Central and Peripheral Analgesia Mediated by the Acetylcholinesterase-Inhibitor Neostigmine in the Rat Inflamed Knee Joint Model

Transcription

1 Central and Peripheral Analgesia Mediated by the Acetylcholinesterase-Inhibitor Neostigmine in the Rat Inflamed Knee Joint Model H. Buerkle, MD, M. Boschin, MS, M. A. E. Marcus, MD, G. Brodner, MD, Pm, R. Wiisten, MD, and H. Van Aken, MD, PhD Klinik und Poliklinik fiir Antisthesiologie und operative Intensivmedizin, Westfslische Wilhelms-Universitst Miinster, Monster, Germany Intrinsic cholinergic inhibitory pathways present a key modulating system in pain perception. The use of intrathecal (IT) acetylcholinesterase-inhibitors, such as neostigmine, result in analgesia in both preclinical and clinical models. However, whether IT neostigmine suppresses tonic persistent pain or has peripheral sites of antinociceptive action has not been determined. Thus, we studied central (IT) and peripheral (intraarticular; IA) neostigmine in a rat inflamed knee joint model. Inhibition of thermal and mechanical hyperalgesia was assessed over 28 h using a modified Hargreaves box and von Frey hairs, respectively. IT neostigmine resulted in a dose-dependent thermal analgesia (50% of maximal effective dose [ED,,] O-4 h: 6.6 pg, h: 9.4 pg) and mechanical analgesia (ED, O-4 h: 3.5 kg, h: 4.3 Fg). IT atropine reversed analgesia by IT neostigmine. IA neostigmine also resulted in an IA atropine reversible dose-dependent increase of thermal analgesia, although it did not exceed 60% of a maximal possible analgesic effect with the largest applied dose (ED,, O-4 h: 76.2 pg, h: pg). Partial suppression of mechanical hyperalgesia was observed after IA neostigmine. We conclude that centrally administered neostigmine modulates thermal and mechanical antinociception in this animal model of inflammatory pain. These data suggest a peripheral site of muscarinic antinociception. Implications: This animal study shows that administration of the acetylcholinesteraseinhibitor neostigmine results in enhanced levels of the endogenous neurotransmitter acetylcholine, which seems to act as one of a group of analgesia-modulating compounds at central and peripheral sites in inflammatory pain. (Anesth Analg 1998;86: ) I n preclinical models corresponding to postsurgical inflammatory pain, such as a kaolin/carrageenan injection (KC) into the knee joint of the rat, an increasing barrage of activity of small myelinated and unmyelinated afferent fibers (A-6 and C- fibers, respectively) result in an amplification of spinal dorsal horn transmission with subsequent development of primary and secondary hyperalgesic pain (1). This hyperalgesia is characterized by increased sensitivity to noxious stimuli (e.g., thermal activation) or to otherwise innocuous mechanical stimuli (allodynia). The initial cascade mediating facilitated behavioral and physiological response is the This study was funded by a grant from the Klinik und Poliklinik fiir Anlsthesiologie und operative Intensivmedizin. Accepted for publication February 4, Address correspondence to Dr. med. Hartmut Buerkle, Klinik und Poliklinik fiir Anlsthesiologie und operative Intensivmedizin, WestMlische Wilhelms-Universitlt Miinster, Albert-Schweitzerstr. 33, Miinster, Germany. Address to burkle@unimuenster.de. result of activation of wide dynamic neurons by a range of excitatory neurotransmitters (2). Cholinergic agonists or cholinesterase-inhibitors administered spinally inhibit nociception in a dose-dependent manner by increasing the endogenous neurotransmitter acetylcholine. The analgesic action of neostigmine is mediated through cerebral cholinergic and noradrenergic pathways, as demonstrated in both preclinical and clinical trials (3-7). The antinociceptive effects of spinal neostigmine have only been assessed for acute postoperative pain. Moreover, muscarinic receptors and their analgesic properties have been studied only in relation to central stimulation of this receptor system (8). Peripheral afferent fibers also express muscarinic receptors, which may result in analgesia (9,lO). Analgesia achieved by peripheral delivery of neostigmine is attractive because it might not be limited by side effects such as nausea, vomiting, and pruritus caused by cephalic spread of centrally applied neostigmine (4). In addition, the inhibition of peripheral afferent input might be an appropriate therapy to reduce concomitant central, i.e., spinal, sensitization by suppressing by the International Anesthesia Research Society /98/$5.00 Anesth Analg 1998;86:

