JOURNAL HOME CME HOME THIS MONTH PAST ISSUES ETOC COLLECTIONS AUTHORS REVIEWERS EDITORIAL BOARD FEEDBACK RSS HELP
		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (35) Citing Articles via Google Scholar Google Scholar Articles by Walker, S. M. Articles by Carr, D. B. Search for Related Content PubMed PubMed Citation Articles by Walker, S. M. Articles by Carr, D. B. Related Collections Regional Anesthesia Pain

---

PAIN MEDICINE

Combination Spinal Analgesic Chemotherapy: A Systematic Review

Suellen M. Walker, MBBS MM(PM), MSc, FANZCA, FFPMANZCA^*, Leonidas C. Goudas, MD PhD, Michael J. Cousins, AM MD, FRCA, FANZCA, FFPMANZCA^*, and Daniel B. Carr, MD FABPM, FFPMANZCA

*Department of Anesthesia and Pain Management, Royal North Shore Hospital and University of Sydney, Sydney, Australia; and Departments of Anesthesiology and Medicine, Tufts University School of Medicine and Tufts New England Medical Center, Boston, Massachusetts

Address correspondence to D. Carr, MD, Department of Anesthesiology, Box 298, New England Medical Center, 750 Washington St., Boston, MA 02111. Address e-mail to daniel.carr{at}tufts.edu Reprints will not be available from the authors.

	Introduction

Top Introduction Combination Therapies: The Best... Commentary and Conclusion References

 
In the 25 years since the first application of spinal opioids for treating cancer pain, this mode of analgesia has grown to enjoy worldwide use (1,2). A natural outgrowth of local-anesthetic spinal anesthesia, "the major advantages of ‘selective’ blockade of pain by spinal opioids [lay] in the absence of sympathetic blockade and postural hypotension, potentially allowing easy ambulation of patients, and avoidance of cardiovascular collapse or convulsions—the major complications of local anesthetic blockade" (3). Perioperative epidural analgesia is increasingly recognized to influence surgical outcomes and the likelihood of developing chronic pain states, and technical refinements in delivery systems (4) allow chronic spinal drug delivery for management of previously refractory cancer pain (5–7). Such progress reflects the rapid maturation of the practice of spinal analgesia since the previous frequently cited (8) review of this topic (3).

During these same two decades, advances in preclinical pain research have led to recognition of the spinal cord as a key target for inhibition of acute nociception and preemption of "pain memory" (9). It is now clear that the customary distinction between acute and chronic pain is an oversimplification because key psychological and physiological responses traditionally associated with persistent pain (gene expression and neuronal sensitization and remodeling) rapidly follow acute injury (10). Persuasive evidence has emerged that persistent pain, regardless of its cause, constitutes a pathologic state per se in which spinal neuronal reorganization ("plasticity") exaggerates and perpetuates nociception and pain (11–13). Insight into spinal cord pathophysiology and pharmacology has spurred novel drug discovery and rekindled interest in spinal delivery of established drugs (14,15).

In an effort to gain better control over pain acutely and to maintain such control over the long term (16,17), anesthesiologists and others involved in the treatment of pain have advanced from single-drug spinal drug therapy to the coadministration of two and three drugs. Unfortunately, the standards of rigorous clinical evidence have almost wholly been ignored during this progress, and uncontrolled case series or case reports form by far the largest proportion of this literature (18). Considering that the number of possible combinations of different analgesics increases as a factorial function of the number of available choices, the optimum choices and relative doses of drugs to apply to different patient populations with different origins of pain require extensive clinical study (19). At the same time, knowledge of the neurotransmitters, membrane receptors, and intracellular mediators involved in dorsal horn nociceptive processing has evolved to reveal a remarkable diversity (Fig. 1). The growing trend of using spinal drug combinations that target multiple mechanisms of analgesia mirrors multidrug therapy in many medical disciplines and has been termed combination analgesic chemotherapy (11).

View larger version (71K): [in this window] [in a new window]   Figure 1. Possible arrangement of pre- and postsynaptic receptors on structures in the dorsal horn of the spinal cord, and potential sites of action of opioid and non-opioid spinal analgesics. Presynaptic release of the neurotransmitter glutamate (Glu) results in activation of the postsynaptic -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor, which controls a rapid-response sodium (Na+) channel. Substance P (SP) interacts with the neurokinin (NK-1) receptor and results in activation of second messengers. With prolonged activation, the N-methyl-D-aspartate (NMDA) receptor is primed, Glu activates the receptor, the magnesium (Mg2+) plug is removed, and the ion channel allows entry of Na+ and calcium (Ca2+) ions. The increase in intracellular Ca2+ then triggers a number of second-messenger cascades. Production of nitric oxide (NO) increases via the Ca2+/calmodulin-dependent enzyme NO synthase. NO may diffuse out of the neuron to have a retrograde action on primary afferents and also activates guanylyl cyclase, leading to increases in intracellular cyclic guanosine monophosphate (cGMP) and activation of cGMP-dependent protein kinases. Activation of the Ca2+-dependent protein kinase C isoform (PKC) leads to phosphorylation of the NMDA receptor, which reduces the Mg2+ block (dotted line II) relating to the development of opioid tolerance. The increase in intracellular Ca2+ also results in the induction of protooncogenes such as c-fos, with a presumed action on target genes of altering long-term responses of the cell to further stimuli. , µ, and = opioid receptors; GABA = -aminobutyric acid; 2 = 2 adrenoceptor; 5-HT = serotonin. Details of the potential analgesics are outlined in the text. NSAID = nonsteroidal antiinflammatory drug; SNX-111 and AM336 = omega conopeptides that block neuronal Ca2+ channels. DAMGO = [D-Ala2,N-Me-Phe4,Gly-ol5]-enkephalin; R-Pia = R-phenyl-isopropyl-adenosine; Neca = N-ethylcarboxamide-adenosine.

