N a u n y n - S c h m i e d e b e r g ' s Arch P h a r m a c o l (1995) 3 5 1 : 6 1 8 - 6 2 3
© Springer-Verlag 1995
John L. Plummer • Patricia L. Cmielewski . Stephen Tallents • Pauline De La M. Hall • John Odontiadis - Geoffrey K. Gourlay Harry Owen
Development of tolerance to antinociceptive effects of an intrathecal morphine/clonidine combination in rats Received: 29 September 1994/Accepted: 25 J a n u a r y 1995
Previous animal studies have shown the antinociceptive effects of intrathecal clonidine and intrathecal morphine to be synergistic. This study investigated the intrathecal administration of multiple doses of this drug combination to examine the rate of development of tolerance and to determine whether there was any toxic effect on the spinal cord. Rats with indwelling intrathecal catheters were given saline, morphine (2.5-7.5 lag), clonidine (17.5 lag), or clonidine (17.5 lag) plus morphine (1 lag) intrathecally twice daily for 4½ days (total of 9 doses). Hot plate and tail flick tests were conducted after the first, fifth and ninth doses. After the ninth dose animals were killed and their spinal cords were removed for histological examination. Tolerance developed to the antinociceptive effects of the drug combination, but at a slower rate than to morphine alone. No evidence of toxicity or injury to the spinal cord was observed other than changes which could be ascribed to the presence of the catheter. Abstract
Key words Clonidine
- Morphine
• Intrathecal
•
Spinal cord • Opioid tolerance • Alpha 2 adrenoceptor
Introduction While pain can be adequately controlled by oral opioids and adjuvant drugs in over 70% of cancer patients with pain (Ventafridda et al. 1987), additional treatment modalities are necessary for the remainder of
patients. One of these modalities is spinal administration of opioids, usually morphine (Cousins et al. 1988). With increasing intensity of the pain stimulus and the development of tolerance, the dose of morphine needs to be increased. However, there is a limit to the extent to which the dose can be increased while still obtaining increased pain relief. At high doses, spinal morphine can produce abnormal movements and hyperaesthesia (Hogan et al. 1991; DeConno et al. 1991). Thus there is a need for alternative treatments enabling maintenance of pain control while reducing the dose of spinal morphine. One possibility is the addition of non-opioid drugs. It has previously been shown in animals that antinociceptive effects of intrathecal morphine and intrathecal clonidine are synergistic (Wilcox et al. 1987; Loomis et al. 1988; Ossipov et al. 1989; Plummer et al. 1992). However, clinical application of this synergism in cancer pain would require that it be maintained over multiple doses. Tolerance develops to antinociceptive effects of both clonidine and morphine (Solomon and Gebhart, 1988; Takano and Yaksh, 1993); it is possible that rapid development of tolerance could occur when both are given together. Furthermore, histological and toxicological studies should be conducted in animals prior to application of the combination in humans (Kitahata 1989; Yaksh and Collins 1989). The present study was conducted to compare the rate of development of tolerance to antinociceptive effects of intrathecal clonidine plus morphine with that to morphine alone, and to determine whether the clonidine-morphine combination leads to histologically discernible changes in the spinal cord following repeated doses.
J.L. Plummer ( ~ ) • P. L. Cmielewski . J. Odontiadis G. K. Gourlay ' H. Owen Pain Management Unit, Department of Anaesthesia and Intensive Care, Flinders Medical Centre, Bedford Park. South Australia. 5042
Materials and methods
S. Tallents - P. De la M. Hall Department Histopathology, Flinders Medical Centre, Bedford Park. South Australia. 5042
Ethical approval. This study was approved by the institutional Animal Ethics Committee.
