Share this article on:

Direct Tramadol Application on Sciatic Nerve Inhibits Spinal Somatosensory Evoked Potentials in Rats

Tsai, Yu-Chuan MD*,; Chang, Pei-Jung MD*,; Jou, I-Ming MD

doi: 10.1097/00000539-200106000-00040
Regional Anesthesia: Research Report

We sought to determine the possible neural conduction blockade of tramadol and whether there is evidence of localized neural toxicity with spinal somatosensory evoked potential (SSEP) measurements. Male Wistar rats were used. SSEP, elicited by supramaximally stimulating the hind paw and recorded from the thoracolumbar and the first and second lumbar interspinous ligaments, was monitored. SSEPs were obtained before drug application as the pretreatment baseline and measured every 15 min after treatment for 2 h and at 60-min intervals thereafter until SSEP returned to baseline or for another 4 h. Two small strips of Gelfoam (0.6 × 1.0 cm2) soaked with the drug were placed under and over the left sciatic nerve for a 30-min period. Gelfoam was prepared with tramadol hydrochloride (Tramal; the US trade name is Ultram) 5, 2.5, and 1.25 mg, diluted if needed with saline to a total volume of 100 μL (5%, 2.5%, and 1.25%, respectively). The control data were obtained from the right side limb with normal saline by following the same method. Spinal SSEPs were measured after 48 h to detect the late neural damage. The results showed that direct tramadol application on sciatic nerves dose-dependently reduced both the amplitude and conduction velocity of SSEPs when compared with the pretreatment baseline. All SSEPs returned to pretreatment baseline, and no significant changes of SSEP between bilateral limbs were noted at the 48-h measurements. No evidence of irreversible conduction blockade indicative of local neural toxicity was seen. Pretreatment with naloxone 1 mg/kg failed to block the changes of SSEP produced by 2.5% tramadol 100 μL. We conclude that tramadol exerts a local anesthetic–type effect on peripheral nerves.

Departments of *Anesthesiology and †Orthopedics, Medical College and Hospital, National Cheng Kung University, Tainan, Taiwan

Supported, in part, by a grant from National Cheng Kung University Hospital, Tainan, Taiwan.

January 30, 2001.

Address correspondence and reprint requests to I-Ming Jou, MD, Department of Orthopedics, National Cheng Kung University Hospital, 138 Sheng-Li Rd., Tainan 704, Taiwan. Address e-mail to

Tramadol hydrochloride (TramalUltram; Ortho-McNeil, Raritan, NJ) is a centrally acting analgesic drug used mainly as an IV analgesic for the treatment of moderate to severe pain (1). Compared with epidural morphine, epidural tramadol was able to provide adequate and prolonged postoperative analgesia without serious side effects (2). The analgesic effects of epidural tramadol have been explained by receptor-mediated actions. Added to mepivacaine, tramadol prolongs the duration of axillary brachial plexus blockade (3). A specific analgesic effect of tramadol on peripheral nerves was supported. Accordingly, tramadol is not a single-mechanism analgesic. In addition to a μ-opioid agonist effect, it exerts a modulatory effect on central monoaminergic pathways, inhibiting the neuronal uptake of noradrenaline and serotonin (4–7). Tramadol was also postulated to have local anesthetic effects because it produced a significant reduction of propofol injection pain similar to lidocaine when used with 1-min retention in veins (8), and a local analgesic effect was shown if it was used in intradermal injection (9). This in vivo study was undertaken to determine the possible local anesthetic-type effects of tramadol and whether there is evidence of localized neural toxicity by observing the spinal somatosensory evoked potential (SSEP) of the sciatic nerve.

Back to Top | Article Outline


This study was approved by the research committee of National Cheng Kung University Medical Center. All animals were treated in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Male Wistar rats weighing from 450 to 500 g were used. Animals were housed at 21°C ± 0.5°C in cages, with free access to food and water. The vivarium was maintained automatically on a 12-h light/12-h dark cycle, with lights on at 7:00 am. The rats were anesthetized with pentobarbital 50 mg/kg intraperitoneally. Subsequent doses of pentobarbital were administered as necessary to maintain adequate anesthetic depth. An incision was made from the left sciatic notch to the distal thigh. The subcutaneous tissue was bluntly dissected under the skin to expose the biceps femoris muscle. The sciatic nerve was freed from its investing fascia. Identical surgery was then performed on the right side.