This article therefore

1. Describes the potential benefits that may follow the coadministration of spinal analgesics. These include
   
   1. a. Improvement in analgesic efficacy. Combinations of analgesic drugs may improve efficacy through either additive or synergistic interactions. Preclinical studies have identified a number of synergistic interactions with intrathecal (IT) coadministration of different compounds (22), but few clinical studies have been performed to rigorously characterize whether various combinations have additive or synergistic interactions (23). Particular types of pain may be better controlled with drug combinations. Delivering a local anesthetic with a spinal opioid improves the control of incident (i.e., movement-related) pain (24), whereas the addition of clonidine to an opioid enhances the control of neuropathic pain (25,26).
      
   2. b. Reduction in side effects. The combination of two drugs with the common desired end point of analgesia but with different side-effect profiles may enhance the therapeutic ratio of the therapy. The analysis of type and incidence of side effects must be carefully investigated to confirm clinical benefits with new combination therapies. In many postoperative studies, patients titrate analgesia to their own level of comfort, and therefore improved analgesic effect with the new drug may not be reflected as a reduction in visual analog pain score (VAS) but instead as a reduction in supplemental analgesic requirements. However, this is of limited clinical benefit unless there is an associated reduction in side effects achieved through the use of smaller doses of that drug (e.g., a reduction in nausea because of a reduced postoperative opioid requirement).
      
   3. c. Reduction in opioid dose escalation. Combination of a non-opioid analgesic with an opioid may reduce the development of tolerance indirectly by reducing the opioid requirement. Alternatively, analgesics such as N-methyl-D-aspartate (NMDA) receptor antagonists may directly affect the development of tolerance (27–29).
      
   
   
2. Summarizes preclinical evidence and, in a series of systematic reviews of randomized controlled trials, the best available clinical evidence on the effects of the coadministration of pairs of these drugs.
   
3. Identifies current deficits and future requirements for evidence-based practice.
   

	Combination Therapies: The Best Available Evidence—Search Strategy and Study Selection

Top Introduction Combination Therapies: The Best... Commentary and Conclusion References

 
To identify clinical trials on spinal delivery of analgesic drug combinations, we used the search strategy outlined in Table 1, supplemented by hand searches based on textbook chapters (11) and review articles (18,30,31).

The selection of articles that are included in this systematic review was based on the following inclusion criteria:

1. Population: patients with acute postoperative pain or chronic pain (cancer or chronic non-cancer-related pain).
   
2. Study design: randomized, double-blinded studies with factorial designs with at least one arm using the combination and one arm for any and each drug component, provided that each component was previously demonstrated to produce sufficient analgesia when used alone. As described previously, we also included trials that compared a control drug, adequate to produce analgesia, with the coadministration of the same drug plus a second drug not considered sufficient for clinical analgesia.
   
3. Intervention: all drugs singly or in combination were administered through the same route, epidural or IT, and at the same time point or interval with respect to the clinical condition studied, regardless of mode (continuous, intermittent, and patient-controlled, etc.).
   
4. Outcomes: pain relief or pain intensity measured at least once after drug administration. Reporting of adverse reactions was desirable but not mandatory for inclusion of an article.
   

Because of the small number of studies in each group and their heterogeneous design, particularly for the management of chronic pain, a mathematical meta-analysis could not be performed.

Opioids and Other Opioids
Spinal coadministration of two opioids with different pharmacokinetic or pharmacodynamic profiles may offer advantages. The combination of an opioid with a rapid onset of action and another with a slower onset but a longer duration of action may improve the quality of early analgesia and prolong its duration. For example, a single epidural injection of morphine plus sufentanil combines the short onset time produced by sufentanil with the long duration of analgesia attributable to morphine, thus providing prolonged analgesia after cesarean delivery (35). Animal studies show that combining opioids with different receptor selectivity has a powerful dosage-sparing effect (22), particularly for µ- plus -selective drugs (36,37). Delta-opioid receptors in the ventral medial medullary reticular formation may be involved in activation of bulbospinal noradrenergic pathways and a descending inhibitory pathway projecting through the dorsolateral funiculus and may thus reinforce other analgesic mechanisms (38,39). Not all opioids induce tolerance at the same rate, and cross-tolerance is incomplete. Tolerance develops more slowly with opioids that have high intrinsic activity than during chronic administration of lower-affinity opioids, such as morphine (11,40). IT D-Ala2-D-Leu5 enkephalin, a moderately selective -receptor agonist, is analgesic in patients tolerant to morphine (41).