2 1028 REGIONAL ANESTHESIA AND PAIN MANAGEMENT BUERKLE ET AL. SPINAL AND PERIPHERAL ANALGESIC ACTION OF NEOSTIGMINE ANESTH ANALG 1998;86: the release of excitatory neurotransmitters through peripheral stimuli blockade. Thus, secondary hyperalgesia might be diminished. To determine the central and peripheral antinociceptive effects of neostigmine, we performed the present study in the rat inflamed knee joint model, a well defined model of persistent inflammatory pain that permits the study of primary and secondary hyperalgesia. Methods Animal surgery and testing protocols were approved by our institutional animal care committee. All procedures were performed according to the guidelines issued by the International Association for the Study of Pain. Rats (male Sprague-Dawley) were kept in individual cages on a 12-h light/dark cycle with water and food ad libitum. Animals were randomly assigned to treatment groups and were only used once. The animals were killed with pentobarbital after each experiment. For the spinal delivery of study drugs, animals were prepared with intrathecal (IT) catheters using a modified version of techniques described previously (11). After recovery from anesthesia, all animals received 20 PL of IT lidocaine 1% to assure proper location of the IT catheter by showing bilateral similar, reversible motor blockade for both hindlimbs. Only animals with normal motor function after this validation of the IT placement were used in the following experiments. A postsurgical recovery period of 5 days was allowed for all animals. After assessing baseline values for thermal and mechanical nociception and supraspinal side effects (see below), animals were briefly anesthetized with halothane 2%-3%, and the right knee joint was injected with 0.1 ml of a mixture of 3% kaolin and 3% carrageenan (KC) (12). Control animals received an injection of 0.1 ml of saline into the right knee joint. Recovery from anesthesia occurred within 5-10 min. For all animals, the degree of the inflammatory response was measured by assessing the circumference of the right injected knee and the left noninjected knee joints 24 h after the induction of inflammation. The threshold for thermal stimuli (thermal withdrawal reflex) was assessed by an observer blinded to the compound injected into the knee joint using a modified technique originally described by Hargreaves et al. (13). Each animal was placed in clear plastic cages (9 X 22 X 25 cm) on top of a glass plate with the surface temperature maintained at 30 C by using a feedback-controlled heater fan. The animal was left on this surface for at least 15 min before any stimulation to allow for proper adaptation to the environment. The thermal stimulus, a halogen bulb, was positioned under the glass and focused either on the plantar surface of the ipsilateral paw of the inflamed or the contralateral paw of the noninflamed knee joint. The halogen bulb was switched on by a timer and automatically switched off when the paw was raised (thermal withdrawal reflex), as sensed by photo diodes or after 20 s (cutoff time). Correct positioning of the bulb was determined by a mirror attached to the bulb, which allowed visualization of the undersurface of the paw. Light beam intensity was monitored by a measurement of bulb current, and the stimulus intensity was calibrated daily by assessing the temperature change after 10 s sensed by an underglass thermocouple (Ti,* = 0.2 s). The intensity of the light was adjusted and maintained at mean (? SD) baseline latenties of s. Withdrawal latency was measured to the nearest 0.1 s. Nagasaka et al. (14) found that the intraarticular (IA) injection of saline alone induces a brief hyperalgesia lasting for 30 min. During this time, no testing was