Sympathetic nerve blockade during lumbar epidural and spinal anesthesia dilates resistance and capacitance vessels of the lower abdomen and lower limbs, and reduces blood pressure. It has been suggested that this vasodilation is counterbalanced by efferent sympathetic vasoconstriction above the level of blockade [1-3]. Most investigators have evaluated sympathetic nerve activity indirectly by measurement of skin temperature, skin blood flow , toe pulse amplitude , pulse-oximetric saturation , or skin conductance response . Lundin et al. [8,9] measured postganglionic impulses to the skin and muscle of the legs to demonstrate that lumbar epidural anesthesia blocks sympathetic nerve activity completely. To date, the effects of epidural anesthesia on sympathetic nerve activity have not been measured directly.
In the present study, we directly measured cardiac and renal efferent sympathetic nerve activity during epidural anesthesia to determine the difference, if any, in the sympathetic responses of the anesthetized versus unanesthetized segments. We also investigated the relationship between the baroreceptor reflex and the response of sympathetic nerve activity to anesthesia-induced hypotension.
With the approval of the Animal Care Committee of Miyazaki Medical College, we studied 13 adult cats of either sex, ranging in weight from 2.7 to 5.0 kg. Animals were divided into three groups: five cats received thoracic epidural anesthesia via epidural catheter at T-4; five received lumbar epidural anesthesia at L-1; and three received lumbar epidural anesthesia at L-1 after the carotid sinus and vagoaortic nerves were severed.
All cats were anesthetized initially by intraperitoneal injection of sodium pentobarbital, 30-40 mg/kg. A catheter was placed in the femoral vein to permit continuous infusion of lactated Ringer's solution at a rate of 10 mL [center dot] kg-1 [center dot] h-1 and sodium pentobarbital at a rate of 3-5 mg [centered dot] kg-1 [center dot] h-1 for basal anesthesia. The anesthetic doses were based on the study reported by Matsukawa and Ninomiya . The depth of anesthesia was assessed by pupillary size and circulatory response to nociceptive stimuli, and the pentobarbital infusion dose adjusted as needed. Tracheotomy was performed. Muscles were paralyzed by intermittent intravenous injection of pancuronium bromide (0.2 mg/kg) and lungs were mechanically ventilated (SN-480-5; Shinano, Tokyo, Japan) using 30% oxygen in air at a respiratory rate of 20-23 breaths/min. E-T CO (2) concentrations were maintained between 35 and 40 mm Hg (Capnomac[TM]; Datex, Helsinki, Finland). A catheter was placed in the femoral artery to permit sampling of arterial blood and measurement of arterial pressure via transducer (Uniflow; Baxter, Irvine, CA) and a carrier amplifier (AP-641G; Nihon-Kohden, Tokyo, Japan). The transducer was zeroed prior to measurement at the right atrium, and mean arterial pressure (MAP) determined by electrical averaging of the pulsatile signals. Heart rate (HR) was monitored by electrocardiography using skin needle electrodes and measured using a heart rate counter (AT600G; Nihon-Kohden). Rectal temperature was maintained at 36-37 degrees C with a heating pad.
After removing the fourth thoracic or the first lumbar spinous process in all cats, a hole 2 mm in diameter was made through the laminar arch. An epidural catheter (20 gauge; Hakko, Tokyo, Japan) was inserted 5 mm cephalad into the epidural space through the hole. With a few drops of an adhesive (Aron Alpha[R]; Toa Synthetic Chemical Industry, Tokyo, Japan), the hole was sealed and the catheter fixed in place.
Cats were fixed in the right lateral position, and a left thoracotomy was performed at the first and second intercostal spaces to expose the left stellate ganglion. The inferior cardiac nerve derived from the ganglion was identified and separated from adjacent tissues. The left kidney was exposed retroperitoneally by a left flank incision, and the renal nerve separated from the renal plexus near the renal artery. The isolated cardiac and renal nerves were mounted on bipolar silver wire electrodes and bathed in warm liquid paraffin.
