alpha2-Adrenergic agonists produce analgesia primarily by a spinal action and hypotension and bradycardia by actions at several sites. Clonidine is approved for epidural use in the treatment of neuropathic pain, but its wider application is limited by hemodynamic side effects. This study determined the antinociceptive and hemodynamic effects of a novel alpha2-adrenergic agonist, MPV-2426, in sheep.
Forty sheep of mixed Western breeds with indwelling catheters were studied. In separate studies, antinociception to a mechanical stimulus, hemodynamic effects, arterial blood gas tensions, cerebrospinal fluid pharmacokinetics, and spinal cord blood flow was determined after epidural, intrathecal, and intravenous injection of MPV-2426.
MPV-2426 produced antinociception with greater potency intrathecally (ED50 = 49 microg) than epidurally (ED50 = 202 microg), whereas intravenous administration had no effect. Intrathecal injection, in doses up to three times the ED95, failed to decrease systemic or central arterial blood pressures or heart rate, whereas larger doses, regardless of route, increased systemic arterial pressure. Bioavailability in cerebrospinal fluid was 7% after epidural administration and 0.17% after intravenous administration. Intrathecal MPV-2426, in an ED95 dose and three times this dose, produced a dose-independent reduction in thoracic and lumbar spinal cord blood flow.
MPV-2426 shares many characteristics of other alpha2-adrenergic agonists examined in sheep, but differs from clonidine and dexmedetomidine by lack of antinociception and minimal reduction in oxygen partial pressure after large intravenous and epidural injections. No hemodynamic depression was observed after intrathecal injection at antinociceptive doses. These results suggest this compound may be an effective spinal analgesic in humans with less hypotension than clonidine, although its relative potency to cause sedation was not tested in this study.
α2-ADRENERGIC agonists produce antinociception in animals by a primary spinal action 1and produce analgesia in humans with acute and chronic pain. 2Although clonidine is frequently administered as part of regional anesthesia in Europe, the European formulation of clonidine does not carry an indication for analgesia or anesthesia. In contrast, studies in the United States on patients with intractable cancer pain 3led to registration of clonidine for epidural administration in combination with morphine for the treatment of refractory neuropathic pain. As a result, in large part, of the restricted clinical development of the drug, the formulation in the United States carries a strong warning against its wider use in perioperative or obstetric settings.
In addition to medicolegal concerns from this warning, limited epidural or intrathecal clonidine administration is a result of its side effects, which are primarily hypotension and bradycardia. α2-Adrenergic agonists produce hypotension by actions in the spinal cord, such that hypotension is more common or severe after spinal injection at thoracic dermatomes, where sympathetic preganglionic neurons reside. 4Additionally, systemic absorption of α2-adrenergic agonists can lead to hypotension and bradycardia by distribution to the brainstem, where these agents reduce sympathetic nervous system activity. At high circulating concentrations, α2-adrenergic agonists produce vasoconstriction, leading to a transient increase in blood pressure after bolus administration in humans. Severe hypertension, bradycardia, and reduced cardiac output also occur in dogs after administration of large doses. 5
MPV-2426 HCl [2,3-dihydro-3-(1H-imidazol-4-ylmethyl)-1H-indan-5-ol] is a novel α2-adrenergic agonist that may obviate some of the concerns of clonidine. MPV-2426 is a full agonist at α2-adrenergic receptors, whereas clonidine is a partial agonist. MPV-2426 binds with low nanomolar affinity (K ivalues of 1–2.1 nM) to all three subtypes of the α2-adrenergic receptor. In preliminary studies in rats, this agent produces potent antinociception after intrathecal administration with less sedation than other α2-adrenergic agents (sedation/analgesia ratio of 43 for MPV-2426 compared with 0.8 for clonidine). The purpose of this study was to characterize more fully the antinociceptive and hemodynamic effects of MPV-2426 in sheep. Although sheep are less sensitive than humans to the hypotensive actions of α2-adrenergic agonists, 6they do respond to thoracic intrathecal injection of these agents with reproducible blood pressure reductions. 7Pharmacokinetics were determined in cerebrospinal fluid (CSF) for comparison with previous studies of clonidine and dexmedetomidine in sheep. 8,9We also determined the effect of MPV-2426 on arterial blood gas tensions, because α2-adrenergic agonists produce hypoxemia in sheep. 10Finally, the effects of intrathecal injection of MPV-2426 on spinal cord blood flow were determined as one component of its preclinical safety evaluation.