Evidence from Randomized Controlled Trials
Three randomized, controlled trials investigated the combination of epidural fentanyl (42,43) or sufentanil (44) with epidural morphine after abdominal surgery in 289 patients (Table 2A–C). Early postoperative analgesia was improved by the use of the combination in all three studies, and this benefit for early analgesia was evident even when morphine was given 45 min before fentanyl (42,43). A reduced dose of morphine (3 mg) was used in the group coadministered sufentanil compared with the group given morphine alone (5 mg). Both the sufentanil-only group and the morphine/sufentanil-combination group had larger supplemental opioid requirements than the morphine-only group in the 24 h after surgery. Therefore, the combination did not allow a reduction in dosage, nor did the incidence of side effects differ between groups (44). In the other two studies, a number of dose combinations of morphine (2–4 mg) and fentanyl (50–100 µg) were investigated, with several conclusions about analgesic efficacy. One of these studies reported no augmentation of analgesia when a second drug was added to morphine doses >2 mg and fentanyl doses >50 µg (42). The other study reported improved late analgesia when a larger dose of 4 mg of morphine (alone or in combination with fentanyl) was compared with 2 mg of morphine (43). A reduction in supplemental opioid requirements in the first 24 h after surgery was reported by Tanaka et al. (42) when combination therapy was compared with either drug as a single therapy. However, in the other studies, combination therapy did not consistently reduce supplemental analgesic requirements when compared with single-drug administration, because it was significant compared with sufentanil alone but not morphine alone (44) or occurred only with a larger dose of combined drugs (42). The reduction in opioid requirements was of little clinical importance because no significant reduction in nausea was reported in the combination compared with single-drug groups (43,44). In the study by Tanaka et al. (42), an additive increase in side effects (vomiting and pruritus) was seen with the largest doses of epidural morphine and fentanyl.

Opioids and Local Anesthetics
Local anesthetics can provide intense sensory anesthesia, but spinal administration may result in motor weakness and postural hypotension because of sympathetic block. The addition of epidural fentanyl to a local anesthetic improves intraoperative surgical analgesia (meta-analysis of 18 controlled trials) (45), and continued infusion of epidural opioids and local anesthetics is common in the management of postoperative pain (46). Combinations of an IT opioid and local anesthetic are most often used in the acute setting to prolong the analgesia after single-shot IT anesthesia, and their use for cesarean delivery has recently been reviewed (47). In the chronic pain setting, clinicians report that opioids are often the first drug of choice for spinal administration, but local anesthetics may be added when pain is refractory to single-drug administration (17).

In animal studies, isobolographic analysis reveals synergistic antinociception with coadministration of morphine and lidocaine by the IT (48) or epidural route (49). Isobolographic analysis is the standard approach to characterization of the interaction of two drugs as additive, synergistic, or antagonistic. In this approach, a straight line—the "line of additivity"—is drawn between the dose of one drug necessary to achieve a specific effect (plotted on the y axis) and the dose of a second drug required to achieve the same effect (plotted on the x axis). If the same effect is produced by coadministration of the two drugs at dosages that, when plotted on the same graph, occur below the line of additivity, the two drugs are said to have a synergistic interaction (50). If doses of the two drugs coadministered to produce the same effect occur above the line of additivity, they are said to have an antagonistic interaction.

Dose-dependent development of tolerance was found after IT morphine infusions; this was not affected by the coadministration of lidocaine. This suggests that lidocaine does not directly influence the development of morphine tolerance. However, because an analgesic response was obtained with smaller doses of morphine in the combination group, and because no cross-tolerance was observed, there was an indirect benefit of lidocaine coadministration because smaller doses of morphine induced less tolerance than larger doses (48). In a prospective clinical study, van Dongen et al. (51) reported a less rapid opioid dose escalation between days 10 and 30 of therapy during IT coadministration of morphine and bupivacaine, compared with opioid alone.

Opioids and Local Anesthetics: Acute Pain
Evidence from Randomized Controlled Trials: Acute Pain.
These studies tested the hypothesis that a combination of a local anesthetic with an opioid coadministered spinally produces equal analgesia but with fewer side effects or produces improved analgesia without increased side effects when compared with either drug administered singly (Table 3A–C). As described previously, we did not include the large number of postoperative pain trials in which opioids were added to local anesthetics (i.e., A versus A + B). Many such trials have shown that the addition of an opioid reduces local anesthetic requirements or vice versa, but they do not allow conclusions as to the nature of their interaction (see the review by Wheatley et al.) (46).

Analgesic efficacy of the combination was better than that of local anesthetic alone but was not different from opioid alone in the study by Cooper et al. (53), whereas Torda et al. (55) found no difference among the effects of boluses of fentanyl alone, local anesthetic alone, or combination therapy. In two studies, overall patient satisfaction was assessed, again with variable results. Cooper at al. (53) reported similar levels of satisfaction after either combination therapy or opioid alone, but both were significantly better than local anesthetic alone. Kopacz et al. (54) reported similar satisfaction with the combination or local anesthetic alone and found both to be significantly better than with opioid alone. Inadequate analgesia was frequently reported after single-drug therapy [either in both the local anesthetic and opioid single-drug groups (54,56) or in only the local anesthetic group (53)].