Efferent signals from the isolated nerves were amplified using a two-channel biophysical preamplifier (AVB-21; Nihon-Kohden) with a bandpass filter (50-3000 Hz). The amplified signals were monitored with an oscilloscope (VC-11; Nihon-Kohden), then passed through a full-wave rectifier circuit and cumulatively integrated (EI-6010; Nihon-Kohden) using a reset time of 5 s. The presence of sympathetic nerve activity was tested by response to intravenous injection of phenylephrine hydrochloride (5 micro g/kg) or sodium nitroprusside (5 micro g/kg) [11,12]. Intravenous phenylephrine decreased cardiac and renal sympathetic activity while increasing arterial pressure. Intravenous nitroprusside increased sympathetic activity while decreasing arterial pressure (Figure 1A).
To permit study of the relationship between the baroreceptor reflex and the response of sympathetic nerve activity to hypotension during lumbar epidural anesthesia, the bilateral carotid sinus and vagoaortic nerves were severed at the midcervical level in three cats (denervated lumbar group). Severance of the carotid sinus and vagoaortic nerves was confirmed by the absence of a response to phenylephrine or nitroprusside administration (Figure 1B).
After completion of the surgical procedures, at least 1 h was allowed for stabilization of hemodynamic variables and nerve activities. After control measurements, a single dose of 1% lidocaine (0.1 mL/kg) was injected through the epidural catheter at a rate of 0.01 mL/s in all groups. This dose and rate were established by preliminary testing to achieve the desired distribution of epidural blockade and mild hypotension (i.e., a blockade of either cardiac or renal sympathetic nerve and a decrease in arterial pressure within 10%-20% of control value). Original and integrated cardiac and renal sympathetic nerve activity waveforms, ECG, and arterial pressure values were recorded on an eight-channel thermal array recorder (RT 2108; NEC San-ei, Tokyo, Japan). HR, MAP, and cardiac and renal sympathetic nerve activity were measured preinjection (control) and at 2, 5, 10, 15, 20, and 30 min after lidocaine injection. Nerve activity was quantified by the average height in five successive integrated waves. Background bioelectrical noise and, if any, afferent sympathetic nerve activity were measured at the end of the experiment after intravenous injection of trimethaphan camsylate (0.1 mg/kg) . After this, cardiac and renal sympathetic nerve activity completely disappeared (Figure 2B), and this level was defined as the noise level for postganglionic sympathetic nerve activity. The background noise level was subtracted from the initially recorded nerve activity levels. Original and integrated waveforms obtained in this experiment would show that cardiac and renal sympathetic nerve activity originated from post-ganglionic efferent fibers.
Plasma lidocaine concentrations were determined from 1.5-mL arterial blood samples obtained through the arterial catheter with a heparinized syringe before and 15 and 30 min after epidural lidocaine injection in four cats each in the thoracic and lumbar epidural groups. Samples were stored at -20 degrees C until determination of lidocaine concentrations by gas chromatography . The internal standard used was mepivacaine.
At the end of the experiment, 0.2% methylene blue solution was injected through the epidural catheter at the same volume and speed as lidocaine solution. Animals were sacrificed by administration of a large dose of sodium pentobarbital and the spine dissected to confirm epidural catheter position and determine the spread of lidocaine via the spread of methylene blue solution.
The values for the upper and lower levels of spread of methylene blue in the epidural space are presented as median +/- interquartile range. The other data are expressed as mean +/- SEM. Changes in HR, MAP, and sympathetic nerve activity were expressed as the percent change from control. Within each group, the results were analyzed by one-way analysis of variance (ANOVA) for repeated measures with individual contrasts, in comparison with the control. Comparisons across the three study groups were made by ANOVA, followed by Dunnett's post hoc procedure to compare results from the thoracic versus nondenervated lumbar epidural groups, and from the lumbar epidural groups with and without denervation. Super ANOVA (Abacus, Berkeley, CA) was used for these statistical analyses. P < 0.05 was considered statistically significant.