Eight animals were used for each treatment group. Two small strips of Gelfoam (0.6 × 1.0 cm2) (Pharmacia-Upjohn, Kalamazoo, MI) soaked with the drug were placed under and over the left sciatic nerve for 30 min as treatment. Gelfoam was prepared with tramadol hydrochloride (Tramal) 5, 2.5, and 1.25 mg diluted if needed with saline to a total volume of 100 μL (5%, 2.5%, and 1.25%, respectively). The control data were obtained from the right side limb with normal saline by following the same method.

The SSEP response was recorded as previously described (10,11). SSEP was elicited by stimulating the hind paw through needle electrodes, with the anode placed directly into both sciatic nerves just proximal to the bifurcation of the peroneal branch, and the cathode 3 mm proximal to the anode. The stimulation was applied at 0.7 impulses per second, with a square pulse of supramaximal intensity and 20 repetitions that were 200 μs long. Bipolar needle electrodes were used to record the potentials. The active electrode for the SSEP recording was placed into the interspinous ligament at thoracolumbar (T-L) and the first and second lumbar (L1-2) junctions. The reference electrode was placed subcutaneously at approximately the same level. Fifty responses were averaged and displayed on a signal averager (MEB-5504K, Neuropack Z; Nihon Koden, Tokyo, Japan). This recording system used a bandpass filter that allowed signals between 50 and 5000 Hz to pass. The analysis window for the signals was between 20 and 40 ms. The animals were kept at a physiologic temperature (38°C–39°C) by a heating blanket and lamp during the study. SSEPs were measured before and every 15 min after treatment for 2 h and at 60-min intervals thereafter until the SSEP returned to pretreatment baseline or for another 4 h. Data from the right limbs treated with normal saline were obtained as well. Amplitude and latency were analyzed to investigate the status of the neural function. The conduction velocity was measured by using the latency of two adjacent recordings (T-L and L1-2 junctions). SSEPs were measured again 48 h later to detect the late neural condition.

To detect opioid receptor–specific effects on nerve conduction, five rats were pretreated with naloxone 1 mg/kg subcutaneously 10 min before local application of 2.5% tramadol 100 μL on the left sciatic nerves and the same volume of normal saline on the right sciatic nerves, followed by SSEP measurement as above.

Data are presented as mean ± sem. A P value of <0.05 was considered statistically significant. Statistical analysis was performed comparing amplitudes, latencies, and conduction velocity after drug exposure with pretreatment baseline values or saline-treated values by using a paired Student’s t-test. Comparison between various dose treatments was accomplished with one-way analysis of variance, followed by Tukey’s test for multiple comparisons when analysis of variance indicated significant results.

Back to Top | Article Outline


In all baseline recordings, we found the wave form of SSEP to be a distinct peak in amplitude. It demonstrated a gradual increase in latency as we increased distance between the stimulation and recording sites. Amplitudes of the SSEP ranged from 43 to 270 μV (mean, 113.4 μV) at the T-L junction and from 9.5 to 25 μV (mean, 20.9 μV) at the L1/L2 junction. The latencies of the initial peak were 2.20–2.58 ms (mean, 2.42 ms) at the T-L junction and 2.00 to 2.40 ms (mean, 2.27 ms) at the L1/L2 junction. There were no significant differences between the data recorded from the experimental (left) and control (right) side and also no differences in the data variability and distribution of amplitudes and latencies during the experimental period on the control side. In addition, the range of pretreatment baseline recordings for amplitudes and latencies was consistent with those obtained from previous studies (10,11).

As Figure 1 shows, direct tramadol application to the sciatic nerve dose-dependently decreased the amplitude of the SSEP when compared with pretreatment baseline and the values of the saline treated limbs. The values of the groups treated with tramadol 2.5 mg (Group 2.5) and 5 mg (Group 5) were significantly lower than those of the groups treated with tramadol 1.25 mg (Group 1.25) and saline (Group Saline) (P < 0.05). SSEPs returned to approximately 90% and 83% of pretreatment baseline 180 min after discontinuation of tramadol treatment in Groups 1.25 and 2.5, respectively. However, no significant differences of SSEP in the bilateral limbs were noted at the 48-h measurement in any group. SSEP latencies also changed obviously in a similar pattern, although no statistical significance was shown.

