Synergistic antinociceptive interaction between ... - Semantic Scholar

8 downloads 0 Views 324KB Size Report
tered spinally can reduce the minimum alveolar concentration (MAC) of inhalation agents in humans and animals11–14 the nature of their com- bined effect, that ...
British Journal of Anaesthesia 1998; 80: 800–806

Synergistic antinociceptive interaction between sevoflurane and intrathecal fentanyl in dogs D. MA, S. M. SAPSED-BYRNE, M. K. CHAKRABARTI, J. G. WHITWAM

Summary This study determined the nature of the antinociceptive interaction between sevoflurane and intrathecal fentanyl on somatosympathetic reflexes in anaesthetized dogs. Afferent A␦- and C-fibre-mediated somatosympathetic reflexes, evoked by supramaximal electrical stimulation of tibial nerves, were recorded from renal sympathetic nerves. The effect of fentanyl alone, administered intrathecally (i.t.) in incremental doses from 2 to 64 ␮g, was compared with the effect of the same doses during the administration of 1.5% sevoflurane. The mean ED50s for the depressant effect of fentanyl (i.t.) on A␦ and C reflexes were 35.6 ␮g and 14.2 ␮g while 1.5% sevoflurane, when administered alone, depressed them by 15.5% (P  0.05) and 27.5% (P  0.01) respectively. During the administration of 1.5% sevoflurane, the mean ED50s of fentanyl (i.t.) for the depression of A␦ and C reflexes were reduced by 76% and 75%, to 8.5 ␮g and 3.5 ␮g respectively. The combined antinociceptive effects of sevoflurane and intrathecal fentanyl were not additive but exhibited a high degree of synergistic interaction. (Br. J. Anaesth. 1998; 80: 800–806) Keywords: anaesthetics volatile sevoflurane; analgesics opioid fentanyl; reflexes somatosympathetic; interactions (drug) synergy; dog

In the 1970s it was shown that the intrathecal (i.t.) administration of opioids could provide pain relief.1 2 These studies contributed to our understanding of the regional role of opioid receptors, which are found not only in supraspinal centres but also, for example, in the dorsal horn of the spinal cord on the terminals of small-diameter primary afferent fibres.3–5 Recent work has suggested that the spinal cord is one of the principal sites of action of inhalation anaesthetic agents,6–9 and this has been summarized by Eger and associates.10 Although previous studies have shown that opioids administered spinally can reduce the minimum alveolar concentration (MAC) of inhalation agents in humans and animals11–14 the nature of their combined effect, that is whether additive or interactive, has not been investigated. We have tested the hypothesis that an inhalation anaesthetic and an opioid administered intrathecally act synergistically to depress nociceptive responses. Single supramaximal electrical stimuli applied to somatic nerves evoke two bursts of activity in sympa-

thetic nerves because of the activation of both afferent myelinated (group III or A␦) and unmyelinated (group IV or C) fibres.15 These reflexes are called “somatosympathetic reflexes”. They can be used to evaluate the antinociceptive effects of drugs16 and are important in the management of anaesthetized, paralysed and ventilated patients. Unsynchronized stimulation of A␦ and C fibres provides the variable nociceptive responses to surgical trauma. A controlled input by electrical stimulation of a specific nerve allows the relative quantitative effects of drugs on the A␦ and C components of nociception to be compared. In the present study, A␦- and C-fibre mediated somatosympathetic reflexes were used as nociceptive indices to examine the type of interaction between sevoflurane and fentanyl administered intrathecally in anaesthetized dogs.