The epidural catheter was located in the epidural space in all cats. Methylene blue solution spread from a median of C-8 to T-6 in the thoracic epidural group, T-8 to L-3 in the lumbar epidural group, and T-7 to L-3 in the denervated lumbar group, signifying no significant differences between groups in segmental spread of lidocaine (Table 1). The mean plasma lidocaine concentration was 0.2 +/- 0.1 micro g/mL and 0.3 +/- 0.0 micro g/mL at 15 min after epidural lidocaine injection in the thoracic versus lumbar epidural groups, respectively, and 0.1 +/- 0.0 micro g/mL at 30 min in the both groups.
Control values for HR were 195 +/- 12, 199 +/- 16, and 220 +/- 5 bpm and for MAP were 122 +/- 6, 109 +/- 11, and 125 +/- 8 mm Hg in the thoracic epidural, lumbar epidural, and denervated lumbar groups, respectively. They did not differ significantly among the groups. HR decreased significantly by 12% +/- 4% at 15 min after epidural lidocaine injection in the thoracic epidural group, 7% +/- 3% at 30 min in the lumbar epidural group, and 7% +/- 2% at 30 min in the denervated lumbar group (Figure 3). The decrease in HR was significantly greater in the thoracic than the lumbar epidural group at 2, 5, and 10 min after lidocaine injection. MAP decreased significantly by 15% +/- 3% at 10 min after epidural lidocaine injection in the thoracic epidural group, 14% +/- 5% at 15 min in the lumbar epidural group, and 27% +/- 5% at 10 min in the denervated lumbar group (Figure 3). The decrease in MAP was significantly greater in the denervated versus nondenervated lumbar epidural group at 2 min after lidocaine injection.
(Figure 2A) shows typical tracings of original and integrated cardiac and renal sympathetic nerve activity and arterial pressure in the lumbar epidural group. Renal sympathetic nerve activity in the unanesthetized segments increased significantly by 53% +/- 14% at 10 min after epidural lidocaine injection in the thoracic epidural group, and cardiac sympathetic nerve activity by 75% +/- 28% at 5 min in the lumbar epidural group (Figure 4). In the denervated lumbar group, however, cardiac sympathetic activity remained unchanged after lidocaine injection, compared with the nondenervated lumbar group where there was indeed a significant difference in cardiac activity at 2 and 5 min after lidocaine injection. Sympathetic nerve activity in the targeted anesthetized segments decreased significantly by 78% +/- 4% in the thoracic epidural group, 68% +/- 7% in the lumbar epidural group, and 58% +/- 12% in the denervated lumbar group at 5 min after lidocaine injection (Figure 4).
Our results demonstrate compensatory enhancement of sympathetic nerve activity in unanesthetized segments during thoracic or lumbar epidural anesthesia in cats, but not during lumbar epidural anesthesia after severance of the carotid sinus and vagoaortic nerves. Specifically, renal sympathetic nerve activity increased during thoracic epidural anesthesia, while cardiac sympathetic nerve activity decreased, and cardiac sympathetic nerve activity increased during lumbar epidural anesthesia while renal sympathetic nerve activity decreased. After the carotid sinus and vagoaortic nerves were severed, renal sympathetic nerve activity decreased during lumbar epidural anesthesia but cardiac sympathetic nerve activity remained unchanged from the control. Hypotension was present throughout these sympathetic nerve activity responses in all three study groups.