After obtaining approval from the Animal Care and Use Committee, 40 sheep of mixed Western breeds were studied. All animals were in good condition on delivery to the laboratory.
After a 24-h fast, 14 adult ewes were anesthetized with 10 mg/kg intramuscular ketamine followed by endotracheal intubation and controlled ventilation with 1–2% halothane in oxygen. Polyvinyl catheters were inserted in a femoral artery and vein under direct vision and advanced 12 cm proximally. The animal was turned prone, and the dura mater at the C3–C4 interspace was exposed by surgical dissection, occasionally assisted with a small laminotomy. A 20-gauge polyethylene catheter was inserted through a small nick in the dura 6 cm caudad so that its tip lay at the lower cervical dermatomes, and a similar catheter was inserted the same distance caudad in the epidural space. Catheters were sutured in place, incisions were closed, and anesthesia was discontinued. All animals were standing, eating, and drinking normally within 4 h of completion of surgery. Prophylactic analgesics were not administered, and in no case was there behavioral evidence of pain or discomfort that would have resulted in administration of analgesics according to protocol.
Animals were trained before surgery with a mechanical device that applies pressure on the forelimb just distal to the knee. The behavioral end point was lifting of the forelimb, at which point the pressure was recorded and released. A maximum of 20 N force was applied to avoid tissue damage. This method has been used previously to examine the antinociceptive effects of α2-adrenergic and opioid agonists in sheep. 11–13
Each animal was studied three to six times, with the first experiment occurring at least 5 days after surgery, and experiments were separated by 2–5 days. On the day of the experiment, withdrawal threshold was determined before and for specified intervals for 2 h after drug injection. Based on pilot experiments, doses studied (dose expressed as hydrochloride throughout the study) were as follows: intrathecal (20, 50, and 100 μg), epidural (30, 100, and 300 μg), and intravenous (2, 20, and 80 μg). Drug was diluted with preservative-free saline and administered in a volume of 2 ml (intrathecal) or 5 ml (epidural and intravenous). Dose of drug was randomized, and the investigator was blinded to dose but not route of injection.
Hemodynamic Effects and CSF Pharmacokinetics
Fifteen adult ewes were anesthetized as previously described. Catheters were inserted in both femoral arteries and a femoral vein and advanced 12 cm proximally. An introducer was inserted in the right common jugular vein and sutured in place for subsequent insertion of a pulmonary artery catheter. The animal was turned prone, and a midlevel lumbar laminotomy was performed. Polyethylene catheters were inserted in the epidural space and, through a small nick in the dura, in the intrathecal space and advanced 15 cm such that the tip of the catheter lay at the midthoracic dermatomes. Catheter tip location was not confirmed in this study because animals were not killed at the end of the experiment. Catheters were sutured in place, incisions were closed, and anesthesia was discontinued.
Each animal was studied three to five times, with the first experiment occurring at least 5 days after surgery, and experiments were separated by 2–5 days. On the day of the first experiment, a pulmonary artery catheter was inserted through the introducer into the pulmonary artery under pressure waveform guidance. Correct location of the catheter was confirmed on subsequent experimental days. Femoral and pulmonary artery catheters were connected to transducers, and mean pressures and heart rate were measured via a computer-based data acquisition system, with average values calculated every minute. After 30 min of baseline recordings, animals received a single injection of MPV-2426 as follows: intrathecal (50, 100, and 300 μg), epidural (200, 700, and 1,400 μg), and intravenous (50, 100, and 300 μg). With the exception of the intravenous dose (which lacked antinociceptive efficacy), these doses represent approximately an ED50, ED95, and two (epidural) or three (intrathecal) times the ED95determined in analgesia experiments. Dose of drug was randomized, and the investigator was blinded to dose but not route of injection. Pressures and heart rate were recorded at specified intervals for 4 h after injection. Cardiac output was determined in duplicate via injection of iced saline and a thermodilution technique, and arterial blood was sampled for blood gas tensions and p H before and at 5, 10, 20, 30, 45, 60, 90, 120, 180, and 240 min after drug injection.
Cerebrospinal fluid was sampled from the in situ catheter (0.2 ml to clear dead space of catheter and 0.5 ml for analysis) at specified intervals after each intrathecal injection and after the largest doses administered epidurally and intrathecally, and was analyzed for MPV-2426 by the bioanalytical laboratory at Orion Corporation.