Materials and methods GENERAL PROCEDURE

The study was approved by the UK Home Office (Licence PPL 90/00852). We studied 16 beagles weighing 13.5–17 kg. Background anaesthesia was induced with methohexitone (Eli Lilly, UK) 10–15 mg kg91 i.v. and maintained with 1% ␣-chloralose (Sigma, Poole, Dorset, UK) in an initial bolus dose of 30 mg kg91 i.v., followed by a continuous infusion of 15–20 mg kg91 h91 i.v. The trachea was intubated and the lungs ventilated with oxygen-enriched air (HV 2000, SLE Ltd, Croydon, UK). Muscle paralysis was maintained using suxamethonium (Wellcome, Crewe, Cheshire, UK) 10 mg i.v. every 20–30 min. The cephalic vein in the right foreleg was cannulated for the administration of drugs and the infusion of fluids. The left femoral artery was cannulated for recording arterial pressure and sampling blood. Oesophageal temperature was measured with a thermistor (Yellow Springs Instruments, OH, USA) and maintained at 37–39⬚C by a heater in the operating table. Arterial blood gas tensions and pHa were determined regularly using a blood gas analyser (Radiometer ABL 3, Copenhagen, Denmark). pHa and PaCO were kept within normal limits, that is, 7.30–7.40 and 4.8–5.6 kPa respectively, and the PaO at 22.7–26.7 kPa; all three were kept close 2

2

D. MA, MD, S. M. SAPSED-BYRNE, BSC, CBIOL-MIBIOL, M. K. CHAKRABARTI, BSC, MPHIL, J. G. WHITWAM, MB, CHB, PHD, FRCP, FRCA, Department of Anaesthetics and Intensive Care, Imperial College School of Medicine, Hammersmith Hospital, Du Cane Road, London W12 ONN. Accepted for publication: January 20, 1998. Correspondence to D. M.

Sevoflurane, fentanyl and somatosympathetic reflexes

801 thetic nerves was processed through a preamplifier (Tektronix type 122) and displayed on a dual-beam oscilloscope (Textronix type 565). Spontaneous sympathetic activity was subjected to full-wave rectification and then integrated with a time constant of 100 ms (Neurolog NL 90). The evoked somatosympathetic responses in renal sympathetic nerves were averaged (⫻16), rectified, and integrated using a neurolog system (NL 90, Digitimer, Welwyn Garden City, UK). Directly recorded and processed signals were displayed on an oscilloscope and also, when required, as “hard” copy on a thermal recorder (Gould 1602). The total electrical activity in the integrated signals of the rectified and averaged evoked sympathetic responses, and also the spontaneous sympathetic outflow during 20-s periods, was measured in arbitrary units. Data were based on the average of three sets of measurements at each point in each preparation and were expressed as a percentage of control values. EXPERIMENTAL DESIGN

Figure 1 Schematic diagram of the preparation. M : Medulla; C : Cervical; T : Thoracic; L : Lumbar; S : Sacral.

to the control values in each preparation by small changes in the FiO and ventilation, and occasional administration of small amounts of sodium bicarbonate. Heart rate and mean arterial blood pressure (MAP) were recorded continuously (Maclab 8 ADInstruments, ADInstruments Pty Ltd, Australia). Each preparation was maintained in a left lateral position. The tibial nerve in the right hindleg was exposed in the upper to middle part of the thigh. A portion was dissected free of the surrounding tissues and a short section (≈ 2 cm) was desheathed, cut distally and mounted on silver–silver chloride electrodes for stimulation. The renal sympathetic nerves were exposed retroperitoneally alongside the renal artery; a single fascicle was desheathed and cut distally, near the kidney, and mounted on silver–silver chloride electrodes for recording sympathetic activity. Prepared nerves were immersed in mineral oil pools. For intrathecal injection of drugs a 22 gauge Y-can cannula (Wallace, Essex, UK) was inserted into the subarachnoid space through the dura exposed by laminectomy at L2–L3. Sevoflurane (Abbott Laboratories Ltd, Queensborough, UK) was administered with a Blease vaporizer (Chesham, Bucks, UK). The expired concentrations were monitored with a Capnomac II (Datex, Helsinki, Finland) which was calibrated before each experiment with standard gas. Twenty min was allowed for equilibration at each concentration. The schema of the preparation is shown in fig. 1. 2

MEASUREMENTS OF EVOKED AND SPONTANEOUS SYMPATHETIC ACTIVITY

Single supramaximal electrical stimuli (frequency 0.33 Hz, intensity 30 V, duration 0.5 ms) were applied to the tibial nerves by a stimulator (Grass S88, Quincy, MA, USA) with a matching isolation unit (Grass 478A). Efferent activity in renal sympa-