Figure 1

Figure 1

The conduction velocities of the major potentials of the pretreatment baseline were 49.6 ± 5.7, 50.8 ± 5.0, and 51.4 ± 7.8 m/s in Groups 1.25, 2.5, and 5, respectively (Fig. 2). There were no significant differences among these values, which indicated the conduction of fast fibers, the major contribution of SSEP in this study. Similar to the changes in reduction of amplitude, the conduction velocity was markedly reduced and remained reduced for some time after cessation of tramadol treatment. It returned progressively to near pretreatment baseline in Groups 1.25 and 2.5 after 180 min. Furthermore, we can identify that the conduction of the fast nerve fibers was blocked completely with the resultant preservation of the slow fiber potentials, characterized by a velocity <10 m/s and an amplitude <10% of the fast fiber potential. In Group 1.25, the recordings in three of the eight animals had fulfilled the above phenomenon, i.e., the fast fibers were blocked but the slow fibers were preserved; this started 15 min after soaking with tramadol and lasted for 60 to 90 min. In Group 2.5, six of the eight animals showed the above phenomenon beginning 15 min after soaking and lasting for 30 to 135 min. One animal showed both fast and slow fibers completely blocked for 90 min. In Group 5, one animal died during the experiment; four of the seven animals showed complete blockade of fast nerve fiber conduction starting 15 min after soaking with tramadol and lasting for 255 to 375 min. Both fast and slow conduction were completely blocked in two of the seven animals for 195 to 375 min.

Figure 2

Figure 2

No significant changes of SSEP between the bilateral limbs were noted at the 48-h measurements. There were no significant changes in SSEP between Group 2.5 and the group treated with tramadol 2.5 mg along with naloxone 1 mg/kg.

Back to Top | Article Outline


The results of this study indicate that direct tramadol application results in dose-dependent inhibition of nerve conduction. SSEPs returned to near pretreatment baseline in the small-dose groups in 180 minutes. No significant changes of SSEP between bilateral limbs were noted at the 48-hour measurements in any group. Thus, no evidence of irreversible conduction blockade indicative of local neural toxicity was seen. Pretreatment with naloxone failed to block the changes of SSEP produced by tramadol. Our data support the fact that tramadol exerts a local anesthetic-type effect on peripheral nerves that is not opioid receptor related.

SSEP, used widely as an intraoperative neuromonitor during invasive spinal and cranial surgery, is a very sensitive and effective tool for the early detection of acute nerve injury (10,11). These potentials are the expression of afferent volley from the posterior root, dorsal ganglia, and dorsal column. The recording of cortical SSEPs from the scalp is frequently replaced by SSEPs from the spine, interspinous ligament, or epidural space cephalad to the area to be manipulated (11–14). Although an electrode, which is smaller than a needle, is safer and less invasive in the epidural space, needle electrodes in the interspinous ligament can detect a greater and more stable amplitude than can a needle in the spine (12). In our previous studies (10,11), spinal SSEP proved to be sensitive and reliable for recording in the lumbar interspinous area.

In our study, any possible systemic effects of tramadol could be excluded because there were no changes in SSEP in the saline-treated control limbs. The reversible depressant effect on SSEPs of the sciatic nerves in tramadol-treated limbs means that the nerve conduction is temporarily interrupted, similar to the effect produced by local anesthetics (15). The phenomenon is not related to the presence of opiate receptors in these nerves (16), because pretreatment with naloxone failed to block the suppression of SSEPs produced by direct tramadol application to the sciatic nerve. This finding is consistent with previous studies on fentanyl, meperidine, and sufentanil; this suggests that these effects are more likely mediated by a nonopioid receptor-dependent mechanism (17,18).

In a randomized, double-blinded study by Kapral et al. (3), addition of 100 milligrams of tramadol to 1% mepivacaine for axillary brachial plexus block resulted in a significant increase in the duration of the blockade without any side effects, whereas IV tramadol (100 mg) showed no effect with 1% mepivacaine for axillary brachial plexus block. The authors concluded that tramadol possesses a specific analgesic effect on peripheral nerves and may be an alternative to epinephrine or clonidine as an adjuvant to local anesthesia for an axillary block. According to our findings, the mechanism of tramadol on 1% mepivacaine for nerve block may be not only epinephrine or clonidine related. Some degree of conduction blockade caused by tramadol should be similar to that produced by local anesthetics. We suggest that tramadol may provide a direct nerve blockade on peripheral nerves when used with mepivacaine for nerve blocks.