The ability to evaluate a dosage-sparing effect of combination therapy depends on the methodology of the study. Similar analgesia was provided by a bolus of 50 mg of bupivacaine alone and by smaller bolus doses of bupivacaine (25 or 12.5 mg) together with fentanyl (55), and therefore the addition of the opioid allowed a reduction in the local anesthetic dose without loss of analgesia. Dosage sparing by combination therapy was assessed in two studies that used a spinal PCA regimen (53,57). In both studies, combination therapy reduced opioid and local anesthetic requirements by 50%–60%, as indicated by a reduction in the number of self-administered spinal boluses (53,57) and use of a less concentrated IT solution in the combination group (53). In a titrated epidural infusion regimen, both cumulative epidural medication requirements and supplemental IV PCA morphine requirements were decreased in the combination group compared with groups that received local anesthetic or opioid alone (56). By contrast, in a study design that used PCEA, there were no significant differences in the number of rescue boluses or cumulative volumes of PCEA solution among groups, despite patients in the single-drug groups reporting increased pain at some time points (54) and having an increased dropout rate.

The ability of combination therapy to reduce analgesic requirements compared with single-drug therapy is clinically useful if there is an associated reduction in side effects. There is limited evidence in these studies to support a reduction in local anesthetic-related side effects with combination therapy. In one study, the degree of hypotension after a thoracic epidural bolus was significantly less when a reduced dose of local anesthetic was used in combination with fentanyl (55). In three studies, the incidence of hypotension was not significantly different among groups (52,54,57). In the study by Kopacz et al. (54), there was more hypotension in all treatment groups, including patients receiving opioid alone. Hypotension may have been a residual effect of the large dose of intraoperative local anesthetic, because the timing of the hypotension was not reported. In the remaining studies, hypotension was not observed in any treatment group (53,56). A significant reduction in the degree of motor and sensory block was reported by Cooper et al. (53) when the dose of local anesthetic was reduced by combination therapy, but many patients in both the local anesthetic and the combination groups had difficulty mobilizing. In other studies, no significant difference in the degree of motor block was seen in the local anesthetic or combination groups (52,56,57). One study reported a reduction in sedation with combination therapy compared with sufentanil alone (57). In these studies, there is no evidence for the ability of combination therapy to reduce the side effects of nausea, vomiting, or pruritus compared with opioid alone (52,54), despite reductions in opioid requirements with combination therapy (53,56,57). The overall incidence of a particular side effect is often reported (e.g., the number of patients who vomited) without an indication of the frequency or severity of the adverse event. This failure to discern differences in side effects associated with dosage sparing may in part relate to the relatively small numbers of patients in each treatment arm (10–24 patients) and consequent underpowering. Indeed, power analysis was reported in only two studies. In the study by Kopacz et al. (54), the sample size was based on estimates of the primary efficacy end point (time to first rescue analgesia). The study by Torda et al. (55) included calculations based on changes in pain scores in addition to changes in one selected side effect (hypotension).

Summary of the Evidence.
Four studies support improved analgesic efficacy with the combination of a local anesthetic and an opioid compared with either drug administered alone. However, in two other studies, no difference in analgesic efficacy was found between the combination and the opioid alone (53) or the combination versus either single drug (55). Most studies indicate that combination therapy reduces dose requirements for either the local anesthetic or the opioid when they are administered as single drugs. This dose reduction is associated with reduced local anesthetic-related side effects (hypotension and motor block) but little (sedation) or no (vomiting and pruritus) reduction in opioid-related adverse effects.

Opioids and Local Anesthetics: Chronic Pain
Spinal coadministration of a local anesthetic and an opioid has been used extensively for the management of chronic pain. Prolonged use of IT combinations of morphine and bupivacaine has been reported in case series of patients with cancer pain (58), with two series reporting adequate pain control until death in 105 patients (59,60). In 51 patients with cancer pain, 17 proceeded from a morphine-only to a morphine/bupivacaine spinal-infusion mixture. Pain intensity subsequently improved in 10 patients, with only moderate improvement in 4 patients, whereas 11 patients required continuation of oral morphine supplementation (61). In patients with noncancer pain, IT opioid provided satisfactory pain relief in 88% (285 of 323 patients in 7 studies), whereas IT opioid plus bupivacaine provided satisfactory pain relief in 93% (96 of 103 patients from 2 studies) (62). In these case series, bupivacaine was added when pain control was inadequate with the opioid alone. Interpretation of these data is hampered by lack of randomization, variable inclusion criteria (particularly type of pain), and variable definitions of satisfactory pain relief. Two prospective studies have shown improvement in analgesia with bupivacaine and morphine combinations compared with opioid alone, although there was neither blinding nor randomization in one study (63) and there was incomplete blinding in the other (51). In both studies, pain intensity at the time of entry varied among patients, who were enrolled for IT therapy when pain was "refractory" to the use of analgesic drugs by other routes (on the basis of the World Health Organization analgesic ladder) or neurolytic techniques. As required in the clinical context, infusions were titrated to effect in individual patients, but the resultant variation in dosing leads to difficulty in the evaluation of the efficacy and side-effect profile of combination versus single-drug therapy.