The preparation was allowed to stabilize for at least 30 min after completion of surgery, before control measurements were made; these were repeated 30 min later to ensure that the preparation was stable. Three groups of data were obtained for spontaneous sympathetic activity, and A␦ and C somatosympathetic reflexes evoked by tibial nerve stimulation: (1) the effect of intrathecal (i.t.) fentanyl alone; (2) the effect of sevoflurane alone, principally to determine accurately the ED50 concentrations for both A␦ and C reflexes and also the effect of a concentration of 1.5%; (3) the effect of fentanyl (i.t.) during administration of 1.5% sevoflurane. DRUG ADMINISTRATION

In five preparations, fentanyl (i.t.) (50 ␮g ml91) (Evans, Horsham, Sussex, UK) was administered in incremental doses of 2, 2, 4, 8, and 16 ␮g, each in 0.5 ml of saline (0.9%), followed by 32 ␮g in 1 ml and 64 ␮g in 2 ml of saline with 10 min intervals between doses. In another five preparations, the ED50s for the concentrations of sevoflurane that reduced the A␦ and C reflexes to tibial nerve stimulation by 50% were determined. This was done by first obtaining control values for these reflexes in the absence of sevoflurane, which was then administered to an equilibrated endtidal concentration of 2.6%. (This had been shown in a previous study to depress the mean C response to radial nerve stimulation by 50%.17) Then, in the present study the “up-and-down” method with step changes of 0.2% was used to titrate accurately the ED50 concentration for depression of the C response to tibial nerve stimulation. The end tidal concentration of sevoflurane was then increased to 3.7% (which had been shown previously to be the ED50 for the A␦ response to radial nerve stimulation,5 and the accurate ED50 for depression of the A␦ response to tibial nerve stimulation was determined using similar methods. These values, together with the ED50 of intrathecal fentanyl alone, were used to construct isobolograms for the combined effects of the two drugs (fig. 5).

802

British Journal of Anaesthesia

Figure 2 Typical recordings of the effects of fentanyl administered intrathecally on evoked responses (ER) and spontaneous sympathetic activity (RSNA) in renal nerves. ER: lower traces : average transient of 16 responses; upper traces : rectified integral of the averaged signals. RSNA: lower traces : transient amplified signals; upper traces : rectified and integrated signals. 1 : control; 2, 3, 4, 5, 6 : recordings at total cumulative doses of fentanyl of 8, 16, 32, 64 and 128 ␮g; 7 : 5 min after naloxone (2 mg i.v.).

In another six preparations, control values were first obtained for spontaneous sympathetic activity and A␦ and C somatosympathetic reflexes evoked by tibial nerve stimulation, in the absence of both the inhalation agent and intrathecal fentanyl. Sevoflurane was then administered in a concentration of 1.5%. (A previous study had shown that sevoflurane at this concentration reduced C reflexes in response to radial nerve stimulation by around 30%,17 and also that its effects on somatosympathetic reflexes remained unchanged during observations lasting more than 3 h.) The effect of an equilibrated concentration of 1.5% sevoflurane on A␦ and C sympathetic reflexes evoked by stimulation of the tibial nerve was observed, stability of the preparation confirmed for 20 min, and new “control” data obtained showing the depressant effect of 1.5% sevoflurane (figs 3 and 4B). Fentanyl (i.t.) was then administered in doses of 2, 2, 4, 8, 16 and 32 ␮g at 10 min intervals until A␦ and C reflexes were abolished as described above for fentanyl alone. When the effect of the last dose of fentanyl had been recorded sevoflurane was withdrawn and measurements were repeated 30 min later. At the end of each experiment, naloxone (2 mg i.v.) was administered to reverse the effects of fenta-

nyl; return of the reflexes to control levels indicated that during the experiment there had been no damage to the spinal cord and no deterioration in recording. Finally i.t. injection of 2 ml plain lignocaine (2%) rapidly abolished not only the tibial nerve reflexes but also efferent spontaneous sympathetic activity, indicating that the spread of the potential effect of the injections of fentanyl encompassed both afferent fibres in the tibial nerve and the efferent sympathetic fibres from which recordings were made. ANALYSIS OF DATA