619
Animals. Male Porton rats, mean weight 422 g (SD = 43) were used. Intrathecal cannulae were placed two weeks prior to drug administration as previously described (Plummer et al. 1990). Drug administration. Rats were randomly assigned to receive by intrathecal administration saline (n = 8), clonidine hydrochloride 17.5 gg (n = 8), clonidine hydrochloride 17.5 pg plus morphine hydrochloride 1 pg (n = 25), or morphine hydrochloride alone in a dose from 2.5 to 7.5 gg (n = 37). The range of morphine doses was chosen to ensure responses covered the same range as those caused by the morphine/clonidine combination. Drugs were given in a volume of 25 gl and flushed in with 5 gl saline, twice daily (approximately 0900 and 1600 h) Monday to Thursday, and on Friday morning. Thus, each rat received 9 doses of the assigned drug. A further 10 rats were not cannulated, did not receive any drugs, and were not subjected to hot plate or tail flick tests, but were housed in the same room as the experimental animals. These rats served as "healthy controls" for histological studies. Drug solutions were freshly prepared each Monday morning, but to ensure constancy of dose the same solutions were used during the whole week, being stored at 4°C when not in use. The first 20 solutions of each drug were assayed for clonidine (Wilczynska-Wojtulewicz and Sadlej-Sosnowska 1986) and/or morphine (McLean et al. 1990) by high performance liquid chromatography on the Monday and Friday to confirm stability of the drugs. Hot plate and tail flick tests were carried out following the Monday (day 1), Wednesday (day 3) and Friday (day 5) morning doses. Animals were killed by overdose of pentobarbitone at approximately 1400 h on the Friday afternoon for removal of their spinal cords. Nociceptive tests. The investigator (PLC) conducting the nociceptive tests was unaware which treatment animals had received. Prior to drug administration on days 1, 3 and 5, duplicate pre-drug hot plate (55 ° C, 25 s cut-off) and tail flick (5 s cut-off) latencies were measured. Following drug administration, latencies were measured at 15, 30, 60 and 90 min and then every 30 min, until the latency had returned to within 25 % of the mean pre-drug latency, but always for at least 90 min. Prior studies indicated that the drugs used did not impair motor performance sufficiently to affect response in the hot plate or tail flick tests (Plummer et al. 1991; Plummer et al. 1992). Histological examination. After animals were killed, Zamboni's fixative (0.1-1.0 ml) was injected via the catheter and the spinal column was removed and immersed in Zamboni's fixative. This was kept at 4 8 ° C for 48 72 h, following which the spinal column was rinsed with water and immersed in neutral EDTA solution (Bancroft and Stevens 1982) until decalcified (2-4 weeks). The cords were removed and stored in 70% alcohol until ready for routine paraffin processing, cutting and staining with haematoxylin and eosin and Luxol fast blue/van Gieson for myelin. Three sections were prepared from each cord; one from above the level of the catheter tip (top section, approximately T10), one at the level of the catheter tip (middle section, approximately T12) and one below the level of the catheter tip (lower section, approximately T13). Coded sections were examined by a histopathologist (ST). Each section was assessed for demyelination, neuronal injury and inflammation. The severity of inflammation was graded as absent, mild, moderate or marked, scored as 0, 1, 2 or 3 respectively. Data analysis. Pre-drug hot plate and tail flick latencies were examined for trend by two factor (assigned drug x day) analysis of variance with repeated measures on one factor (day). Post-drug hot plate and tail flick latencies were converted to percent maximum possible effect (% MPE) using the equation (Yaksh et al. 1986): post-drug latency % MPE =
cut-off time -
mean pre-drug latency
mean pre-drug latency
x 100
where mean pre-drug latency is the mean of the latencies measured on the same day prior to drug administration, and cut-off time is
5 s for the tail flick and 25 s for the hot plate test. The area under the % M P E vs time curve (AUC) was calculated by the trapezoidal rule from the time of drug administration until either 90 min after drug administration or the first time at which test latency returned to within 25 % of the mean pre-drug latency, whichever was the larger. As indices of tolerance, we took the ratios of the day 3 and day 5 AUC to the AUC on day 1. These ratios may be interpreted as the fraction of initial activity remaining at day 3 and day 5. To compare these ratios for morphine plus clonidine and morphine only, we used a rank-based randomized blocks analysis of variance. The blocking was necessary due to- the variability of responses on day 1, and an apparent relationship between rate of decline of response and response on day 1. Responses were grouped into three approximately equal sized blocks on the basis of the day 1 AUC. Within each block, the Mann-Whitney U statistic, its expected value under the null hypothesis of no treatment difference, and its variance were calculated. These were summed over the three blocks to provide a single test of the null hypothesis (Lehmann 1975). This approach reduces the influence of variability in response on day 1, which would otherwise reduce the power to detect a difference between treatments. The inflammation severity scores were compared by Friedman's test (within-animal comparisons) or Mann-Whitney U tests (among animal comparisons). Associations were examined by Spearman's rank correlation coefficient.