performed. To test the presence of tactile allodynia, rats were placed on an elevated wire mesh bottom in clear plastic cages and were allowed to adapt to the new environment for at least for 5 min before testing. Calibrated von Frey hair filaments were used in a technique using the Dixon up and down method described by Chaplan et al. (15). Briefly, calibrated von Frey hairs were applied to the ipsilateral paw of the inflamed knee joint and to the contralateral paw of the noninflamed knee joint from below the mesh floor. The minimal stimulation period was 5 s, with a maximum of six stimuli per trial and paw. Stimuli were delivered in lo- to 15-s intervals. A sharp, distinct withdrawal (mechanical withdrawal reflex) was considered a positive response. To score the behavioral changes before and during treatment of the animal, a supraspinal index was used. The index consisted of four behavioral variables that are dose-dependently blocked by opiates (16). These measured responses were pinna reflex, cornea reflex (both evoked by light touch of the surface of pinna or cornea with a small piece of PE-10 tubing), evoked movement (startle reflex evoked by tapping on the cage wall), signs of spontaneous movement (e.g., grooming, chewing, ambulating). Each variable was scored as 0 = normal (brisk pinna/cornea reflex response or startle reflex, spontaneous movement within 30 s of the assessed time point), 1= attenuated (touch with tubing of pinna or cornea reflex or knocking on the cage wall results in a slow reflex behavior and has to be repeated; touch or knocking has to be repeated at least twice), or 2 = completely absent (no reflex was evident after touching the cornea or pinna three times on both sides, no startle behavior was displayed after knocking against the cage wall three times, and no spontaneous movement observed for >l min). To permit a quantitative assessment of the behavioral effect, the supraspinal side effects index,

3 ANESTH ANALG 1998;86: REGIONAL ANESTHESIA AND PAIN MANAGEMENT BUERKLE ET AL SPINAL AND PERIPHERAL ANALGESIC ACTION OF NEOSTIGMINE which consisted of summing the individual scores for the four measurements at each time point, permitting a total score of 8 (4 X 2), was used. Tests for the supraspinal side effects were performed concomitantly with the antinociceptive testing. Treatment groups received different doses of neostigmine IT (2,5,10,30 pg) in volumes of 10 /JL of IT followed by 10 PL of saline to flush the catheter or neostigmine IA (10,30,100 kg) in volumes of 100 FL (IA), respectively. Atropine was used as an antagonist either IT (30 pg) or IA (300 pg). Control groups received either IT saline 20 PL or IA saline 100 pl. All experiments were performed in a randomized manner, and each measurement was obtained by an observer blinded to the treatment. Dose-response curves were derived from thermal antinociceptive effects. Each animal was tested twice on two consecutive days after receiving the KC injection on Day 1. Testing intervals were 30 min before KC injection and at specific intervals thereafter. The drug was delivered 30 min after KC. On both days, tests were performed 30,60,120, and 240 min after drug delivery (IT or IA). Experiment 1 To assess the effect of an IA injection of KC mixture on thermal nociception and mechanical stimuli and supraspinal side effects, rats were divided into four groups (n = 6-S/group): IT saline + IA saline, IT saline + IA KC, IA saline, or IA KC, and tests were performed. Experiment 2 To assess the effect of IT neostigmine on thermal nociception and mechanical stimuli and supraspinal side effects in rats with knee joint inflammation, four groups of rats (n = 5-6/group) were assigned to receive IT neostigmine. In a separate group (n = 4), the antagonist atropine was delivered 10 min before the largest applied dose of IT neostigmine. Experiment 3 To assess the effect of IA neostigmine on thermal nociception and mechanical stimuli and supraspinal side effects in rats with knee joint inflammation, three groups of rats (n = 5-S/group) were assigned to receive IA neostigmine. The antagonist atropine was delivered IA 10 min before the largest