MPV-2426 concentrations were determined by liquid chromatography–tandem mass spectrometry. The limit of quantitation for sheep plasma samples was 5 pg/ml when 1 ml plasma was used. The interassay coefficient of variation of the method in sheep plasma was 4.7% at the concentration of 20 pg/ml, 1.2% at 200 pg/ml, and 3.3% at 5,000 pg/ml. The intraassay coefficient of variation of the method in sheep plasma was 3.0% at 5 pg/ml, 1.3% at 20 pg/ml, 1.4% at 200 pg/ml, and 0.77% at 5,000 pg/ml. The limit of quantitation was 25 pg/ml for sheep CSF samples when 0.2 ml CSF was used. The intraassay coefficient of variation of the method in sheep CSF was 0.84% at 3.14 ng/ml and 0.17% at 6.08 ng/ml.
Noncompartmental pharmacokinetic techniques were used to determine the mean residence time and CSF bioavailability of the drug after intravenous and epidural injections. CSF bioavailability was defined, as in previous studies, 8,9as the fraction of the administered dose appearing in CSF. Protein binding was not determined. Compartmental models were fit using the nonlinear mixed-effects regression techniques of the NONMEM software program (NONMEM Project Group, University of California–San Francisco, San Francisco, CA). 14Model fits were made using several variations of one-, two-, and three-compartment models with and without absorption rates and lag times. All pharmacokinetic models were parameterized using the microrate constants k i,j. These microrate constants equal the fraction of drug flowing from one compartment to another per unit time (minutes). The subscripts i,j indicate the model's compartment. For example, k 12is the fraction of drug in compartment 1 that distributes into compartment 2 per minute. Assay data from the same sheep on different occasions were, for modeling purposes, treated as separate individuals. Volumes and clearance were calculated from the microrate constants for those more familiar with the volume/clearance parameterization.
The models that simultaneously minimized both the extended least-squares objective function and parameter number were judged best. Selection was based on the Swartz–Bayesian criterium and graphic examination of residual plots and individual model fits. Median performance error and median percent bias were calculated by standard formulae.
Spinal Cord Blood Flow Effects
Eleven adult ewes were anesthetized and prepared surgically as described in the previous section. In addition, a 7-French soft catheter was advanced under pressure waveform guidance into the left ventricle through the left internal carotid artery. The pulmonary artery catheter was inserted under pressure waveform guidance on the day of surgery, and all animals recovered for at least 5 days before study.
Each animal was studied on two occasions, separated by 48 h. On the day of each study the location of pulmonary arterial and left ventricular catheters were confirmed by examination of the pressure waveforms. Femoral and pulmonary artery catheters were connected to transducers, and mean pressures and heart rate were measured via a computer-based data acquisition system, with average values calculated every minute. After 30 min of baseline recordings, colored microspheres were injected through the left ventricular catheter. Animals were randomized to receive either 100 or 300 μg MPV-2426 by slow intrathecal injection, followed in 30 and 120 min by left ventricular injection of microspheres of a different color. Each left ventricular injection consisted of 20 × 10 6presonicated colored microspheres (15.5 μm in diameter; Ultrasphere, Los Angeles, CA). Reference sampling from both femoral arteries (8 ml/min) was begun 30 s before microsphere injection and lasted 120 s. At the time of each microsphere injection, cardiac output was determined by thermodilution in triplicate with injection of 5 ml iced dextrose solution, and arterial blood samples were obtained for blood gas tension and p H analysis.
At the end of the second study day, deep anesthesia was induced with pentobarbital (10 mg/kg intravenously) followed by intravenous injection of saturated potassium chloride to produce cardiac arrest. A dorsal laminectomy from the lumbar to the cervical spine was performed, the intrathecal catheter tip position was verified, and the entire spinal cord was removed to determine spinal cord blood flow. The spinal cord was divided into lumbar, thoracic, and cervical segments. After the dura and pia mater were removed, the spinal tissue was weighed and analyzed as described below. In addition, samples of the left and right kidneys and the left ventricular endocardium and epicardium were removed for blood flow analysis.
Tissue and reference blood samples were processed according to an established protocol 15using tissue digestion by alkaline hydrolysis and counting of spheres using a true color digital image acquisition and custom-designed software that allows online computation of regional blood flow. Spinal cord and renal blood flow, represented as ml · min−1· 100 g tissue−1were determined by: where Qt = tissue blood flow (ml/min), Qr = reference sample blood flow (ml/min), Ct = number of microspheres in the tissue sample, and Cr = number of microspheres in the reference sample.