Reductions in blood pressure due to sympathetic blockade are often seen in the first 24 h of IT bupivacaine infusion, but after this initial stabilization, postural hypotension is rarely significant. Side effects of bowel or bladder dysfunction and motor weakness have not been observed at IT bupivacaine doses of <30 mg/d (61), but motor weakness occurred with IT doses of >45 mg/d (60). These side effects are more likely to be tolerated by patients who are bedridden with terminal disease, and further studies in ambulatory patients are required. A reduction in opioid-related side effects was reported in one study after the initiation of combination therapy (64), but in most series the numbers studied are too few to identify any difference in the incidence of side effects with single or combination therapy.

Evidence from Randomized Controlled Trials: Chronic Pain
Only one trial satisfied the selection criteria applied in this systematic review for the use of opioid/local anesthetic combinations in the chronic pain setting (Table 3A–C). van Dongen et al. (51) reported a diminished progression of IT morphine dose during the coadministration of morphine and bupivacaine when compared with IT morphine alone. However, because infusions were titrated to effect in individual patients who had progressive disease, no assessment of analgesic efficacy with combination therapy was possible. Similarly, many patients had preexisting adverse effects related to analgesic regimens, the underlying disease, or both, rendering direct comparison of the incidence of side effects impossible.

Opioids and Clonidine
Much basic research confirms the antinociceptive properties and the mechanisms involved in the production of analgesia after spinal administration of ₂-adrenergic drugs (65,66). Clonidine’s analgesic activity is mediated through pre- and postsynaptic ₂ receptors localized in the superficial layers of the spinal dorsal horn (22). In healthy volunteers, the reduction of pain intensity after epidural clonidine correlates with its concentrations in the cerebrospinal fluid, but not in serum (65). Much smaller bolus doses of clonidine are needed through the IT route to produce potent and long-lasting analgesia than via the epidural or systemic routes (67). Preclinical and clinical evidence suggests that clonidine’s actions at the spinal cord simulate the release of spinal norepinephrine that is also observed after systemic administration of opioids (68) or during painful labor (69). These and other data indicate that opioids and clonidine act on different receptor systems and distinct but overlapping neural pathways at spinal and supraspinal sites to produce potent antinociception. Investigations of this interaction in animal models support the hypothesis that clonidine, the most widely used ₂-adrenergic agonist, produces at least an additive and in some cases a synergistic antinociceptive effect when coadministered with opioids at the spinal cord (67). The prevalence of adverse effects of opioids has led clinical investigators to test combinations of opioids with clonidine for the management of postoperative, labor, or chronic pain.

Evidence from Randomized Controlled Trials
Seven randomized, controlled trials satisfied our selection criteria (Table 4A–D). These trials investigated the analgesic effect produced by combinations of opioids (i.e., morphine, fentanyl, and sufentanil) with clonidine compared with the analgesic effect produced by the opioids or clonidine administered singly through the epidural or IT route for acute postoperative or chronic (cancer or spinal cord injury) pain. In aggregate, 461 patients enrolled in 7 clinical trials were randomized to receive an opioid with clonidine, the same opioid alone, or clonidine alone. In five of the seven trials, investigators used the epidural route (25,70–73), whereas in the remaining two they used IT injections (26,74). Clonidine was combined with morphine in five studies, sufentanil in one study, and fentanyl in one study.

The epidural route was used in four trials (70–73) and the IT route in one trial (74). The same doses of clonidine and morphine used singly were combined in the mixture in two studies (70,73), but larger doses of single drugs than those used in the mixture were used in one study (72). Another study tested three doses of each drug administered singly (71), whereas in the remaining study, combination doses of the two drugs in a fixed ratio of doses were given (74).

Morphine and Clonidine.
Carabine et al. (70) compared bolus epidural injections of clonidine (150 µg) followed by continuous epidural infusion of clonidine 25 or 50 µg/h, with a bolus injection of morphine (1 mg) followed by epidural infusion of morphine (0.1 mg/mL) and a bolus injection of a mixture of clonidine (150 µg) and morphine (1 mg) followed by continuous epidural infusion of morphine 0.1 mg/h. At both 30 and 60 min after the injections, all three groups had significantly lower values for pain intensity compared with the morphine group. Supplemental IV PCA morphine requirements in the combination group were significantly less than in the morphine and clonidine 25-µg groups. IV PCA morphine requirements in the clonidine 50-µg group were also less than in the epidural morphine group. However, the combination and the clonidine 50-µg groups did not differ in IV PCA morphine requirements for supplemental analgesia. Hypotension was significantly more pronounced in the combination group compared with the other groups from 5 until 20 min after injection. At 18 and 24 h after surgery, arterial blood pressure was significantly less in both clonidine groups than in the morphine and combination groups. No differences in other side effects were demonstrated. Collectively, the observations of this trial suggest that there is no demonstrable benefit in the use of the mixture as compared with the 50 µg of clonidine. The incidence of hypotension in the combination group was more pronounced than in the clonidine 50-µg group, whereas these two groups did not differ in sedation. These results must be interpreted cautiously because in fact all groups received a mixture of clonidine and morphine (the two clonidine groups received supplemental IV PCA morphine). Rockemann et al. (72) showed that the combination of a minimally effective epidural morphine dose (2 mg) with a marginally effective clonidine dose (280 µg, calculated according to patient weight) produced analgesia that was not significantly different from that produced by morphine alone (3.35 mg, calculated according to patient weight). In this study there were no differences in side effects between study arms. It is noteworthy that the investigators rightly excluded 6 of 15 patients in the morphine group from data analysis because of requests by these patients for supplemental analgesia. The study demonstrates that the combination of clonidine and morphine is better compared with morphine alone only because of the faster onset of pain relief. van Essen et al. (73) compared clonidine (70 µg), morphine (3 mg), and their combination given as bolus epidural injections 60 min after surgery in 28 patients for postoperative pain control. The authors found no difference in pain intensity (verbal analog pain score) in any of the three treatment groups. Statistically significant reductions in blood pressure were observed in the morphine-with-clonidine group but were considered of no clinical importance by the authors. No significant differences were observed in other side effects (urinary retention, nausea, vomiting, and pruritus) after the combination as compared with morphine alone. No supplemental opioid was administered to any of the patients in this study, although they could have requested it if they wished.