dose of IA neostigmine used in a separate group (n = 4). The percentage of maximal possible effect (%MPE) of analgesia in animals with an inflamed knee joint was calculated. Dose-response curves are presented as the %MPE of thermal or mechanical analgesia. Statistical analysis of the data was performed by using nonparametric tests. The Mann-Whitney ranked sum test and the Kruskal-Wallis test were used. Multiple comparisons after the Kruskal-Wallis test were performed using Scheffe s test. P < 0.05 was considered statistically significant. For all drugs, the doseresponse analysis was performed as described by Tallarida and Murray (17). The 50% effective dose (ED,,) and the 95% confidence intervals were calculated using the least squares linear regression model for log dose values (17). Results Experiment 1 Injection of KC into the knee joint resulted in a reliable, persistent thermal hyperalgesia with a mean decrease of thermal withdrawal responses of the ipsilatera1 paw of the inflamed knee joint to 5.9 s s for control groups regardless of IT catheter implantation, IT saline injection, or IA saline delivery (P < 0.05). Thermal hyperalgesia was seen within min after induction and remained stable until testing was completed on Day 2 (28 h after KC injection) (Fig. 1A). Similarly, a persistent decrease in the mechanical withdrawal thresholds of the ipsilateral paw of the inflamed knee joint were observed with the onset time min after drug delivery (maximal possible effect of inhibition of mechanical withdrawal response in percentage; % MPI) (Fig. 1B). Inflammation increased the circumference of the right injected knee joint by 1.7? 0.7 cm compared with the left noninjected joint (P < 0.05) over 28 h. Assessment of the different supraspinal activities revealed no increase in side effects for all control groups (Index = 0). Experiment 2 IT neostigmine resulted in a dose-dependent inhibition of the thermal withdrawal response (Fig. 2). The peak effect on Days 1 and 2 was observed after 30 min (Fig. 3). Mechanical withdrawal response was also dose-dependently suppressed by IT neostigmine (Fig. 4). The peak effect of IT neostigmine was observed at 30 min on Days 1 and 2 (Fig. 4). The area under the curve (AUC) of inhibition of the thermal withdrawal response after the largest dose of IT neostigmine (30 Fg) was significantly different (P < 0.05) for Days 1 (59.2%) and 2 (46.6%). No significant difference for the AUC was observed for inhibition of the mechanical withdrawal response: 50% (Day 1) and 48.2% (Day 2). After IT neostigmine, there was no increase in the supraspinal index at the doses tested. However, 70% of the animals in the IT 30 pg neostigmine group showed agitation and

4 1030 REGIONAL ANESTHESIA AND PAIN MANAGEMENT BUERKLE ET AL. ANESTH ANALG SPINAL AND PERIPHERAL ANALGESIC ACTION OF NEOSTIGMINE 1998;86: a b Thermal Hyperalgesia on day 1 and day 2 zloo u left paw day1 $ 80 + left paw day2 z c right paw day1 P z 40 + right paw day2 E 5 20 g 0 E Mechanical Hyperalgesia on day 1 and day 2 2 loo g 80 -!J- left paw day1 c 5 60 h left paw day2 5 + right paw day & right paw day2 5 c 20 P 0 E s Figure 1. A, Time course of thermal hyperalgesia after induction of knee joint inflammation into the right knee. %MPI = percent maximal possible inhibition (100% MI 1 = 20 s of thermal stimulation). The left paw of the contralateral noninflamed knee joint (open symbols) showed no decrease in response latency on Days 1 and 2. The right ipsilateral paw of the inflamed knee joint (filled symbols) showed a decrease in thermal response latency, i.e., thermal hyperalgesia, on Days 1 and 2. Each point presents the mean -c SD data for 14 rats over 28 h. **Significant difference between the ipsilateral and contralateral paws(p < 0.05). B, Time course for mechanical hyperalgesia after induction of knee joint inflammation into the right knee. Mechanical hyperalgesia is defined as the maximal applied force of the von Frey hairs as a percent maximal possible inhibition (%MPI) to withdraw the paw (100% MI 1 = 15.1 g). Each point presents the mean 2 SD data of 14 rats over 28 h. **Significant difference between the