Conduction block in nerves can be produced by a variety of chemical compounds, including peptides, alcohols, barbiturates, anticonvulsants, and narcotics (19). The nerve conduction block effects of opioids have been demonstrated in both clinical and animal studies (15,17,20). However, previous studies concerning the effects of opioids on nerve conduction have made use of in vitro preparations either with peripheral nerves or dorsal roots (17–21). Compared with previous in vitro experiments, which detected the drug effects on nerve conduction, this in vivo study provided long-term follow up and late outcome monitoring to determine whether neural toxicity did occur. Although the SSEP technique suffices for noninvasive clinical recording of nerve, root, or cord conduction, it is by no means the equal of direct-contact recording of the compound nerve action potential. The extraskeletal technique would be justified were the entire experiment noninvasive; it was not, because the nerve was exposed with a long posterior thigh incision. Hence, direct application of recording electrodes on the nerve would have been better yet. The experiment could have been performed entirely noninvasively by using the percutaneous rodent sciatic nerve block method of Stichartz’s laboratory (22,23). However, exposure of the sciatic nerve in this study was performed to ensure not only that the drug was located precisely, but also that the sciatic nerve was covered in whole layers by two strips of Gelfoam, which was soaked with the drug in various doses.

Opioids such as morphine, codeine, meperidine, dextropropoxyphene, fentanyl, and sufentanil have local anesthetic effects in in vitro studies (17–21). Tramadol is a centrally acting analgesic, structurally related to codeine; both drugs have a relatively low affinity for μ-opioid receptors. Unlike those of codeine, the analgesic effects of tramadol are mediated via indirect modulation of the central monoaminergic inhibitory pain pathways as well. Clinically, several studies have shown that tramadol may have local anesthetic-type properties (8,9), but no study has described its effects on nerve conduction. In this study, conduction block in a dose-related manner on sciatic nerves was produced by direct application of tramadol. However, classification as a local anesthetic requires not only clinical local anesthesia, but also characteristic blockade of voltage-dependent Na+ channels (24). In this study, we did not assess whether tramadol has this property.

The drug was applied extraneurally in our study. It is possible that differential fiber blockade resulted from radial drug penetration where the largest drug concentration was at the surface and where the smallest concentration was within the nerve center. To best avoid this situation, two strips of Gelfoam soaked with tramadol were placed under and over the sciatic nerve for 30 minutes. It was shown that the greater depressant effect on the conduction velocity of fast fibers (around 50 m/s, compatible with A fibers), as compared with that of slow fibers (<10 m/s, compatible with C fibers), was consistent with previous reports. This indicated that A fibers might be more sensitive than C fibers for conduction blockade when they are exposed to local anesthetics (25,26). In the in vitro study by Gissen et al. (17), fentanyl and sufentanil showed similar effects.

Another aspect of this study concerned the possible neurotoxicity of tramadol. Although tramadol has been administered into the epidural space of humans, little information is available concerning the possible early or late neurotoxicity of epidurally administered tramadol. The SSEPs in the two small-dose groups returned to near baseline in 180 minutes, and there were no significant differences between the SSEPs in the tramadol-treated and saline-treated limbs at the 48-hour measurement in the three tramadol-treated groups. No abnormal motor impairment was found in any rat. These findings indicate that direct tramadol application on the sciatic nerve did not produce deleterious neurological effects.

In conclusion, direct application of tramadol on the sciatic nerve inhibits SSEP in a dose-dependent and reversible manner that is not affected by naloxone. Our data suggest that tramadol exerts a local anesthetic type effect on peripheral nerves.

The authors thank Miss Ya-Ting Tsai for her excellent technical assistance.