It has been suggested that the increase in sympathetic nerve activity in unanesthetized segments during epidural anesthesia is mediated by the arterial baroreceptor reflex . Cardiopulmonary baroreceptors also contribute, by minimizing changes in arterial pressure in response to changes in blood volume . In the present study, MAP decreased significantly in all test groups, but significantly more so in the denervated lumbar epidural group in which the arterial and cardiopulmonary baroreceptor reflexes were absent. The lack of sympathetic response and hypotension in the denervated group indicates that the compensatory excitation of the sympathetic nerves during epidural anesthesia results from the baroreceptor reflex response to anesthesia-induced hypotension. Further, other sympathetically mediated compensatory mechanisms would occur. The adrenal catecholamine, renin-angiotensin, and vasopressin contribute to maintaining blood pressure during epidural anesthesia [14,16,17], although we did not examine the humoral system.
Hopf et al.  demonstrated that upper thoracic epidural anesthesia involving only a few dermatomes evokes an increase in foot skin temperature, i.e., high thoracic segmental epidural anesthesia produces a widespread sympathetic nerve block extending to the most caudal parts of the sympathetic nervous system. Accordingly, we chose a small epidural lidocaine dose (1% 0.1 mL/kg) to avoid widespread blockade and facilitate direct measurement of sympathetic nerve activity in both anesthetized and corresponding unanesthetized segments. When we used 0.2 mL/kg (1% lidocaine) as preliminary testing, both the cardiac and renal nerve activity decreased and severe hypotension appeared. Furthermore, methylene blue of the same volume extended to both the cardiac and renal innervated spinal segments. Although a large fluid volume injected in the epidural space increases in blood pressure by the increase in cerebrospinal fluid pressure , a small volume of epidural lidocaine and the slow rate of injection speed used in the present study would not increase cerebrospinal fluid and blood pressure. Therefore, we had no vehicle control group to account for the effects of increased epidural pressure. We assessed the spread of epidural lidocaine using methylene blue solution. The spread of methylene blue dose not necessarily correspond with the spread of sensory or sympathetic blockade . However, the spread of the dye would represent the diffusion of lidocaine solution as liquid in the epidural space. Median spread in all three study groups did not differ significantly, indicating that distribution of lidocaine solution did not contribute to the differences in sympathetic nerve activity response.
Valley et al.  described a rapid onset of sympathetic blockade after epidural anesthesia in humans. Using laser Doppler flowmetry to measure the vasoconstrictive reflex to deep inspiration, they demonstrated that sympathetic nerve activity decreased within 6 min after epidural administration of a 3-mL test dose of lidocaine. In the present study, we measured sympathetic nerve activity directly and found that activity decreased within only 2 min after epidural injection in the targeted anesthetized segments. The difference in the rate of sympathetic depression may be due to the differences in species or in sites and methods studied.
Intravenous lidocaine depresses sympathetic nerve activity in a dose-dependent manner . Thus, one explanation for the loss of activity in anesthetized segments may be a systemic effect of lidocaine. However, arterial baroreflex control of renal sympathetic nerve activity is preserved at plasma lidocaine concentrations less than 5 micro g/mL in alpha-chloralose-anesthetized dogs . In the present study, plasma lidocaine concentrations in all samples were too small (<0.5 micro g/mL) to demonstrate a systemic effect. The loss of sympathetic nerve activity in the anesthetized segments therefore likely reflects a true effect of nerve blockade.
Epidural injection of lidocaine at T-4 rapidly decreased HR by blocking the response of the cardiac sympathetic nerves. Injection of lidocaine at L-1 increased cardiac sympathetic nerve activity, but produced no change in HR for 15 min, followed only by a slow decline. This apparent anomaly may be explained  by a mild anesthesia-induced reduction in venous return which increases vagal activity, thereby offsetting any increase in sympathetic activity to leave HR unchanged or slightly decreased. Therefore, during epidural anesthesia, HR is controlled predominantly by parasympathetic modulation. In this experiment, control values of HR and blood pressure were not significantly different between the lumbar epidural and denervated lumbar groups, although HR was slightly different. Baroreceptor denervation resulted in the increases in HR and blood pressure, but they decreased slowly during the period for stabilization of hemodynamic variables and nerve activity. Consequently, control values did not differ significantly. Nishikawa et al.  also observed that control values for HR and blood pressure were not increased by baroreceptor denervation in nitrous oxideanesthetized cats. The mechanism of the decrease in HR in the denervated lumbar epidural group, which had neither afferent sympathetic nor parasympathetic signals, is unclear.