ipsilateral and contralateral paws (P < 0.05). tremor, which were not assessed by the index used. No difference between the IT neostigmine and control groups was observed regarding the knee joint circumference over 28 h. Experiment 3 IA neostigmine resulted in a dose-dependent inhibition of the thermal withdrawal response (Fig. 5). Concomitantly, a minor increase of thermal antinociception was observed for the contralateral paw of the noninflamed, noninjected knee joint. However, the latency increase for the ipsilateral paw of the treated, Jl Dose response curve: Thermal analgesia after IT neostigmine day1 /day2 r T rlghi paw day1 1 - right paw day2 10 IT dose of neostigmine (in pg) Figure 2. The 50% effective dose of IT neostigmme for thermal analgesia. Analgesic response to the thermal stimulus was performed on Days 1 and day 2 after the IT delivery of neostigmine. %MPE = percent maximal possible effect of analgesia (100% MPE = no withdrawal after 20 s of thermal stimulation). **Significant difference for the contralateral paw of the noninflamed knee joint versus the right ipsilateral paw of the inflamed knee joint for 5 pg and 10 pg of IT neostigmine (P < 0.05). Time response curve for IT neostigmine (3Opg): Thermal analgesia + right paw day 1 & right paw day 2 Figure 3. Thermal analgesia time effect curves for the largest applied dose of IT neostigmine (30 pg). Testing was performed on Days 1 and 2. %MPE = percent maximal possible effect (100% MPE = no withdrawal after a time of 20 s of thermal stimulation). Each point presents the mean t SD of five animals. inflamed knee joint was always statistically significant different (P < 0.05). The analgesic effect of neostigmine for the thermal withdrawal reflex did not exceed 60% of the %MPE of analgesia, due to systemic adverse effects (see below). No difference between the IA neostigmine and control groups was observed regarding the knee joint circumference over 28 h. As for the IT delivery, the peak effect of IA neostigmine was observed on Days 1 and 2 after 30 min (Fig. 6) for suppression of the thermal withdrawal reflex. The AUC for thermal analgesia after the largest dose of IA neostigmine was 24.3% on Day 1 and 13.9% on Day 2 (P < 0.05). The supraspinal index revealed a dosedependent increase after IA neostigmine with maximal scores of (P < 0.05). This increase of adverse

5 ANESTH ANALG REGIONAL ANESTHESIA AND PAIN MANAGEMENT BUERKLE ET AL ;86: SPINAL AND PERIPHERAL ANALGESIC ACTION OF NEOSTIGMINE Dose response curve: Mechanical analgesia after IT neostigmine day 1 /day2 Time response curve for IA neostigmine: Thermal analgesia - left paw clay1 + lelt paw day right tpaw day1 & right paw day IT dose of neostigmine (in pg) Figure 4. The 50% effective dose of IT neostigmine for mechanical analgesia. Mechanical analgesia is defined as no withdrawal to the maximal applied force of the von Frey hairs as a percent maximal possible effect of analgesia (%MPE) (100% MPE = 15.1 g). Testing was performed on Days 1 and 2. Each point presents the mean t- SD of five animals. The test was performed only on the ipsilateral right paw of the inflamed knee. There were no significant differences between the days Dose response curve: Thermal analgesia after IA neostigmine dayliday2.m --n-- Ibit pawoayt I w 80 + Ien paw clay* i? --R-- right paw day1 ** 5 -;L- right paw day2 iij 60 E *?I h 20 a 8 ;>,,,, 4,,.,,,,, IA dose of neostigmine (in pg) Figure 5. The 50% effective dose of IA neostigmine for thermal analgesia. Testing was performed on Days 1 and 2. %MPE = percent maximal oossible I effect of analgesia v (100% ~ MPE = no withdrawal after 20 s of thermal stimulation). Each point presents the mean I SD of four to eight animals. *Significant difference (P < 0.05) between paws at each dose on both days. **Significant difference for the inflamed paw on both days compared with the noninflamed paw. effects was accompanied by initial agitation, hyperactivity, and hypersalivation, which were not assessed by the index used. Because of the systemic side effects, no reliable measurements of IA delivery of neostigmine and mechanical withdrawal response could be accomplished in this model. Discussion Animal and human trials demonstrate that central (spinal or epidural) administration of a cholinesterase inhibitor such as neostigmine results in dosedependent analgesia (5,6,8). However, its analgesic ** Figure 6. Thermal analgesia time effect curves for the largest dose of IA neostigmine (100 pg) applied in this model and the 50% effective dose of IA neostigmine for thermal analgesia. Testing was performed on Days 1 and 2. %MPE = percent maximal possible effect of analgesia (100% MPE = no withdrawal after 20 s of thermal stimulation).