Fentanyl and Clonidine.
Eisenach et al. (71) found a slight and insignificant benefit of the combination of epidural clonidine and fentanyl for obstetric pain relief. This was demonstrated as a reduction in the 50% effective dose (ED₅₀) for pain relief of the mixture compared with a theoretical additive ED₅₀ calculated from data acquired from the single-drug groups, suggesting an additive interaction between the study drugs. No differences in side effects were demonstrated. The pain outcome was assessed at 10 min after epidural injections. Supplemental opioids were used later during the study.

Sufentanil and Clonidine.
Gautier et al. (74) compared the effects of a single IT bolus of sufentanil (2.5 or 5 µg) and clonidine (15 or 30 µg) with those of their combination (all four possible dose combinations) administered in women in active labor. The combination of 30 µg of clonidine with 2.5 or 5 µg of sufentanil produced a significantly longer duration of analgesia as compared with 5 µg of sufentanil alone. The dose of 15 µg of clonidine combined with 2.5 or 5 µg of sufentanil produced analgesia of similar duration to that of 5 µg of sufentanil. The mixtures of clonidine and sufentanil did not result in a significant reduction in the incidence of side effects as compared with sufentanil alone. Collectively, there was a significant improvement in the analgesic outcome with use of the mixture of sufentanil and clonidine as compared with sufentanil alone, with no significant difference in side effects.

Summary of the Evidence.
These randomized trials provide the best available clinical evidence concerning the combination of clonidine and morphine, fentanyl, or sufentanil at the spinal cord for acute pain. Weaknesses of these trials may be the relatively small number of patients enrolled (range, 28–100) and the use of a supplemental opioid in two of the four trials (70,72), with resulting "impure" treatment groups that potentially influence the results. The improved pain outcomes recorded in most cases are at single time points with lower pain scores for the combination as compared with the opioid alone, a reduction in supplemental opioid requirement after the combination, or an increase in the duration of analgesia. None of the studies demonstrated a reduction in the incidence or severity of side effects (e.g., hypotension or sedation) after the combination as compared with the single drug. These trials provide at best weak and scattered evidence that administering clonidine with an opioid spinally is more effective than either clonidine or the opioid used singly for acute pain management. One may attribute these results to the marginal doses of epidural clonidine used in these trials (range, 70 to 300 µg) and to the small number of patients enrolled per study group. It is also interesting that only one of the studies discussed here used the IT route (72), which is presumed to be in the proximity of the site of interaction between opioids and ₂-adrenergic agonists.

Opioids and Clonidine: Chronic Pain
We identified two randomized placebo-controlled trials investigating the combination of epidural or IT morphine with clonidine—one for the management of chronic cancer pain (25) and the other for pain after spinal cord injury (26). The Epidural Study Group (25) compared in a blinded fashion the analgesic efficacy of epidural morphine (0.05 mg/kg) plus clonidine (3 µg/kg) with that of epidural morphine (0.05 mg/kg) plus saline in 85 patients with intractable cancer pain despite large doses of opioids. Success was defined as a reduction of pain intensity (on a 0–10 VAS) or a reduction of morphine use, with the alternative variable either decreasing or remaining constant. According to this definition, the success rate was significantly increased in the combination group as compared with the morphine-only group. This difference was more pronounced in patients with neuropathic pain. Hypotension was increased in the morphine-plus-clonidine group, whereas nausea was slightly but significantly greater in the morphine-only group. Measurements of quality-of-life indices at baseline and at the end of the study did not differ between groups. These data suggest that the addition of epidural clonidine to morphine for cancer pain management is beneficial, particularly in those patients with a significant neuropathic component. Siddall et al. (26), in a study that used a design especially suitable for combination analgesic drug trials, compared the efficacy of IT administration of saline, morphine (0.2–1 mg), clonidine (50–100 µg), and the combination of clonidine and morphine (see Fig. 2). The study consisted of two phases. Each patient received saline, clonidine, and morphine in a random sequence, and one dose per day of each drug was titrated over 3 days toward a positive response (defined as a >50% reduction from baseline pain score) or the occurrence of side effects. The starting doses of IT morphine and clonidine were calculated on the basis of previous use of these drugs. If the patient was regularly taking opioids, then the dose was calculated as 1:100 of the daily IM dose or 1:200 of the daily oral dose. Titration of clonidine and morphine was performed as follows: if there was inadequate pain relief without substantial side effects (sedation or effect on respiratory function), the subject received a 50% larger dose of the same drug on the second day and double the initial dose on the third day. During the second phase of the study, each patient received a combination consisting of 50% of the final dose of morphine combined with 50% of the final dose of clonidine. The authors compared the proportion of those patients who had a positive response at any time during the assessment. Of the 15 patients tested, 5 responded positively to saline, 3 to the largest dose of clonidine alone, 4 to the largest dose of morphine alone, and 7 to the combination of half the largest dose of clonidine plus half the largest dose of morphine. These data suggest that morphine and clonidine are a worthwhile combination but do not permit a distinction between additivity and synergy in their analgesic effect.