Ninomiya et al.  found that the effect of the baroreceptor reflex on sympathetic nerve activity was quantitatively nonuniform for functionally different organs. For example, the baroreceptor-dependent component of sympathetic nerve activity to the heart or kidney exceeds 95%, whereas that to the skin is only 5%. In the present study, we increased the likelihood of detecting a baroreceptor reflex component by focusing on the heart and kidney, and directly measuring the response of isolated cardiac and renal sympathetic nerves to epidural anesthesia and hypotension. In addition, we measured sympathetic response in both anesthetized and unanesthetized segments and included an epidural group in which the carotid sinus and vagoaortic nerves were severed, strengthening further the detection of a baroreceptor reflex effect on sympathetic nerve activity.
Alterations of baroreceptor input produced by common carotid occlusion or sinus nerve stimulation have minimal effect on renal blood flow in anesthetized animals . However, in patients with hypovolemia, compensatory excitation of renal sympathetic nerves during thoracic epidural anesthesia may induce renal vasoconstriction and reduce urinary output, thereby increasing risk of further adverse changes in fluid volume. We did not examine the effect of changes in sympathetic activity on cardiac function. However, cases of coronary vasospasm during epidural anesthesia have been reported . Although the relationship of coronary vasospasm to epidural anesthesia is unclear, it is possible that the imbalance in autonomic nervous system activity induced by epidural anesthesia may produce coronary vasospasm.
It was difficult to measure cardiac and renal sympathetic nerve activity simultaneously for more than 30 min. However, arterial baroreceptor control begins within seconds , leaving adequate time to measure the compensatory response to sympathetic blockade. Our experimental protocol was conducted using sodium pentobarbital anesthesia as basal anesthesia. Although pentobarbital affects sympathetic nerve activity [10,25], all cats received the same loading dose and almost same duration (4-5 h), thereby reducing the impact of this variable on the results. However, we cannot completely exclude an effect of the barbiturate alone, nor the possibility of pentobarbital interaction with lidocaine or with the epidural technique per se. Given the necessity for a humane protocol, this remains an unavoidable limitation of our study design.
In conclusion, sympathetic nerve activity in corresponding unanesthetized segments increases during thoracic or lumbar epidural anesthesia and hypotension. However, under conditions of sinoaortic and vagal denervation and lumbar epidural anesthesia, cardiac sympathetic nerve activity remains unchanged. These findings indicate that compensatory excitation of the sympathetic nerves results from the baroreceptor reflex response to anesthesia-induced hypotension.
The authors wish to thank Ms. Winifred von Ehrenburg for editorial assistance.
1. Bonica JJ, Berges PU, Morikawa K. Circulatory effects of peridural block. I. Effects of level of analgesia and dose of lidocaine. Anesthesiology 1970;33:619-26.
2. Sjogren S, Wright B. Circulatory changes during continuous epidural blockade. Acta Anaesthesiol Scand 1972;16:5-25.
3. Arndt JO, Hock A, Stanton-Hicks M, Stuhmeier KD. Peridural anesthesia and the distribution of blood in supine humans. Anesthesiology 1985;63:616-23.
4. Bengtsson M, Nilsson GE, Lofstrom JB. The effect of spinal analgesia on skin blood flow, evaluated by laser Doppler flowmetry. Acta Anaesthesiol Scand 1983;27:206-10.
5. Meijer J, de Lange JJ, Ros HH. Skin pulse wave monitoring during lumbar epidural and spinal anesthesia. Anesth Analg 1988;67:356-9.
6. Peduto VA, Tani R, Pani S. Pulse oximetry during lumbar epidural anesthesia: reliability of values measured at the hand and the foot. Anesth Analg 1994;78:921-4.