-each point presents the mean t SD of four to eight animals. Significant difference (P < 0.05) between the paws of the injected and noninjected knee joint on *one day or on **both days. effect is limited by adverse effects such as nausea and pruritus caused by spinal rostra1 spread of neostigmine (4). Thus, neostigmine has been used only in acute pain states and not for repeated administrations in persistently painful conditions. As for opioids, there might be time-dependent differences in the analgesic efficacy of neostigmine because of inflammatory processing (18). Animal models simulating clinical pain and analgesia are numerous (19). The inflammatory knee joint model presents a reliable model of prolonged persistent pain (12). It is a useful model for the assessment of the central component of tissue injury. After the injection of the inflammatory mixture (KC), there is an initial acute barrage of activity of small myelinated and unmyelinated fibers, which results in the release of glutamate and other endogenous neurotransmitters at the spinal dorsal horn. This release of excitatory modulators of pain results in thermal and mechanical hyperalgesia, as observed in our study and described by others (12). In this study, we hypothesized that the local application of an analgesic would inhibit primary hyperalgesia; the subsequent decrease of afferent input to the spinal cord would then decrease secondary hyperalgesia. For more conclusive studies of postoperative pain in animals, thermal and mechanical pain behavior should be assessed. Therefore, we evaluated the analgesic properties of centrally or peripherally administered neostigmine in this model of the inflamed knee joint and assessed the subsequent inhibition of two different modalities of inflammatory pain-thermal hyperalgesia and mechanical hyperalgesia. Central delivery of neostigmine potently inhibited both pain modes, which suggests the suppression of different afferent fiber systems-a& C, and AP fibers.

6 1032 REGIONAL ANESTHESIA AND PAIN MANAGEMENT BUERKLE ET AL. ANESTH ANALG SPINAL AND PERIPHERAL ANALGESIC ACTION OF NEOSTIGMINE 1998;86: Moreover, the small doses of neostigmine necessary to achieve the obtained ED,, for mechanical analgesia suggested that IT neostigmine preferentially suppressed mechanical allodynia. Interestingly, the spinal delivery of neostigmine seemed to have a greater effect on the analgesic response of the ipsilateral paw of the inflamed knee joint than on the contralateral paw of the noninflamed knee. It is conceivable that, like the unilateral neuronal barrage by activated afferent fibers, there might be a unilateral acetylcholine release at the level of the spinal dorsal horn. Inflammation clearly activates opioid receptor systems (20); there is a breakdown of the perineurium through inflammatory processes, which results in a facilitated accessibility of opioid receptors. Silent opioid receptors are activated by inflammation, and axonal transport of newly synthesized opioid receptors from the dorsal horn to the peripheral nociceptor terminal is induced (l&21,22). These changes provide an enhanced analgesic response to exogenous opioid delivery. Despite some similarities between opioidmediated analgesia and muscarinic analgesia, in our model of inflammatory pain, the analgesic efficacy of neostigmine was not altered by repeated injection of an inflammatory agent. The observed decrease of the AUC for the thermal antinociception after the second day of delivery of IT neostigmine could be due to the development of acute tolerance to IT neostigmine or, more