Antinociception data were converted to percent maximum possible effect (%MPE), where %MPE =(postdrug threshold − baseline threshold)/(20 N − baseline threshold)× 100 (where 20 N is the maximum applied force). All antinociception data were non-normally distributed and are presented as median ± 25th and 75th percentiles. ED50and ED95were determined by log-linear regression on the entire data set for that route of administration. Other data are presented as mean ± SEM.
Hemodynamic data were averaged within each animal for the 5-min period surrounding each cardiac output determination and were analyzed by one-way analysis of variance for repeated measures followed by Dunnett's test. P < 0.05 was considered significant.
All animals recovered normally from surgery, with absolute hemodynamic values at rest similar to multiple previous studies from this laboratory. Animals tolerated all drug doses without distress. Mild sedation was present after the larger doses of MPV-2426 by each route, but level of sedation was not quantified in this study. Two animals that received 300 μg intrathecal MPV-2426 showed transient labored breathing. In one case, this began 10 min after injection and lasted for 10 min. Arterial blood analysis showed a p H of 7.26, oxygen partial pressure (PO2) of 34 mmHg, and carbon dioxide partial pressure of 57 mmHg during the episode. In the other case, labored breathing was noted 80 min after injection and lasted for 20 min. Arterial blood analysis revealed a p H of 7.41, PO2of 65 mmHg, and carbon dioxide partial pressure of 38 mmHg. There were no cases of respiratory distress as evidenced by rapid, labored breathing after intravenous MPV-2426.
In contrast to pilot studies, which suggested a brief (< 15 min) period of antinociception from intravenous MPV-2426 at doses > 10 μg, the drug produced no antinociception when administered intravenously in doses up to 80 μg in this randomized, blinded study (fig. 1). However, the wide interquartile range, representing a > 50% MPE in three, one, and three animals after 2-, 20-, and 80-μg doses, respectively, reflects a wide variability in response to intravenous injection. The regression of %MPE versus dose for intravenous injection was nonsignificant. In contrast to intravenous injection, MPV-2426 produced dose-dependent antinociception after intrathecal and epidural administration (fig. 1, insert). Regression analysis showed an ED50for MPV-2426 after epidural injection of 202 μg and after intrathecal injection of 49 μg. Antinociception from intrathecal injection of 100 μg MPV-2426 was still present 2 h after injection (fig. 1).
Hemodynamic Effects and Pharmacokinetics
Intrathecal injection, even at a dose of three times the ED95, produced no statistically significant decrease in systemic arterial pressure, heart rate, cardiac output, central venous pressures, or arterial PO2(figs. 2–4and tables 1 and 2). However, as previously noted, there were two sheep in the antinociception part of the study that showed transient respiratory effects after intrathecal injection.
Intravenous injection of the largest dose (300 μg) produced a transient (< 15 min) significant increase in systemic arterial pressure, decrease in heart rate, decrease in cardiac output, increase in central venous pressures, and decrease in arterial PO2(figs. 2–4and tables 1 and 2). Similarly, epidural injection resulted in a dose-dependent, statistically significant increase in systemic arterial blood pressure, reduction in heart rate and cardiac output, increase in central venous pressures, and decrease in arterial PO2(figs. 2–4and tables 1 and 2). No route or dose of MPV-2426 injection altered arterial p H or carbon dioxide partial pressure (data not shown).
All routes of injection resulted in an increase in systemic arterial blood pressure, as previously described. This was likely a result of systemic effects because there was no difference in the dose–response curve for this effect with route of injection.
The intravenously administered drug maintained constant CSF drug concentrations after the initial redistribution and thus could not be modeled within the time frame of the study. CSF mean residence time by noncompartmental analysis varied from approximately 50 min after spinal infusions to 87 and 101 min after epidural and intravenous injections, respectively (table 3). CSF bioavailability was 7% after epidural and 0.17% after intravenous drug injections (table 3).
Compartmental modeling was performed on each data set (spinal and epidural) separately and on the combined data set. The data from two dosing studies were excluded when fitting the models because of dose–response curves that were inconsistent with all of the other study data (nonsensical, repetitive, and large increases and decreases in concentration over time).