Back to Top | Article Outline


1. Bamigbade TA, Langford RML. Tramadol hydrochloride: an overview of current use. Hosp Med 1998; 59: 373–6.
2. Baraka A, Jabbour S, Ghabash M, et al. A comparison of epidural tramadol and epidural morphine for postoperative analgesia. Can J Anaesth 1993; 40: 308–13.
3. Kapral S, Gollmann G, Waltl B, et al. Tramadol added to mepivacaine prolongs the duration of an axillary brachial plexus blockade. Anesth Analg 1999; 88: 853–6.
4. Driessen B, Reimann W. Interaction of the central analgesic tramadol with the uptake and release of 5-hydroytrytamine in the rat brain in vitro. Br J Pharmacol 1992; 105: 147–51.
5. Driessen B, Reimann W, Giertz H. Effects of the central analgesic tramadol on the uptake and release of noradrenaline and dopamine in vitro. Br J Pharmacol 1993; 108: 806–11.
6. Giusti P, Buriani A, Cima L, Lipartiti M. Effect of acute and chronic tramadol on [3H]-5-HT uptake in rat cortical synaptosomes. Br J Pharmacol 1997; 122: 302–6.
7. Raffa RB, Friderichs E, Reimann W, et al. Opioid and non-opioid components independently contribute to the mechanisms of action of tramadol, an atypical opioid analgesic. J Pharmacol Exp Ther 1992; 260: 275–85.
8. Pang WW, Huang PY, Chang DP, Huang MH. The peripheral analgesic effect of tramadol in reducing propofol injection pain: a comparison with lidocaine. Reg Anesth Pain Med 1999; 24: 246–9.
9. Pang WW, Mok MS, Chang DP, Huang MH. Local anesthetic effect of tramadol, metoclopramide, and lidocaine following intradermal injection. Reg Anesth Pain Med 1998; 23: 580–3.
10. Jou IM, Lai KA. Neuromonitoring of an experimental model of clip compression on the spinal nerve root to characterize acute nerve root injury. Spine 1998; 23: 932–40.
11. Jou IM, Lai KA, Shen CL, Yamano Y. Changes in conduction, blood flow, histology, and neurological status following acute nerve-stretch injury induced by femoral lengthening. J Orthop Res 2000; 18: 149–55.
12. Jones SJ, Edgar MA, Ransford AO, Thomas NP. A system for the electrophysiological monitoring of the spinal cord during operations for scoliosis. J Bone Joint Surg Br 1983; 65: 134–9.
13. Matsuzaki H, Tokuhashi Y, Matsumoto F, et al. Problems and solutions of pedicle screw plate fixation of lumbar spine. Spine 1990; 15: 1159–65.
14. Shimoji K, Kano T, Higashi H, et al. Evoked spinal electrograms recorded from epidural space in man. J Appl Physiol 1972; 33: 468–71.
15. Fernandez-Galinski SM, Monells J, Espadaler JM, et al. Effects of subarachnoid lidocaine, meperidine and fentanyl on somatosensory and motor evoked responses in awake humans. Acta Anaesthesiol Scand 1996; 40: 39–46.
16. Fields HL, Einson PC, Keigh BK, et al. Multiple opiate receptor sites on primary afferent fibers. Nature 1980; 284: 351–3.
17. Gissen AJ, Gugino LD, Datta S, et al. Effects of fentanyl and sufentanil on peripheral mammalian nerves. Anesth Analg 1987; 66: 1272–6.
18. Power I, Brown DT, Wildsmith JAW. The effect of fentanyl, meperidine and diamorphine on nerve conduction in vitro. Reg Anesth 1991; 16: 204–8.
19. Staiman A, Seeman P. The impulse-blocking concentrations of anesthetics, alcohols, anticonvulsants, barbiturates, and narcotics on phrenic and sciatic nerves. Can J Physiol Pharmacol 1974; 52: 535–50.
20. Jaffe RA, Rowe MA. A comparison of the local anesthetic effects of meperidine, fentanyl, and sufentanil on dorsal root axons. Anesth Analg 1996; 83: 776–81.
21. Brodin P, Skoglund LA. Dose-response inhibition of rat compound nerve action potential by dextropoxyphene and codeine compared to morphine and cocaine in vitro. Gen Pharmacol 1990; 21: 551–3.
22. Kohane DS, Lu NT, Gokgol-Kline AC, et al. The local anesthetic properties and toxicity of saxitonin homologues for rat sciatic nerve block in vivo. Reg Anesth Pain Med 2000; 25: 52–9.
23. Popitz-Bergez FA, Leeson S, Strichartz GR, Thalhammer JG. Relation between functional deficit and intraneural local anesthetic during peripheral nerve block: a study in the rat sciatic nerve. Anesthesiology 1995; 83: 583–92.
24. Wagner LE II, Eaton M, Sabnis SS, Gingrich KJ. Meperidine and lidocaine block of recombinant voltage-dependent Na+ channels: evidence that meperidine is a local anesthetic. Anesthesiology 1999; 91: 1481–90.
25. Gissen AJ, Covino BG, Gregus J. Differential sensitivity of mammalian nerve fibers to local anesthetic agents. Anesthesiology 1980; 53: 467–74.
26. Huang JH, Thalhammer JG, Raymond SA, et al. Susceptibility to lidocaine of impulses in different somatosensory afferent fibers of rat sciatic nerve. J Pharmacol Exp Ther 1997; 282: 802–11.
© 2001 International Anesthesia Research Society