All data are expressed as mean (SD). Statistical analysis was performed by analysis of variance (ANOVA) followed, when there were significant changes, by paired t tests. Differences were considered statistically significant when the P was 0.05. A regression analysis was used to calculate the slope and 95% confidence intervals of the fentanyl (i.t.) dose–response curves using a computer program Statview + Graphic (Macintosh, Abacus Concepts, Berkeley, CA, USA); the ED25,50,75 values, together with the 95% confidence intervals for fentanyl’s depressive effects, were calculated from the dose-response curves.

Sevoflurane, fentanyl and somatosympathetic reflexes

803

Figure 3 Recordings of the effects of fentanyl administered intrathecally, during the administration of 1.5% sevoflurane, on evoked responses (ER) and spontaneous sympathetic activity (RSNA) in renal nerves. ER: lower traces : average transient of 16 responses; upper traces : rectified integral of the averaged signals. RSNA: lower traces : transient amplified signals; upper traces : rectified and integrated signals. 1 : control; 2 : 20 min after 1.5 % sevoflurane administration; 3,4,5 : recordings at total cumulative doses of 16, 32 and 64 ␮g fentanyl i.t. combined with continued sevoflurane administration (1.5%); 6 : 30 min after sevoflurane withdrawal; 7 : 5 min after naloxone (2 mg i.v.).

Results

Effect of 1.5% sevoflurane

Figures 2 and 3 show examples of typical recordings of spontaneous sympathetic activity and A␦- and C-fibre-mediated somatosympathetic reflexes, evoked by supramaximal electrical stimulation of the tibial nerves, from two preparations. They show that whereas sevoflurane causes a major increase in the effect of i.t. fentanyl on somatosympathetic reflexes, there was no measurable effect on spontaneous sympathetic activity attributable to fentanyl and sevoflurane alone or in combination.

Mean A␦ and C evoked responses were depressed to 84.5% (P0.05) and 72.5% (P0.001) of baseline by 1.5% sevoflurane (figs 3 and 4B). This confirms the results of a previous study, which showed that the differentially greater effect of sevoflurane on C reflexes compared with A␦ reflexes was comparable to that of ␮ opioids.17

SOMATOSYMPATHETIC REFLEXES: SEVOFLURANE

ED50 for A␦ and C reflexes The mean ED50s with 95% confidence intervals were 3.8 (3.5–4.4) % and 2.7 (2.4–3.0) % for A␦_and C reflexes respectively. These values are slightly higher than those for A␦ (3.7%) and C reflexes (2.6%) evoked by radial nerve stimulation,17 which could be because of the larger size of the tibial nerve and the different site of access of the nerve fibres to the spinal cord.

SOMATOSYMPATHETIC REFLEXES: FENTANYL

Both C- and A␦-evoked responses were depressed by fentanyl in a dose-dependent manner. It can be seen in figs 2 and 4A that fentanyl has a greater effect on C reflexes compared with A␦ and the ratio of the mean doses required to cause the same percentage reduction in C and A␦ responses was approximately 1:2.5. However at a total dose of 128 ␮g the C reflexes were completely abolished while the mean value of A␦ responses was reduced to 10% of control, with complete abolition in two of five preparations. Naloxone (2 mg i.v.) reversed the effects of fentanyl on both C and A␦ reflexes. Its ED25,50,75, calculated from the dose response curves, for A␦ reflexes were

804

Figure 4 The effect of i.t. fentanyl alone (A, n : 5) and i.t. fentanyl combined with sevoflurane (1.5%) (B, n : 6) on A␦- and C-fibre mediated somatosympathetic reflexes evoked by supramaximal electrical stimuli (30 V intensity, 0.5 ms duration, 0.33 Hz frequency) of tibial nerves. Results are means expressed as percentage of control values (SD). Comparison with control: *P  0.05; **P  0.001; ***P  0.001.

10.5 (8.7–16.2), 35.6 (23.4–46.5) and 120.7 (95.8– 135.8) ␮g and for C reflexes they were 4.3 (2.1–6.5), 14.2 (11.3–16.4) and 46.5 (35.7–54.3) ␮g respectively.

British Journal of Anaesthesia

Figure 5 Isobolograms for sevoflurane coadministration with i.t. fentanyl in (A) A␦-fibre-mediated and (B) C-fibre-mediated somatosympathetic reflexes evoked by supramaximal electrical stimulation of tibial nerves. The solid diagonal lines represent the additive line constructed by joining the ED50 (95% confidence intervals) dose of sevoflurane with that of fentanyl. The points on the lines (closed circles) are the predicted ED50 doses of fentanyl if its interactions with sevoflurane are additive. The points below the left of the diagonal line are the observed ED50 doses of fentanyl during sevoflurane administration. The horizontal lines are the 95% confidence intervals of the ED50 dose for fentanyl.

and 4.8 ␮g respectively instead of the observed values of 8.5 and 3.7 ␮g; also the predicted/observed dose ratios for the A␦ and C reflexes were 2.1 and 1.3 respectively. These observations indicate marked synergism between the two drugs, which is supported by the isobolograms shown in fig. 5.

COMBINATION OF 1.5% SEVOFLURANE AND FENTANYL

The effect of sevoflurane was to cause a major increase in the effectiveness of fentanyl such that C and A␦ responses were completely abolished in all preparations by mean total doses of 32 ␮g and 64 ␮g of fentanyl respectively; A␦ and C reflexes returned to 38.2 % and 27% of control respectively 30 min after the withdrawal of sevoflurane and the remaining effect of fentanyl was subsequently reversed by naloxone (2 mg i.v.) (figs 3 and 4B). In the presence of 1.5% sevoflurane, ED25,50,75 with 95% confidence intervals for fentanyl, calculated from the dose– response curves, for A␦ reflexes were 2.4 (1.3–3.3), 8.5 (6.3–10.9) and 30 (23.4–46.8) ␮g respectively, and the ED50,75 for C reflexes were 3.7 (2.6–4.9) and 11.9 (11–19.5) ␮g respectively. These values were markedly less than those for intrathecal fentanyl alone, and showed a disproportionately larger increase in the effect of fentanyl in the presence of sevoflurane than would have been predicted from the sum of their effects. If the effects of combinations of sevoflurane (1.5%) and intrathecal fentanyl on A␦ and C reflexes were additive, the theoretical predicted mean ED50 doses of fentanyl for A␦ and C reflexes would be 18

SPONTANEOUS SYMPATHETIC ACTIVITY

There were no significant changes in spontaneous sympathetic activity in any preparation throughout the present study, that is, with i.t. fentanyl, which abolished both A␦ and C reflexes, with sevoflurane alone in ED50 concentration for A␦ and C reflexes, and i.t. fentanyl used to abolish A␦ and C reflexes during the administration of 1.5% sevoflurane (figs 2 and 3). RESTING HEART RATE AND MEAN ARTERIAL PRESSURE

Fentanyl There were no significant changes in heart rate or MAP throughout the study. Combination of 1.5% sevoflurane and fentanyl Administration of 1.5% sevoflurane caused a decrease in mean heart rate from 168 beats min91 (bpm) to 144 bpm (ns), after which there was no further reduction during the administration of fentanyl. The mean MAP was reduced from 120 mm Hg, that is

Sevoflurane, fentanyl and somatosympathetic reflexes the control level, to 98 mmHg (P 0.01) by 1.5% sevoflurane without any further reduction during the administration of i.t. fentanyl. Both mean heart rate and MAP returned to within control values 30 min after sevoflurane withdrawal, that is before the administration of naloxone (2 mg i.v.).

Discussion This study shows for the first time that the combined antinociceptive effects of an inhalation anaesthetic agent and a spinally administered analgesic drug are not merely additive but highly interactive. It reveals the full extent of the synergistic interaction between the antinociceptive effects of sevoflurane and intrathecally administered fentanyl. It also confirms that sevoflurane, like fentanyl, has a greater effect on C compared with A␦ reflexes.17 The observation of the greater sensitivity of C-fibre responses to opioids is in keeping with previous work on morphine, fentanyl and alfentanil administered intrathecally,18–20 which also applies to fentanyl, alfentanil, sufentanil and morphine when administered intravenously.16 21–23 The ratio of the doses of i.t. fentanyl that depress A␦ and C reflexes, that is, 2.5:1, is of a similar order of magnitude to that for fentanyl during i.v. administration.16 Previous studies have shown that a reduction of the MAC of volatile agents is caused by opioids administered intrathecally or extradurally in both rats and humans.11–14 For example, Drasner and colleagues reported that 0.75 mg morphine injected intrathecally in rats produced a 40% reduction in the MAC of halothane,11 and more recent work has indicated that the extradural administration of morphine (4 mg) decreased the MAC of halothane by 28% in humans.14 It has also been reported that the MAC of halothane is reduced in a dose-related manner by fentanyl administered extradurally in man.13 The present study shows not only that sevoflurane causes a large augmentation of the effect of intrathecal fentanyl on A␦- and C-fibre-mediated somatosympathetic reflexes but also that their interaction is highly synergistic. Opioid receptors are found in the dorsal horn of the spinal cord on the terminals of small-diameter primary afferent fibres,3–5 and it is well known that opioids, when administered intrathecally, depress nociceptive reflexes because of their actions in the dorsal horn area.1 24 25 It has been suggested that in the spinal cord of the rat, which is the animal model most often used in experimental pain studies, 70% of the opioid receptor population are type ␮, 24% are ␦ receptors and 6% are ␬.26 However, in the spinal cord of humans ␬ receptors are the predominant type at 50%, followed by ␮, with 40%, with relatively fewer ␦ receptors.27 The dog model used in the present study has a spinal cord length of comparable scale to that of the human, with an opioid-receptor distribution also similar to that in humans,28 and is therefore a better model for clinically relevant experimental studies. Interaction between the brain and spinal cord with respect to the action of inhalation agents has been the subject of a limited amount of work, summarized by Eger and colleagues.10 The concentration required to produce amnesia and unconsciousness is approximately 25–40% of that needed to suppress purpose-

805 ful movement.6 Rampil and colleagues reported that in rats, decerebration does not decrease the anaesthetic concentrations required to produce immobility.7 8 It has also been shown that if only supraspinal structures are exposed to inhalation anaesthetic agents, the concentrations required to produce anaesthesia and reflex depression are more than double those required in the normal situation when the spinal cord is also involved.9 It would appear that the spinal cord is a principal site of action for inhalational anaesthetic agents. Synergy between the actions of drugs depends on interactive pharmacodynamic effects, and these will clearly be influenced temporally by pharmacokinetics and spatially by functional interactions at different anatomic sites.29 However, hitherto, there are no published reports on the type of pharmacodynamic interaction that occurs between an analgesic drug and an inhalation anaesthetic agent, such as fentanyl and sevoflurane. At cellular level it has been demonstrated that opioid ␮ agonists increase potassium conductance in presynaptic neuronal membranes, causing membrane hyperpolarization and reduced excitability.30 Opioids may reduce neuronal transmission by both inhibition of postsynaptic cell firing and through presynaptic reduction of neurotransmitter release.31 Volatile anaesthetic agents reduce voltage-operated calcium currents in the central nervous system so as to cause a reduction in neurotransmitter release at presynaptic terminals.32 It has also been shown that sevoflurane activates postsynaptic GABAA receptors in CA1 pyramidal neurons of rat hippocampus.33 Therefore, sevoflurane and fentanyl may activate neurotransmitter systems, which may couple the same second-messenger system or G proteins34 and therefore provide a basis for their synergistic interaction. However, for a single anaesthetic agent, the output of neuronal networks, modified by changes in the function of the cells within them, may show interactive amplification of effect to cause the anaesthetic state. The combined effects of anaesthetic and analgesic drugs will also reflect mechanisms fundamental to that state and any unified hypothesis of anaesthesia must embody the phenomenon of synergism. Fentanyl, injected intrathecally in the lumbar region, depresses somatosympathetic responses evoked by stimulation of tibial nerves and recorded in renal nerves; this action occurs with no effect on spontaneous efferent sympathetic activity or responses to radial nerve stimulation, indicating that fentanyl’s effects are localized to the afferent pathway.19 This is true for other ␮ opioids unless there is significant systemic absorption, when there are effects on all afferent pathways and higher centres that regulate sympathetic activity.20 For sevoflurane, it has been shown that compensatory baroreflexes, although depressed, retain sufficient activity at concentrations up to 3–4% to prevent effects on either spontaneous efferent sympathetic pathways or the heart rate.35 Hence neither fentanyl nor sevoflurane would be expected to cause a reduction in spontaneous sympathetic activity, and this was confirmed for both drugs when administered alone and in combination throughout this study. In conclusion, it has been shown that intrathecally administered fentanyl and inhaled sevoflurane

806 interact with a high level of synergism to depress somatosympathetic reflexes. The dorsal horns of the spinal cord are likely to be a principal site for their interaction. In the concentration and doses used, either alone or in combination, they did not modify spontaneous sympathetic activity.

References 1. Yaksh TL, Rudy TA. Analgesia mediated by a direct spinal action of narcotics. Science 1976; 192: 1357–1358. 2. Wang JK, Nauss LA, Thomas JE. Pain relief by intrathecally applied morphine in man. Anesthesiology 1979; 50: 149–151. 3. Pert CB, Kuhar MJ, Snyder SH. Autoradiographic localization of the opiate receptor in rat brain. Life Science 1975; 16: 1849–1853. 4. Atweh SF, Kuhar MJ. Autoradiographic localization of opiate receptors in rat brain. I. Spinal cord and lower medulla. Brain Research 1977; 124: 53–67. 5. Fields HL, Emson PC, Leigh BK, Gilbert RFT, Iversen LL. Multiple opiate receptor sites on primary afferent fibres. Nature 1980; 284: 351–353. 6. Dwyer R, Bennett HL, Eger EI II, Heilbron D. Effects of isoflurane and nitrous oxide in subanesthetic concentration on memory and responsiveness in volunteers. Anesthesiology 1992; 77: 888–898. 7. Rampil IJ, Mason P, Singh H. Anesthetic potency (MAC) is independent of forebrain structures in the rat. Anesthesiology 1993; 78: 707–712. 8. Rampil IJ. Anesthetic potency is not altered after hypothermic spinal cord transection in rats. Anesthesiology 1994; 80: 606–610. 9. Antognini JF, Schwartz K. Exaggerated anesthetic requirements in the preferentially anesthetized brain. Anesthesiology 1993; 79: 1244–1249. 10. Eger EI II, Koblin DD, Harris RA, Kendig JJ, Pohorille A, Halsey MJ, Trudell JR. Hypothesis: inhaled anesthetics produce immobility and amnesia by different mechanisms at different sites. Anesthesia and Analgesia 1997; 84: 915–918. 11. Drasner K, Bernards C, Ciriales R. Intrathecal morphine reduces the minimum alveolar concentration of isoflurane in the rat. Anesthesiology 1987; 67: A670. 12. Drasner K, Bernards CM, Ozanne GM. Intrathecal morphine reduces the minimum alveolar concentration of halothane in humans. Anesthesiology 1988; 69: 310–312. 13. Inagaki Y, Mashimo, Yoshiya I. Segmental analgesic effect and reduction of halothane MAC from epidural fentanyl in humans. Anesthesia and Analgesia 1992; 74: 856–864. 14. Schwieger IM, Klopfenstein CE, Forster A. Epidural morphine reduces halothane MAC in humans. Canadian Journal of Anaesthesia 1992; 39: 911–914. 15. Whitwam JG, Kidd C, Fussey IV. Responses in sympathetic nerves of the dog evoked by stimulation of somatic nerves. Brain Research 1979; 165: 219–233. 16. Niv D, Whitwam JG. Selective effect of fentanyl on group III and IV somatosympathetic reflexes. Neuropharmacology 1983; 22: 703–709. 17. Ma D, Wang C, Pac Soo CK, Chakrabarti MK, Whitwam JG. Effect of sevoflurane on spontaneous and evoked sympathetic reflexes. British Journal of Anaesthesia 1997; 78: 471P–472P.

British Journal of Anaesthesia 18. Sato A, Sato Y, Suzuki A, Swenson RS. The effects of morphine administered intrathecally on the somatosympathetic reflex discharges in anesthetized cats. Neuroscience Letters 1986; 21: 345–350. 19. Wang C, Chakrabarti MK, Galletly DC, Whitwam JG. Relative effects of intrathecal administration of fentanyl and midazolam on A␦ and C fibre reflexes. Neuropharmacology 1992; 31: 439–444. 20. Wang C, Chakrabarti MK, Whitwam JG. Effect of low and high concentrations of alfentanil administered intrathecally on A␦ and C fibre mediated somatosympathetic reflexes. British Journal of Anaesthesia 1992; 68: 503–507. 21. Swenzen GO, Whitwam JG. Selective tolerance of group III and IV somatosympathetic reflexes to the effects of alfentanil. Neuropharmacology 1986; 25: 1379–1385. 22. Swenzen GO, Chakrabarti MK, Sapsed-Byrne SM, Whitwam JG. Selective effect of sufentanil on group III (A␦) and IV (C) somatosympathetic reflexes. Acta Anaesthesiologica Scandinavica 1986; 30: 545–548. 23. Ness TJ, Gebhart GF. Differential effect of morphine and clonidine on visceral and cutaneous spinal nociceptive transmission in the rat. Journal of Neurophysiology 1989; 62: 220–230. 24. Dickenson AN, Sullivan AF. Electrophysiological studies on the effects of intrathecal morphine on nociceptive neurones in the rat dorsal horn. Pain 1986; 24: 211–222. 25. Dickenson AH, Sullivan AF, Knox R, Zajac JM, Roques BP. Opioid receptor subtypes in the rat spinal cord: electrophysiological studies with ␮- and ␦-opioid receptor agonists in the control of nociception. Brain Research 1987; 413: 36–44. 26. Besse D, Lombard MC, Besson JM. Autoradiographic distribution of ␮, ␦ and ␬ opioid binding sites in the superficial dorsal horn over the rostrocaudal axis of the rat spinal cord. Brain Research 1991; 548: 287–291. 27. Czlonkowski A, Costa T, Przewlocki R, Pasi A, Herz A. Opiate receptor binding sites in human spinal cord. Brain Research 1983; 267: 392–396. 28. Hunter JC, Birchmore B, Woodruff R, Hughes J. Kappa opioid binding sites in the dog cerebral cortex and spinal cord. Neuroscience 1989; 31: 735–743. 29. Solomon RE, Gebhart GF. Synergistic antinociceptive interactions among drugs administered to the spinal cord. Anesthesia and Analgesia 1994; 78: 1164–1172. 30. McFadzean I. The ionic mechanisms underlying opioid actions. Neuropeptides 1988; 11: 173–180. 31. North RA, Williams JT. How do opiates inhibit neurotransmitter release? Trends in Neuroscience 1983; 6: 337–339. 32. Miao N, Frazer MJ, Lynch C III. Volatile anesthetics depress Ca2; transients and glutamate release in isolated cerebral synaptosomes. Anesthesiology 1995; 83: 593–603. 33. Hirota K, Roth SH. Sevoflurane modulates GABAA and GABAB receptors in area CA1 of rat hippocampus. British Journal of Anaesthesia 1997; 78: 60–65. 34. Nicoll RA, Malenka RC, Kauer JA. Functional comparison of neurotransmitter receptor subtypes in mammalian central nervous system. Physiological Reviews 1990; 70: 513–565. 35. Ma D, Sapsed-Byrne SM, Chakrabarti MK, Whitwam JG. Effect of sevoflurane on spontaneous sympathetic activity and baroreflexes in rabbits. British Journal of Anaesthesia 1998; 80: 68–72.