Two compartment models best fit our epidural, spinal, and combined CSF data sets. Analysis of the epidural data in both the epidural and combined CSF models included a bioavailability constant (F 1) and a first-order absorption rate (k a) for modeling the additional movement of the drug from the epidural space to the CSF space. Analysis of the intrathecal data in both intrathecal and combined CSF models were modeled to directly enter the CSF space without adjustments for F 1or k a. The bioavailability constant was fixed at 7% based on the aforementioned determination, whereas the rate of epidural to CSF space absorption was fitted by the compartmental models. Our best models characterized the intersheep variability as log normal in distribution and modeled the intrasheep residual error using a combined additive and constant coefficient of variability model (residual error =ς1+ς2* predicted concentration). We were unable to determine the standard errors of our estimated parameters with NONMEM because of its inability to estimate the covariance matrix.
Our combined epidural/spinal model results were in reasonable agreement with the data, as evidenced by examination of the time versus concentration plots (fig. 5), the observed/predicted plots (fig. 6), and low median performance error (21%) and bias (0%). Table 4lists our best model parameter estimates for CSF after spinal and epidural injections. In both the spinal and epidural models, the pharmacokinetic models grossly underestimate CSF drug concentrations for the first 5 min after injection (2–50-fold), and these data are not shown and were not modeled. However, after this initial period, the models performed well, correctly predicting CSF concentrations within half to double the true drug concentrations (fig. 6).
Spinal Cord Blood Flow Effects
Intrathecal injection of 100 and 300 μg MPV-2426 significantly increased mean arterial pressure 30 min after injection and decreased cardiac output 30 and 120 min after injection, but only the 300-μg dose decreased heart rate (table 5). Renal cortical blood flow was similar in tissue from right and left kidneys, and neither dose of MPV-2426 affected renal blood flow or left ventricular endocardial or epicardial blood flow (table 5).
Intrathecal MPV-2426 did not significantly alter the blood flow in cervical spinal cord, but decreased thoracic spinal cord blood flow in a dose-independent manner (fig. 7). The 100-μg but not the 300-μg dose reduced lumbar spinal cord blood flow (fig. 7).
The α2-adrenergic agonist clonidine has unique efficacy in patients with neuropathic pain but is sometimes not tolerated by this patient population because of hypotension after epidural and intrathecal administration. 3,16Similarly, hypotension and bradycardia occur with epidural and intrathecal administration of this agent in postoperative and obstetric patients, 2thereby limiting the usefulness of this therapy. The current study suggests that the novel α2-adrenergic agonist MPV-2426 may exert less of these adverse effects than does clonidine, although this conclusion is not certain even in sheep because this was not a comparative trial.
α2-Adrenergic agonists produce antinociception in animals after systemic, intrathecal, and intraventricular administration. However, these agents act primarily by actions in the spinal cord. Thus, spinal cord transsection does not abolish antinociception from systemic clonidine in animals, 17and analgesia in human volunteers is greater after intrathecal than intravenous injection of clonidine. 18There is a close correlation between clonidine concentration in CSF and analgesia to experimental pain stimuli in humans, and analgesia after lumbar epidural injection is restricted to the lower extremity, both consistent with a spinal site of action in humans. 19Finally, analgesia is more profound with less drug when clonidine is administered intrathecally than epidurally, and more profound epidurally than intravenously, in postoperative patients. 20,21
MPV-2426 was more potent intrathecally than epidurally in the current study, suggesting a spinal site for antinociception. MPV-2426 only had a limited bioavailability after epidural administration in sheep (7%) compared with clonidine (14%)8and dexmedetomidine (22%). 9We failed to observe antinociception from intravenous administration of this compound in the current study. It is possible that the dose levels chosen, based on evanescent but measurable antinociception after small intravenous doses in pilot studies, were inadequate. Alternatively, perhaps MPV-2426 poorly penetrates the blood–brain barrier after intravenous injection. This would be consistent with the pronounced peripheral hemodynamic effects of this compound after intravenous injection (vasoconstriction), with little subsequent central effect (hypotension). Indeed, ST-91, a hydrophilic α2-adrenergic agonist, has a similar profile to MPV-2426 in rats. 22
The ED50of MPV-2426 after intrathecal administration is approximately 30% less than that of clonidine and dexmedetomidine in sheep. 23However, because the relative potencies of clonidine and dexmedetomidine differ considerably between rats and sheep, 22and because only one of them has been systematically administered to humans, it is not certain whether this increased potency of MPV-2426, compared with clonidine, will be observed clinically.
The hemodynamic profile of MPV-2426 is qualitatively similar to other α2-adrenergic agonists. Thus, large intravenous or epidural doses result in increased systemic and pulmonary arterial pressure, because of increased vascular resistance in both beds, and a reflex reduction in heart rate and cardiac output. 24This effect is less pronounced with decreased dose and with intrathecal compared with intravenous administration. 9
However, there are quantitative differences in the hemodynamic effects of MPV-2426 compared with other α2-adrenergic agonists. Of particular importance is the lack of hypotension of intrathecal MPV-2426 at the ED95level for antinociception. This dose produced a minor increase in blood pressure, which was statistically significant in only one of the two studies involving blood pressure measurement (the spinal cord blood flow study). The dose–response curve for α2-adrenergic agonists on blood pressure is U-shape after intrathecal administration, reflecting initially increasing sympatholysis and, at high doses, a counterbalancing peripheral vasoconstrictive effect. 7Clonidine produces clear decreases in blood pressure at analgesic doses in humans. 25It is possible that the relative potency of MPV-2426 at antinociceptive sites compared with sites for hypotension differs from other α2-adrenergic agonists.
These data suggest that intrathecal injection of MPV-2426 might produce less hypotension than clonidine in humans. It could be argued that sheep are less sensitive to the hypotensive actions of α2-adrenergic agonists than humans. As such, the maximal reduction in systemic arterial pressure usually achieved with intrathecal clonidine is approximately 15% in sheep, 7but can be 30–40% in humans. 2In addition, α2-adrenergic agonists decrease blood pressure more in hypertensive than in normotensive subjects, and this issue was not addressed in the current study. Power analysis showed that we were unable to observe a decrease as large as 10% in the current study, or two thirds as great as clonidine. Thus, one would anticipate that the degree of hypotension from MPV-2426 will be less than that from clonidine in humans, a hypothesis that will clearly be tested.
α2-Adrenergic agonists produce transient hypoxemia and labored breathing in sheep but not in other species, including humans. 9,10There were only two cases of transient hypoxemia noted (both in the antinociception part of this study) in the 40 sheep studied with MPV-2426. The reasons for this discrepancy in sheep are unclear.
The effects of intraspinally administered α2-adrenergic agonists on spinal cord blood flow are not consistent in previous studies; increases, decreases, 26,27or no change 28,29have been observed. The magnitude of the reduction in spinal cord blood flow observed in the current study is similar to that observed from intrathecal clonidine in pigs 27and rats 26and is less than that observed after intrathecal injection of some local anesthetics 30and less than that observed from known neurotoxins. 31The reduction produced by MPV-2426 in spinal cord blood flow is clearly a drug effect because it is greater near the site of injection (thoracic) than further away (lumbar and cervical). The reduction in spinal cord blood flow observed after intrathecal clonidine in rats was consistent with an autoregulatory reduction commensurate with the reduction in metabolic rate. 26However, the effect of MPV-2426 on spinal cord metabolism was not examined in the current study; therefore, it is unknown whether this reduction in flow is secondary to metabolic-flow coupling, to loss of autoregulation in the face of reduced cardiac output, or to active vasoconstriction. Thus, although one cannot exclude the possibility that MPV-2426 could under pathologic conditions (e.g. , severe hypotension, marginal spinal cord blood flow before drug) produce spinal cord ischemia, the degree of reduction in blood flow from this agent is small and in the range previously observed with other α2-adrenergic agonists and local anesthetics.
In summary, the novel α2-adrenergic agonist MPV-2426 produces antinociception with a potency order of intrathecal > epidural >> intravenous. Within the power limits of the study, MPV-2426 at antinociceptive and greater doses does not decrease blood pressure, whereas large doses intravenously or epidurally increase blood pressure and decrease heart rate and cardiac output. These results are consistent with the hypothesis that MPV-2426 penetrates the blood–brain barrier poorly after systemic administration. Similar to other α2-adrenergic agonists, intrathecal MPV-2426, at maximal antinociceptive dose levels and above, reduces local spinal cord blood flow, but whether this is because of active vasoconstriction or in response to reduced metabolism was not tested in the current study. Sedation, a major and sometimes therapy-limiting side effect of α2-adrenergic agonists, was not assessed in this study; therefore, whether MPV-2426 might cause more or less sedation than clonidine is unknown. Although other preclinical toxicity studies must be performed before introduction into humans, these results suggest that MPV-2426 may be a useful spinal analgesic in humans.