View larger version (32K): [in this window] [in a new window]   Figure 2. Pain relief (expressed as a percentage of the pretest baseline numerical pain rating score) after the administration of saline (placebo), morphine, clonidine, and a mixture of morphine and clonidine in the group of subjects with spinal cord injury. Lines represent the change in individual subjects, and the heavy lines represent the means for the group, with the vertical line indicating SEM.

The ability of combination therapy to reduce the incidence of side effects varies. Reductions in opioid-related side effects with combination therapy have been reported in a single-bolus epidural study, although the comparison was limited by the much larger dose of morphine being used as a single drug (2 mg of morphine given alone, versus 0.5 mg of morphine and 10 mg of ketamine in combination) (80).

In animal models, NMDA antagonists have significant effects on the development of opioid tolerance. Chronic spinal administration of MK801, an NMDA receptor antagonist, attenuates the development of tolerance to IT morphine in a dose-dependent manner (81). IT co-infused ketamine attenuates morphine tolerance in animal models of both somatic and visceral antinociception (82,83). These data suggest that a combination of an NMDA antagonist and opioid may have advantages for long-term infusions in clinical practice. Improved analgesia and reduced opioid requirements have been reported with addition of S(+)-ketamine to an IT infusion of morphine in a patient with chronic back and leg pain (84). Further trials are required to establish the efficacy, side-effect profile, and safety of spinally administered ketamine.

Evidence from Randomized Controlled Trials
Two postoperative randomized, controlled trials have investigated the effect of the addition of ketamine in an infusion plus bolus PCEA regimen of morphine (85) or morphine, bupivacaine, and epinephrine (86) (Table 5A–C). The relative concentrations of drugs in the infusions were kept constant, but the size of the bolus dose and the rate of background infusion were adjusted according to the VAS for pain intensity. The interval and duration of pain assessment varied markedly between the two studies. When VAS was assessed daily, the addition of ketamine was reported to significantly reduce VAS at rest on Day 1 and 2 and VAS with coughing for 3 days after surgery (86). When assessed more frequently in the first 24 h (at 30 min and 3, 6, 12, 18, and 24 h), pain VAS was significantly reduced at 30 min and 3 h only (85). Both studies reported a reduction in PCEA opioid consumption by the addition of ketamine. However, the incidence of nausea and pruritus was not significantly reduced by combination therapy in either study, although a reduction in the incidence of vomiting in the combination group was reported in one study (85).

Summary of the Evidence
The addition of ketamine to an opioid-based PCEA regimen reduces overall opioid requirements, but without clear clinical benefit in terms of a reduction in associated adverse effects. Current evidence supports the analge-sic efficacy of single-bolus spinal ketamine added to morphine in patients with cancer pain, but there are no controlled data for chronic therapy with this combination.

Neostigmine in Combination Therapy
IT neostigmine, an acetylcholinesterase inhibitor, produces dose-related analgesia through a muscarinic action (89). When administered as a single drug to volunteers, IT neostigmine resulted in a large number of side effects, including nausea, vomiting, and leg weakness, at doses >50 µg, and it increased blood pressure and heart rate at doses >200 µg (90). Therefore, interest is now focused on the potential benefits of small doses of neostigmine coadministered with other IT analgesics.

Neostigmine and Local Anesthetic: Evidence from Randomized Controlled Trials
The effect of the addition of 25–100 µg of neostigmine to an IT local anesthetic has been studied in the perioperative period in patients undergoing lower limb orthopedic surgery (91) and vaginal hysterectomy (92) (Table 6A–C). Minor reductions in pain VAS were seen at some time points in the 24 h after surgery, and postoperative analgesic requirements were reduced. However, a dose-dependent increase in the incidence of nausea occurred in patients receiving neostigmine. In a similar study, prophylactic antiemetics (0.5 mg of IV droperidol, 10 mg of IV metoclopramide, or a 2–4 mg · kg^-1 · h^-1 infusion if IV propofol) did not decrease the nausea and vomiting experienced by patients receiving 100 µg of neostigmine added to the IT local anesthetic (92). Because all groups received the same dose of local anesthetic, it is unclear from both studies whether neostigmine has dose-sparing effects when administered with local anesthetic. There were no differences in hemodynamic variables with combination therapy (93).