likely, to a decrease of endogenous spinal acetylcholine release by ongoing inflammation. The IA administration of neostigmine revealed a dose-dependent, peripherally mediated analgesic effect of neostigmine. However, this was only seen for inhibition of the thermal stimulation. This analgesic effect was reversed by applying the atropine IA, but not IT, which supports the peripheral site of action. Because of some systemic reabsorption of large doses of IA neostigmine, the %MPE in thermal and mechanical antinociception were limited and could not be evaluated in this animal model. In conclusion, our data suggest that neostigmine might be useful in inflammatory pain if administered spinally. Neostigmine displays peripheral analgesic activity in addition to central analgesic activity. The authors thank Thomas Briissel, MD, PhD, for valuable comments for the revision of this manuscript. References 1. Woolf CJ. Evidence for a central component of post-injury pain hypersensitivity. Nature 1983;306: Zimmermann M, Herdegen T. Plasticity of the nervous system at the systematic, cellular and molecular levels: a mechanism of chronic pain and hyperalgesia. Prog Brain Res 1996;110: Hood DD, Eisenach JC, Tong C, et al. Cardiorespiratory and spinal cord blood flow effects of intrathecal neostigmine methylsulfate, clonidine, and their combination in sheep. Anesthesiology 1995;82: Hood DD, Eisenach JC, Tuttle R. Phase I safety assessment of intrathecal neostigmine methylsulfate in humans. Anesthesiology 1995;82: Lauretti GR, Reis MI, Prado WA, Klamt JG. Dose response study of intrathecal morphine versus intrathecal neostigmine, their combination, or placebo for postoperative analgesia in patients undergoing anterior and posterior vaginoplasty. Anesth Analg 1996;82: Naguib M, Yaksh TL. Antinociceptive effects of spinal cholinesterase inhibition and isobolographic analysis of the interaction with mu and alpha 2 receptor systems. Anesthesiology 1994;80: Pedigo W. Determination and characterization of the antinociceptive activity of intraventricularly administered acetylcholine in mice. J Pharmacol Exp Ther 1975;193: Yaksh TL, Grafe MR, Malkmus S, et al. Studies on the safety of chronically administered intrathecal neostigmine methylsulfate in rats and dogs. Anesthesiology 1995;82: Duarte IDG, Lorenzetti BB, Ferreira SH. Peripheral analgesia and activation of nitric oxide-cyclic GMP pathway. Eur J Pharmacol 1990;186: Lauretti GR, Lima IC. The effects of intrathecal neostigmine on somatic and visceral pain: improvement by association with a peripheral anticholinergic. Anesth Analg 1996;82: Yaksh TL, Rudy TA. Analgesia mediated by a direct spinal action of narcotics. Science 1976;192: Yang LC, Marsala M, Yaksh TL. Characterization of time course of spinal amino acids, citrulline and PGE, release after carrageenan/kaolin-induced knee joint inflammation: a chronic microdialysis study. Pain 1996;67: Hargreaves K, Dubner R, Brown F, et al. A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. Pain 1988;32: Nagasaka H, Awad H, Yaksh TL. Peripheral and spinal actions of opioids in the blockade of the autonomic response evoked by compression of the inflamed knee joint. Anesthesiology 1996;85: Chaplan SR, Bach FW, Pogrel JW, et al. Quantitative assessment of tactile allodynia in the rat paw. J Neurosci Methods 1994;53: Buerkle H, Yaksh TL. Comparison of the spinal actions of the mu-opioid remifentanil with alfentanil and morphine in the rat. Anesthesiology 1996;84: Tallarida RJ, Murray RB. Manual of pharmacologic calculations with computer programs. New York: Springer Verlag, Schafer M, Imai Y, Uhl GR, Stein C. Inflammation enhances peripheral mu-opioid receptor-mediated analgesia, but not muopioid receptor transcription in dorsal root ganglia. Eur J Pharmacol 1995;279: Abram SE. Necessity for an animal model of postoperative pain. Anesthesiology 1997;86: Stem C, Yassoudris A. Peripheral morphine analgesia. Pam 1997;71: Stein C. Peripheral mechanisms of opioid analgesia. Anesth Analg 1993;76: Antonijevic I, Mousa SA, Schafer M, Stein C. Perineurial defect and peripheral opioid analgesia in inflammation. J Neurosci 1995;15: