Clonidine, an α2-adrenergic agonist prescribed for the treatment of arterial hypertension (1), is now undergoing intensive investigation as an opioid substitute and as an opioid potentiator in local and regional anesthesia (2-7) and for the treatment of acute and chronic pain (8-10). Pretreatment with oral clonidine before induction of anesthesia produces a more stable hemodynamic course (11), reduces the requirement for general anesthetic drugs during the anesthetic, and reduces the dose of morphine required to produce adequate perioperative analgesia (12). Co-administration of clonidine also prolongs the duration of action of mepivacaine administered for regional anesthesia (4), and alters the hyperpolarizing response of mammalian C fibers to repetitive stimulation (5).
As is the case for opioids, α2-adrenergic agonists have been administered intrathecally and extradurally. Although there is evidence that α2-adrenergic agonists produce their effects after intradural or extradural application by binding to α2-adrenergic receptors in the spinal cord, this topic remains the subject of investigation and debate (13). We hypothesized that the analgesia seen after neuraxial application of α2-adrenergic agonists might result (at least in part) from direct inhibition of impulse conduction in primary afferent nerve fibers. The present study tests the efficacy of the α2-adrenergic agonists clonidine and guanfacine (Figure 1) at differential inhibition of impulse conduction through Aα (large, myelinated neurons with conduction velocity ≥70 m/s responsible for somatic motor function and proprioception) and C (small, unmyelinated neurons with conduction velocity ≤2 m/s responsible for pain, reflex responses, and postganglionic sympathetic transmission) fibers in rat sciatic nerves (14).
Male rats (Charles River Laboratories, Wilmington, MA) weighing 350-500 g were anesthetized with 25 mg/kg intraperitoneal pentobarbital. Additional pentobarbital was given (25 mg/kg) as indicated by clinical signs of light anesthesia (movement or alterations in respiratory patterns) during dissection. The skin and gluteal musculature were incised and reflected laterally to expose the sciatic nerve, which was removed intact from its origin in the lumbar plexus to its bifurcation at the knee. Upon its excision, the nerve was submerged in oxygenated Tyrode solution and desheathed using jeweler's forceps, Castroviejo scissors (Roboz Surgical Instruments, Rockville, MD), and sharpened sewing needles held in pin vises. After the nerve was removed, rats were given a lethal dose of pentobarbital (50 mg/kg intraperitoneally). In some cases rats remained anesthetized with pentobarbital while one nerve was studied so that the contralateral nerve could be excised later and also studied. Our study was reviewed and approved by the animal care committee of Harvard Medical School; it was conducted within the guidelines of relevant local, state, and federal statutes.
The Tyrode solution contained (in mM): NaCI 140, NaHCO3 12, KCl 4, CaCl2 2, MgCl2 2, NaH2PO4 0.4, glucose 5, and N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (HEPES) 5. The Tyrode solution was titrated with concentrated NaOH to pH 7.4. All chemicals were reagent grade or better; all were purchased from commercial sources. Clonidine was purchased from Sigma Chemical Company (St. Louis, MO). Guanfacine was a gift from Dr. Claes Post, formerly of Astra Pain Control AB, Södertälje, Sweden.
Nerves were mounted in a four-pool sucrose gap chamber as described previously (15-17). A model S44A stimulator (Grass Instrument Company, Quincy, MA) or a model 1800 isolated generator (World Precision Instruments, New Haven, CN) delivered supramaximal stimuli through Ag/AgCl electrodes. Stimulus duration was normally either 50 or 100 μs when Aα fibers were being studied; longer durations (500-1000 μs) and three to four times higher intensities were needed to supramaximally stimulate C fibers. Compound action potentials generated in either Aα or C fibers were identified by their characteristic conduction velocities and waveforms (Figure 2). A differential electrometer (model AK47UU, MetaMetrics, Carlisle, MA) amplified (× 50) the alternating current (AC) -coupled compound action potentials. (AC-coupling is often used in electrophysiology to assist in holding a tracing in a stable position on an oscilloscope screen in settings where the direct current (DC) potential may drift. This is particularly important when high-gain amplification is required, e.g., when recording from C fibers.) Waveforms were displayed on an analog storage oscilloscope (model PM3234, Phillips Inc., The Netherlands), and photographed on Polaroid® film.
No nerve was subjected to drug testing before it had yielded constant, reproducible action potentials (<5% variation) in drug-free Tyrode solution for at least 30 min. Drugs were dissolved in Tyrode solution before application to nerves. When the nerves were re-exposed to drug-free Tyrode solution, they rapidly recovered from block. However, complete recovery from inhibition sometimes took as long as 40 min.
Tonic block was measured as the percent reduction in the compound action potential (compared with its baseline amplitude) measured at steady-state after 20 min of drug application, with infrequent (<0.1 Hz) stimulation (18). Phasic block was measured as the additional decline in the action potential during highfrequency (≥10 Hz) stimulation. Thus, phasic block was always compared with concurrently measured tonic block (16,18). Dose-response data were fit to a generalized, saturating, binding isotherm by nonlinear least squares, using the equation:
(where γ = percent inhibition of the action potential, X = mean drug concentration, KD = EC50, and N = Hill coefficient). EC50 values and Hill coefficients are reported as means ± standard deviation.
Both clonidine and guanfacine produced concentration-dependent, reversible, inhibition (tonic block) of Aα compound action potentials (Figures 3 and 4). The time from application of either clonidine or guanfacine until steady-state tonic block was achieved was comparable to that seen after application of local anesthetics (i.e., about 20 min). Guanfacine appeared slightly more potent than clonidine at producing tonic block. Due to a limited supply of guanfacine, the upper range of its concentration-response curve could not be as carefully defined as that for clonidine. Least square fits of curves to these data yield EC50 values of 1.2 ± 0.2 mM (Hill coefficient = 1.0 ± 0.1) for guanfacine and 2.0 ± 0.8 mM (Hill coefficient = 1.4 ± 0.1) for clonidine.
Both clonidine and guanfacine produced concentration-dependent, reversible inhibition (tonic block) of compound action potentials in C fibers (see Figures 3 and 4). C fibers appeared more sensitive than Aα fibers to inhibition by both guanfacine and clonidine (Figures 3 and 4); the estimated EC50, values for C fibers were 0.45 ± 0.01 mM (Hill coefficient = 1.2 ± 0.3) for clonidine and 0.17 ± 0.06 mM (Hill coefficient = 1.0 ± 0.4) for guanfacine.
Compared to clonidine, guanfacine demonstrated a greater tendency to produce phasic block of compound action potentials when the two drugs were compared at comparable degrees of tonic block and identical stimulation frequencies (Figures 5 and 6). Phasic block was assessed at either 10 or 40 Hz (stimulation frequency) and usually with moderate (20-80%) degrees of tonic block. Both the rate of onset (inversely related to the number of impulses in a train required to achieve steady-state) and the steady-state reduction in action potential amplitude (with frequent stimulation) were greater with guanfacine than clonidine (Figures 5 and 6). The relationship between drug concentration and the extent of phasic block was not monotonic; drug concentrations yielding tonic block of 40-60% usually yielded greater phasic block than drug concentrations yielding either greater or lesser extents of tonic block.
We were unable to assess phasic block of C fibers in this preparation due to the relative large stimulus requirements and small amplitude of the C fiber signal in rat sciatic nerves, and to the tendency of compound C fiber signals to increase in amplitude (in the absence of drugs) with repetitive stimulation (19,20).
Our data demonstrate that the α2-adrenergic receptor agonists clonidine and guanfacine inhibit conduction in Aα and C fibers of rat sciatic nerves studied in vitro. Unlike traditional local anesthetics, both clonidine and guanfacine tended to inhibit C fibers more potently than Aα fibers (14,21,22). (In this context, comparisons of potency refer to drug concentrations which produce a common effect. A more potent agent is effective at lower concentration than a less potent agent. Similarly, an agent that more potently inhibits C fibers than Aα fibers would inhibit C fibers at a lower concentration than would be required to inhibit Aa fibers.) We therefore speculate that some of the efficacy of α2-adrenergic agonists at producing analgesia following their regional injection may result from their “local anesthetic” actions on Aα and, especially, C fibers.
A distinct separation between concentrations of clonidine or guanfacine inhibiting motor and sensory fibers (unlike the case for local anesthetics, including bupivacaine) would be consistent with the relatively greater potency at inhibition of C fibers compared to Aα fibers which we report for clonidine and guanfacine. The absence of motor block in clinical and animal studies with clonidine argues against our explanation of the analgesia which follows epidural or spinal administration of α2-adrenergic agonists. However, side effects of epidurally administered clonidine (e.g., somnolence and hypotension) have limited its dosage in patients, preventing use of higher doses which might have produced motor block as significant Aα fiber inhibition was achieved (3,9,13,23). In humans, the concentration of clonidine in cerebrospinal fluid peaks at 4.5 μM after a large (700 μg) epidural dose (Dr. James Eisenach, Winston-Salem, NC, personal communication). We note that reduced concentrations of bupivacaine provide intense analgesia without producing perceptible motor block in patients (e.g., in parturients receiving epidural analgesia) despite the absence, in careful experimental studies, of a tendency to inhibit C fibers preferentially (24). If α2-agonists exhibit the same relationship between concentrations inhibiting nerve excitability (micromolar range) and impulse conduction (millimolar range) as do procaine and lidocaine (25-27), cerebrospinal fluid concentrations after epidural drug administration may be more than sufficient to perturb normal patterns of impulse activity.
The analgesic efficacy of oral and intravenous clonidine also argues against our explanation for α2-adrenergic agonist-induced analgesia. Is this evidence that all analgesia produced by α2-adrenergic receptor agonists results from drug binding to specific α2-adrenergic receptors, rather than from binding to and inhibition of (presumably) Na+ channels, even when drugs are applied regionally near nerves? We suspect not, but recognize that studies in which specific antagonists are present will be required to resolve the issue. We note that there are a number of compounds whose mechanism of action may change depending on their site of injection. Meperidine provides analgesia when administered systemically by binding to opioid receptors in the brain and spinal cord. When administered as a spinal anesthetic, meperidine produces sensory and motor block resembling that produced by local anesthetics (28). Meperidine may also provide greater analgesia than saline when used during intravenous regional anesthesia (29). Moreover, lidocaine, which is thought to produce regional anesthesia by inhibiting voltage-gated Na+ channels, relieves neuropathic pain and reduces the requirements for general anesthetic agents in humans when administered intravenously (30,31). Although the mechanism of these latter two effects remains unclear, low systemic concentrations of lidocaine and procaine profoundly reduce the excitability of central and peripheral neurons (26,27).
Finally, substance P antagonists, which initially were thought to produce spinal analgesia by specifically inhibiting binding of substance P to its receptor, produce reversible inhibition of impulse conduction along peripheral nerve axons similar to that produced by local anesthetics (32). In short, α2-adrenergic agonists, like other drug classes, may have multiple possible mechanisms of action, depending on the concentration and the site of administration.
Guanfacine demonstrated greater potency than clonidine at both tonic and phasic block. It may be important that at any given degree of tonic block, guanfacine produced greater phasic block than clonidine. Although a relationship between phasic block and clinical analgesia (from local anesthetics) has not been established, increased inhibition during trains of impulses remains an attractive explanation for the intense analgesia provided by “subanesthetic” doses of bupivacaine, as, for example, during labor (24). Using that argument, one would predict that guanfacine would be a better agent for clinical analgesia than clonidine. However, studies comparing the antinociceptive potency of clonidine and guanfacine have yielded conflicting data (7,33).
In most circumstances, the actions of α2-adrenergic agonists can be blocked by co-administration of an appropriate concentration of a specific antagonist. If α2-adrenergic agonists were to demonstrate potency in inhibiting voltage-gated Na+ channels, it would be of special interest to determine whether specific antagonists (e.g., phentolamine) might be used to reverse the inhibition. In the unlikely event that this were true, α2-adrenergic agonists would represent the first regional anesthetic agents which afforded the anesthetist the potential opportunity of reversing the block when it was no longer required.
Interaction between local anesthetics and/or α2-adrenergic agonists with spinal neurons may not be well-modeled by the nerve preparation we used in these experiments. Spinal neurons contain an array of receptors and ion channels that is unmatched in peripheral axons (18). Sensitivity to α2-adrenergic agonists may not be constant in different neuronal populations. Furthermore, it is likely that highly branched fibers, such as are common in the central nervous system, will be more susceptible than unbranched axons, such as those we studied here, to agents acting by inhibition of Na+ channels (34). Such regional differences in conduction safety and in susceptibility to inhibition have been demonstrated in several other preparations (35,36). Additionally, we also suspect that greater inhibition would be measured from a given concentration of α2-adrenergic agonist dissolved in spinal fluid compared to drug exposure over only a short length of axon (such as the ∼1 cm used in this study). Raymond et al. have shown in amphibian axons that as the length of exposure to local anesthetic increased from 6 to 25 mm, the blocking concentration diminished by roughly 50% (37). The importance of length of nerve compared to drug concentration required during spinal and epidural anesthesia has been considered carefully by Fink (38,39) and Raymond and Strichartz (40). We suspect that α2-adrenergic receptor agonists, much like local anesthetics, will demonstrate a length dependence of their blocking concentration, making a “local anesthetic” contribution from epidural clonidine administration a distinct possibility.
We conclude that the α2-adrenergic agonists clonidine and guanfacine inhibit conduction through Aα and C fibers in rat sciatic nerve fibers in vitro. Guanfacine appears to be more potent than clonidine at producing both tonic and phasic block. Whether these agents produce nerve block by a receptor-specific, antagonizable mechanism (e.g., by binding to α2-adrenergic receptors) remains untested as does the possibility that traditional local anesthetics bind to α-adrenergic binding sites to alter synaptic transmission in the spinal cord.
1. Schmitt H. The pharmacology of clonidine and related products. In: Gross F, ed. Antihypertensive agents. Berlin: Springer-Verlag, 1977:299-396.
2. Eisenach JC, Castro MI, Dewan DM, Rose JC. Epidural clonidine analgesia in obstetrics: Sheep studies. Anesthesiology 1989;70:51-6.
3. Castro MI, Eisenach JC. Pharmacokinetics and dynamics of intravenous, intrathecal, and epidural clonidine in sheep. Anesthesiology 1989;71:418-25.
4. Singelyn FJ, Muller G, Gouverneur JM. Adding fentanyl and clonidine to mepivacaine results in a rapid in onset and prolonged anesthesia and analgesia after brachial plexus blockade. Anesthesiology 1991;75:A653 (Abstract).
5. Gaumann D, Brunet P, Jirounek P. Effects of clonidine and lidocaine on after hyperpolarization of rabbit C-fibers. Anesth Analg 1992;74:S106 (Abstract).
6. Gaumann DM, Brunet PC, Jirounek P. Clonidine enhances the effects of lidocaine on C-fiber action potential. Anesth Analg 1992;74:719-25.
7. Smith BD, Baudendistel LJ, Gibbons JJ. Schweiss JF. A comparison of two epidural α2
-agonists, guanfacine and clonidine, in regard to duration of antinociception, and ventilatory and hemodynamic effects in goats. Anesth Analg 1992;74:712-8.
8. Eisenach JC, Rauck RL, Buzzanell C, Lysak SZ. Epidural clonidine analgesia for intractable cancer pain: Phase I. Anesthesiology 1989;71:647-52.
9. Eisenach JC, Lysak SZ, Viscomi CM. Epidural clonidine analgesia following surgery: Phase 1. Anesthesiology 1989;71:640-6.
10. Mendez R, Eisenach JC, Kashtan K. Epidural clonidine analgesia after cesarean section. Anesthesiology 1990;73:848-52.
11. Ghignone M, Calvillo O, Quintin L. Anesthesia and hypertension: The effect of clonidine on perioperative hemodynamics and isoflurane requirements. Anesthesiology 1987;67:3-10.
12. Flacke JW, Bloor BC, Flacke WE, et al. Reduced narcotic requirements by clonidine with improved hemodynamic and adrenergic stability in patients undergoing coronary bypass surgery. Anesthesiology 1987;67:11-9.
13. Eisenach JC, Tong C. Site of hemodynamic effects of intrathecal α2
-adrenergic agonists. Anesthesiology 1991;74:766-71.
14. Raymond SA, Gissen AJ. Mechanisms of differential nerve block. In: Strichartz GR, ed. Local anesthetics. Berlin: Springer-Verlag, 1987;95-164.
15. Butterworth JF IV, Moran JR, Whitesides GM, Strichartz GR. Limited nerve impulse blockade by “leashed” local anesthetics. J Med Chem 1987;30:1295-302.
16. Butterworth JF IV, Lief PA, Strichartz GR. The pH-dependent local anesthetic activity of diethylaminoethanol, a procaine metabolite. Anesthesiology 1988;68:501-6.
17. Strong PN, Smith JT, Keana JF. A convenient bioassay for detecting nanomolar concentrations of tetrodotoxin. Toxicon 1973;11:433-8.
18. Butterworth JF IV, Strichartz GR. Molecular mechanisms of local anesthesia: a review. Anesthesiology 1990;72:711-34.
19. Ritchie JM, Straub RW. The after-effects of repetitive stimulation on mammalian non-medullated fibres. J Physiol 1956;134:698-711.
20. Ritchie JM, Straub RW. The hyperpolarization which follows activity in mammalian non-medullated fibres. J. Physiol 1957;136:80-97.
21. Gissen AJ, Covino BG, Gregus J. Differential sensitivities of mammalian nerve fibers to local anesthetic agents. Anesthesiology 1980;53:467-74.
22. Gissen AJ, Covino BG, Gregus J. Differential sensitivity of fast and slow fibers in mammalian nerve: III. Effect of etidocaine and bupivacaine on fast/slow fibers. Anesth Analg 1982;61:570-5.
23. Eisenach JC. Intravenous clonidine produces hypoxemia by a peripheral alpha-2 adrenergic mechanism. J Pharmacol Exp Ther 1988;244:247-52.
24. Vanderick G, Geerinckx K, van Steenberge AL, de Muylder E. Bupivacaine 0.125% in epidural block analgesia during child-birth: Clinical evaluation. Br J Anaesth 1974;46:838-41.
25. Raymond SA, Thalhammer JG. Endogenous activity-dependent mechanisms for reducing hyperexcitability of axons: Effects of anesthetics and CO2
. In: Chalazonitis N, Gola M, eds. Inactivation of hypersensitive neurons. New York: Alan R. Liss, Inc., 1987:331-43.
26. Butterworth JF IV, Cole LR. Low concentrations of procaine and diethylaminoethanol reduce the excitability but not the action potential amplitude of hippocampal pyramidal cells. Anesth Analg 1990;71:404-10.
27. Butterworth JF IV, Cole LR. Tonic and phasic local anesthetic effects in rat hippocampal pyramidal cells. Anesthesiology 1990;73:A763.
28. Sangarlangkarn S, Klaewtanong V, Jonglerttrakool P, Khankaew V. Meperidine as a spinal anesthetic agent: A comparison with lidocaine-glucose. Anesth Analg 1987;66:235-40.
29. El-Bakry MS, El-Shafei SB, Seyam EM, et al. Use of pethidine as an intravenous regional anesthetic. M E J Anesth 1989;10:189-94.
30. Boas RA, Covino BG, Shahnarian A. Analgesic responses to i.v. lignocaine. Br J Anaesth 1982;54:501-5.
31. De Clive-Lowe SG, Desmond J. North J. Intravenous lignocaine anaesthesia. Anaesthesia 1958;13:138-46.
32. Post C, Butterworth JF IV, Strichartz GR, et al. Tachykinin antagonists have potent local anaesthetic actions. Eur J Pharmacol 1985;117:347-54.
33. Post C. Gordh T, Minor BG, et al. Antinociceptive effects and spinal cord tissue concentrations after intrathecal injection of guanfacine or clonidine into rats. Anesth Analg 1987;66:317-24.
34. Stoney SD Jr. Limitations on impulse conduction at the branch point of afferent axons in frog dorsal root ganglion. Exp Brain Res 1990;80:512-24.
35. Galindo A. Effects of procaine, pentobarbital and halothane on synaptic transmission in the central nervous system. J. Pharmacol Exp Ther 1969;169:185-95.
36. Richards CD. The action of pentobarbitone, procaine and tetrodotoxin on synaptic transmission in the olfactory cortex of the guinea-pig. Br J. Pharmacol 1982;75:639-46.
37. Raymond SA, Steffensen SC, Gugino LD, Strichartz GR. The role of length of nerve exposed to local anesthetics in impulse blocking action. Anesth Analg 1989;68:563-70.
38. Fink BR. The long and short of conduction block. Anesth Analg 1989;68:551-5.
39. Fink BR. Mechanisms of differential axial blockade in epidural and subarachnoid anesthesia. Anesthesiology 1989;70:851-8.
© 1993 International Anesthesia Research Society
40. Raymond SA, Strichartz GR. The long and short of differential block. Anesthesiology 1989;70:725-8.