7. Malmqvist LA, Tryggvason B, Bengtsson M. Sympathetic blockade during extradural analgesia with mepivacaine or bupivacaine. Acta Anaesthesiol Scand 1989;33:444-9.
8. Lundin S, Wallin BG, Elam M. Intraneural recording of muscle sympathetic activity during epidural anesthesia in humans. Anesth Analg 1989;69:788-93.
9. Lundin S, Kirno K, Wallin BG, Elam M. Effects of epidural anesthesia on sympathetic nerve discharge to the skin. Acta Anaesthesiol Scand 1990;34:492-7.
10. Matsukawa K, Ninomiya I. Anesthetic effects on tonic and reflex renal sympathetic nerve activity in awake cats. Am J Physiol 1989;256:R371-8.
11. Nishikawa K, Fukuda T, Yukioka H, Fujimori M. Effects of intravenous administration of local anesthetics on the renal sympathetic nerve activity during nitrous oxide and nitrous oxide-halothane anesthesia in the cat. Acta Anaesthesiol Scand 1990;34:231-6.
12. Yoneda I, Nishizawa M, Benson KT, et al. Attenuation of arterial baroreflex control of renal sympathetic nerve activity during lidocaine infusion in alpha-chloralose-anesthetized dogs. Acta Anaesthesiol Scand 1994;38:70-4.
13. Takasaki M. Blood concentrations of lidocaine, mepivacaine and bupivacaine during caudal analgesia in children. Acta Anaesthesiol Scand 1984;28:211-4.
14. Cousins MJ, Bromage PR. Epidural neural blockade. In: Cousins MJ, Bridenbaugh PO, eds. Neural blockade in clinical anesthesia and management of pain. Philadelphia: JB Lippincott, 1988:253-360.
15. Guyton AC. Nervous regulation of the circulation, and rapid control of arterial pressure. In: Guyton AC, Hall JE, eds. Textbook of medical physiology. 9th ed. Philadelphia: WB Saunders, 1996:209-20.
16. Hopf HB, Schlaghecke R, Peters J. Sympathetic neural blockade by thoracic epidural anesthesia suppresses renin release in response to arterial hypotension. Anesthesiology 1994;80:992-9.
17. Carp H, Vadhera R, Jayaram A, Garvey D. Endogenous vasopressin and renin-angiotensin systems support blood pressure after epidural block in humans. Anesthesiology 1994;80:1000-7.
18. Hopf HB, Wei beta bach B, Peters J. High thoracic segmental epidural anesthesia diminishes sympathetic outflow to the legs, despite restriction of sensory blockade to the upper thorax. Anesthesiology 1990;73:882-9.
19. Galindo A. Hemodynamic changes in the internal carotid artery produced by total sympathetic block and carbon dioxide. Anesth Analg 1964;43:276-84.
20. Feuk U, Jakobson S, Norlen K. Central haemodynamics and regional blood flows during thoracic epidural analgesia combined with positive pressure ventilation. An experimental study in the pig. Acta Anaesthesiol Scand 1987;31:479-86.
21. Valley MA, Bourke DL, Hamill MP, Raja SN. Time course of sympathetic blockade during epidural anesthesia: laser Doppler flowmetry studies of regional skin perfusion. Anesth Analg 1993;76:289-94.
22. Ninomiya I, Irisawa H. Non-uniformity of the sympathetic nerve activity in response to baroceptor inputs. Brain Res 1975;87:313-22.
23. Kirchheim HR. Systemic arterial baroreceptor reflexes. Physiol Rev 1976;56:100-76.
24. Krantz EM, Viljoen JF, Gilbert MS. Prinzmetal's variant angina during extradural anaesthesia. Br J Anaesth 1980;52:945-9.
25. Wagner PG, Eldridge FL, Dowell RT. Anesthesia affects respiratory and sympathetic nerve activities differentially. J Auton Nerv Syst 1991;36:225-36.