pain states. The intrathecal administration of nonsteroidal anti-inflammatory drugs (NSAIDs) prevents the development of hyperalgesia and inhibits the release of PGE ₂ . In an animal experiment using a peripheral nerve injury model, the nonselective COX inhibitor ketorolac provided significantly longer antinociception than the COX-2 selective NSAID, NS-398 (6 days versus 2 hours). These findings are consistent with recent experiments showing potent antinociceptive effects without neurotoxicity following intrathecal ketorolac administration. In a study by Parris and colleagues investigating intrathecal ketorolac and morphine in an animal model of neuropathic pain, the authors demonstrated that both drugs possess antinociceptive properties, with morphine being more potent than ketorolac for all outcome measures except cold allodynia, in which the effects of the two drugs were found to be similar.

Despite the promising evidence produced in animal studies, several well-designed studies have failed to demonstrate the efficacy of intrathecal ketorolac in clinical or experimental pain. Eisenach and associates tested the efficacy of intrathecal ketorolac in a randomized controlled, chronic pain study involving patients suffering primarily low back and lower extremity pain with a combination of somatic and neuropathic components. This study failed to demonstrate a difference between patients receiving ketorolac intrathecally compared with saline. In a second clinical study, Eisenach and collaborators found that 2 mg of intrathecal ketorolac combined with 15 mg bupivacaine failed to prolong postoperative analgesia when compared with normal saline. These results are disappointing and imply that intrathecal ketorolac are of limited utility for alleviating pain in humans. However, these studies included a small number of subjects, and more studies involving larger numbers of subjects are needed before conclusions can be made about the effectiveness of intrathecal ketorolac. Two studies examined the analgesic effects of intrathecal aspirin in patients with chronic refractory pain. In a large study conducted in 60 cancer patients with intractable pain, a single intrathecal dose of isobaric lysine acetylsalicylate (doses ranged from 120 to 720 mg) resulted in excellent pain relief in 78% of cases, with the duration of analgesia lasting from 3 weeks to 1 month on average. Neuraxial ketorolac is not FDA approved and is considered investigational or only for use in palliative care.

Cholinergic Agonists ( Box 43.8 )

Cholinergic receptors are divided into the G-protein-coupled (metabotropic) receptors (i.e., the muscarinic subtype) and ion channel (ionotropic) receptors (i.e., the nicotinic subtype). Pharmacologic molecular cloning studies have led to the classification of muscarinic acetylcholine receptors (mAChRs) in central and peripheral tissues into five distinct subtypes: M1, M2, M3, M4, and M5. Studies of radioligand binding and analysis of mRNA have demonstrated the existence of M1, M2, M3, and M4 receptors in the spinal cord. Neuronal nicotinic acetylcholine receptors (nAChRs) are pentameric ligand-gated ion channels, and molecular cloning has identified nine alpha and three beta subunits. These subunits assemble to form functional receptors in heteromeric combinations or as homopentamers. Receptor subunit composition underlies the differences in functional properties, and there is considerable variation in subunit expression throughout the spinal cord.

In animal models, the spinal administration of muscarinic cholinergic agonists results in antinociception, an effect that is reversed by muscarinic antagonists. In contrast, the intrathecal administration of nicotinic agonists results in a decrease in nociceptive threshold (hyperalgesia) or increase in spontaneous pain behaviors in most studies, but antinociception in others. The increase in pain-related behaviors after the intrathecal administration of nicotonic agonists may be the result of an associated increase in excitatory transmitter release. This hyperalgesia is reversed by the administration of intrathecal nicotinic or glutamate receptor antagonists. The diversity of nAChR subunits likely contributes to the seemingly paradoxical analgesic and hyperalgesic effects of nicotinic agonists. However, the predominant analgesic effects of neuraxial cholinergic drugs are thought to be mediated by muscarinic M1 and M3 receptors found in the dorsal root ganglia and superficial laminae of the dorsal horn, with a more modest contribution by nicotonic receptors. Cholinesterase inhibitors, including neostigmine and physostigmine, increase the amount of available acetylcholine by inhibiting its metabolism. These drugs have been found in numerous animal studies to produce antinociception, anti-allodynia, and antihyperalgesia after neuraxial administration.

Because of the lack of available selective muscarinic agonists and the paradoxical and unpredictable effects of neuraxially administered nicotinic agonists, human research has focused on the administration of cholinesterase inhibitors, primarily neostigmine. Despite promising data from animal studies, the administration of intrathecal neostigmine alone in humans has been somewhat disappointing. After preclinical toxicity screening, neostigmine was introduced into clinical trials for intrathecal administration. Spinal administration was found to produce analgesia to experimental pain stimuli in naïve volunteers and patients suffering from cancer and postoperative pain. Unfortunately, the pain relief was accompanied by severe and debilitating nausea. Thus, the use of intrathecal neostigmine as a sole analgesic is not currently recommended. However, the addition of neostigmine to intrathecal opioids and local anesthetics has been found to prolong and enhance analgesia, with only a modest increase in the incidence of nausea and vomiting. The most recent literature has focused on using neostigmine in combination with other neuraxial agents and has helped to better define the use of neostigmine in neuraxial analgesia. One finding is that neostigmine, when used in combination with an alpha-adrenergic agonist such as clonidine, might provide superior labor analgesia while reducing the risk of side effects when compared with either of these two drugs given alone.

Roelants and coworkers compared a single epidural dose of clonidine (150 µg) together with a single dose of neostigmine (750 µg) to three combinations of clonidine (75 µg) and neostigmine (250, 500, and 750 µg). The study found that only the combinations of 75 µg clonidine with 500 or 750 µg neostigmine provided visual analog scale (VAS) pain scores significantly lower than baseline, and that the effect was significantly longer than in the three other groups. Furthermore, the combination of these two drugs did not result in any increase in maternal adverse outcomes, such as nausea, hypotension, or sedation, or any neonatal adverse outcome. In another study, Van de Velde and colleagues found that a combination of clonidine 75 µg and neostigmine 500 µg, administered epidurally as part of a combined spinal-epidural anesthetic (CSE) technique with ropivacaine and sufentanil, did not result in maternal adverse effects. The authors found that the combination prolonged the initial analgesic effect of the spinal component of the CSE and decreased the requirement for local anesthesia. Furthermore, the investigators were able to demonstrate that this combination prolonged the initial analgesic effect of the spinal component of the CSE and provided a subsequent local anesthetic-sparing effect.

Neostigmine has also been coadministered for caudal analgesia in children with bupivacaine or ropivacaine, where it was reported to provide prolonged analgesia without any adverse effects. This benefit was demonstrated in a double-blind, randomized, prospective study by Karaaslan and coworkers, whereby 60 male patients between the ages of 5 months and 5 years undergoing genitourinary surgery were allocated randomly to one of three groups. One group received caudal 0.25% levobupivacaine (1 mL/kg) alone (note that levobupivacaine is no longer available in the United States). The next two groups of patients received neostigmine (2 and 4 µg/kg, respectively) together with levobupivacaine. Both groups receiving intrathecal neostigmine combined with levobupivacaine showed decreased pain scores postoperatively, a longer duration of analgesia, and lower analgesic consumption compared with the group administered levobupivacaine alone. There was no difference in analgesic efficacy between groups receiving different doses of neostigmine, nor were the adverse effects different among the three groups. The benefit of combining neostigmine with local anesthetics is not limited to levobupivacaine. Batra and colleagues demonstrated that combining neostigmine at a dose of 0.75 µg/kg with bupivacaine significantly extends spinal anesthesia duration, reduces postoperative pain scores, and decreases the need for rescue analgesia in infants undergoing lower abdominal and urogenital procedures. The study showed no benefit by increasing the dose to 1 µg/kg. Animal studies have shown that the coadministration of neostigmine does not decrease the neurotoxicity of lidocaine.

The combination of neostigmine with morphine via the epidural route has been shown to decrease the incidence of postoperative urine retention and prolong analgesia. The synergistic effect of combining neuraxial neostigmine with local anesthetics was confirmed in a randomized double-blind study by Kumar and colleagues, which assessed the addition of neostigmine, ketamine, and midazolam to bupivacaine in 80 children administered a single-shot caudal injection for inguinal hernia repair. The duration of complete analgesia was significantly longer in the neostigmine-bupivacaine group than in patients who received midazolam-bupivacaine, ketamine-bupivacaine, and bupivacaine alone. In addition to enhancing sensory blockade, combining neostigmine with an assortment of other intrathecal medications may also prolong muscle weakness and increase sedation. The most commonly encountered side effects of intrathecal neostigmine are nausea, vomiting, and, at doses exceeding 150 µg, sedation and leg weakness. At low doses, neostigmine is devoid of significant hemodynamic effects, but at higher doses (750 µg), increases in blood pressure, heart rate, respiratory rate, and anxiety may occur.

The epidural administration of neostigmine alone, or in combination with local anesthetics or opioids, has been found to effectively decrease postoperative pain. In contrast to intrathecally administered neostigmine, neostigmine administered via the epidural route enhances opioid and local anesthetic analgesia without increasing the incidence of nausea. In a prospective, randomized, double-blind study, Caliskan and collaborators demonstrated that patients who received 1 µg/kg of neostigmine in addition to 20 mL of bupivacaine experienced faster restoration of bowel sounds and a shortened duration of postoperative ileus after abdominal aortic surgery compared to patients receiving 20 mL of bupivacaine and an equal volume of normal saline. In the first of a two-phase randomized controlled study evaluating bolus (40 and 80 mcg) epidural neostigmine in women scheduled for elective cesarean section, Ross and colleagues found that epidural neostigmine boluses did not alter baseline fetal heart rate, induce contractions, or produce nausea. In the second, randomized phase comparing patient controlled epidural analgesia with either bupivacaine alone (1.25 mg/mL) or bupivacaine combined with neostigmine (4 mcg/mL), those who received the combination treatment had a 19% decrease in bupivacaine requirements. In parturients receiving treatment for longer than 4 hours, the decrease was 25%. Similar to a previous study, those patients who received combination did experience a slight increase in sedation.

Unfortunately, there are few studies evaluating the long-term use of neuraxial neostigmine for chronic pain. Lauretti and coworkers found that the epidural administration of a low dose of neostigmine (100 µg) in combination with morphine was associated with improved analgesia when compared to opioid treatment alone in terminal cancer patients followed for 20 days. This improvement was not associated with an increased incidence of adverse effects. Neostigmine is considered investigational or only for use in palliative care of terminal patients without further options and full consent of the patient.

Other Experimental Neuraxial Agents