The efficacy of neostigmine in a nerve-injury model suggests a potential role in chronic pain management. In a controlled trial, the addition of a bolus of epidural neostigmine (100 µg) to epidural morphine increased the duration of analgesia in patients with terminal cancer (87). Current case reports are also limited to effects after single-bolus administration. In two patients with lower limb ischemic pain, single doses of hyperbaric neostigmine 50 µg produced analgesia lasting longer than 6 h but were associated with nausea and vomiting (97). Two patients with metastatic abdominal cancer achieved relief of pain for approximately 20 h after single IT injections of neostigmine (100 and 200 µg) (98). Controlled trials of chronic IT neostigmine administration have not yet been conducted, and the effect of coadministration on the development of opioid tolerance has not been evaluated.

Evidence from Randomized Controlled Trials
An increased duration of analgesia and reduction in postoperative analgesic requirement has been reported with combinations of neostigmine and opioid when compared with either drug given singly (99–101) (Table 7A–C). Initial phases of a study conducted in women during labor by Nelson et al. (96) demonstrated a 25% reduction in ED₅₀ for sufentanil when combined with neostigmine. The subsequent randomized phase of the study found similar analgesic efficacy when sufentanil 9 µg was compared with a smaller dose of sufentanil (6 µg) given in combination with neostigmine, but there was no reduction in side effects. Therefore, although equivalent analgesia has been achieved with smaller doses of opioid when combined with neostigmine, no study has shown clinical advantages of the combination therapy. The number of patients in each treatment group tends to be small in these studies, and evaluation of the relative incidence of side effects is difficult. The incidence of nausea and vomiting is often increased when neostigmine is added, and only one study reported a significant reduction in the incidence of an opioid-related side effect, i.e., pruritus, when IT morphine 50 µg plus neostigmine 12.5 µg was compared with morphine 100 µg (101).

Because sympathetic nervous system activity is reduced by spinally administered clonidine and is increased by neostigmine, a potential advantage of combination therapy is the inhibition of clonidine-induced hypotension by neostigmine (105). In a volunteer study, IT neostigmine alone (50–200 µg) did not affect blood pressure, but the coadministration of neostigmine and clonidine prevented the dose-dependent hypotension induced by epidural clonidine infusion (104). The incidence of other side effects associated with single-drug use (nausea and leg weakness with neostigmine, and sedation with clonidine) was not equivalent during combination therapy. During labor, the incidence of hypotension was less when patients received neostigmine in combination with clonidine, but the number studied was too small to detect a statistically significant difference (106).

Evidence from Randomized Controlled Trials.
The addition of IT neostigmine 50 µg and clonidine 150 µg to bupivacaine spinal anesthesia for cesarean delivery resulted in prolonged postsurgical analgesia when the combination was compared with neostigmine or clonidine alone plus bupivacaine (Table 8A–C). However, the dose of drugs was not reduced in the combination group (neostigmine plus clonidine plus bupivacaine), and an increased duration of motor block and an increased incidence of nausea and vomiting were also seen in the combination group (107). Although prolongation of analgesia was maintained with smaller doses of IT clonidine 30 µg and neostigmine 10 µg added to bupivacaine and fentanyl 25 µg during labor, nausea was more common in all patients who received neostigmine (106).

Midazolam in Combination Therapy
Animal studies provide evidence of analgesic interactions between -aminobutyric acid (GABA)-A agonists and morphine (108) after IT coadministration. Midazolam binds to the benzodiazepine site of the GABA-A receptor complex and increases the amplitude and duration of the GABA-induced synaptic current (109). When administered at a spinal level, midazolam displays additive or synergistic interactions with opioids (110,111), whereas midazolam inhibition of opioid antinociception has been reported at a supraspinal level (112). Synergistic analgesia has also been shown between spinally administered midazolam and glutamate receptor antagonists acting at the NMDA receptor (AP-5) or the -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor. Analgesia was achieved at smaller doses when the drugs were combined, and this was associated with a reduction in the untoward behavioral changes and motor disturbances seen at doses required for single-drug analgesia (113).

An early clinical trial reported prolonged postoperative analgesia after IT midazolam in a small number of patients (114). In chronic low back pain, a single dose of IT midazolam was reported to provide prolonged analgesia comparable to that achieved with epidural steroids (115). A subsequent randomized placebo-controlled trial (although reported only in abstract form) found no analgesic effect and no difference from placebo when IT midazolam 2 mg was administered to patients with chronic mechanical low back pain.¹ Infusion of IT midazolam and clonidine improved analgesia in four patients with chronic benign pain (116), and isolated case reports of severe cancer pain have reported improvements with the addition of IT midazolam to IT opioid, local anesthetic, or both (117,118).

Investigation of the neurotoxicity of spinally administered midazolam has yielded conflicting results (119–121). In relation to combination therapy, repeated administration of fentanyl 0.005% and midazolam 0.1% was not associated with a more prolonged neur