Results Assay of the drug solutions confirmed stability of the morphine and clonidine over the five day study period. On the Friday, morphine solutions contained a mean of 100.9% (95% confidence interval 98.7-103.1%, n = 20) and clonidine solutions a mean of 100.2% (98.4-102.0%, n = 20) of the concentration initially present on the Monday. During the 5 days of drug administration cannulated rats lost an average of 8 g (SD = 15) of body weight, but the loss was similar in the different treatment groups. During the corresponding period, the "healthy controls", without cannulae gained a mean of 12 g (SD = 14). No animals showed any obvious signs of withdrawal. None developed any thermal injuries from the nociceptive tests. Pre-drug hot plate latencies were significantly (P < 0.05) lower on days 3 and 5 than on day 1. The difference was, however, quantitatively small and did not differ over the different assigned drugs. Pre-drug tail flick latencies did not differ significantly over the period of the study (Table 1). Clonidine given alone had only a weak antinociceptive effect in the hot plate test at the dose we used. The response diminished as repeated doses were given (Fig. 1). Saline had little effect on hot plate latencies and response changed little over five days. The pattern of response was similar in the tail flick test except that clonidine (mean response on day 1 = 9500, SD = 6400) had a greater effect in this test relative to the other drugs. In the tail flick test, response on day 1 was only weakly associated with morphine dose over the dose range used (mean response 10700 at 2.5 btg morphine, 13500 at 7.5 btg morphine). In the hot plate test, the
620 Table 1 Mean pre-drug latencies in the hot plate and tail flick tests
Mean pre-drug tail flick latency (s)
Mean pre-drug hot plate latency (s) Assigned treatment
Day 1
Day 3
Day 5
Day 1
Day 3
Day 5
Saline
11.4 (1.3)
10.5 (1.4)
10.6 (2.1)
2.2 (0.5)
2.2 (0.2)
2.3 (0.2)
Morphine
11.2 (1.8)
10.9 (1.8)
11.1 (1.7)
2.1 (0.2)
2.1 (0.2)
2.0 (0.2)
Clonidine
12.9 (1.1)
11.7 (1.7)
11.9 (2.3)
2.1 (0.3)
2.0 (0.2)
2.1 (0.3)
Morphine plus Clonidine
11.7 (1.7)
10.7 (2.2)
10.8 (2.4)
2.1 (0.3)
2.1 (0.2)
2.0 (0.2)
Overall
11.6 (1.7)
10.9" (1.9)
11.1" (2.0)
2.1 (0.3)
2.1 (0.2)
2.1 (0.2)
Data are mean (SD) for n = 8 37 rats. * P < 0.05 compared to pre-drug latency on day 1.
combination on day 1 was considerable (AUC range of 2400-17200 in the hot plate and 6400-23000 in the tail I flick test) and so for purposes of comparison of develo opment of tolerance, all data for morphine alone where 10 the day 1 response fell within these ranges were used. This led to the exclusion of data from three rats in the 0 --1 hot plate test and one rat in the tail flick test. ,=C Table 2 shows the ratios day 3/day 1 and day 5/day 1 response for the hot plate test for rats receiving morphine only or morphine plus clonidine, stratified a. according to the level of response on day 1. At day 3, rats receiving the combination retained a greater pro0 -r portion of their initial response than those receiving morphine alone, but this difference was not statistically 0 1 2 3 4 5 significant (P > 0.05). At day 5, however, rats receiving the combination did retain significantly more response Days (P < 0.05). In the case of the tail flick test (Table 3), retention of response was again similar at day 3 for the Fig. 1 Response in the hot plate test (mean, SD for n = 8 25 rats) on the first, third and fifth days of multiple intrathecal dosing with two treatments (P > 0.05) but on day 5 the mormorphine (7.5 ~tg), morphine (1 lag) plus clonidine (17.5 lag), phine/clonidine combination retained a greater fracclonidine (17.5 lag) or saline. Only the 7.5 lag dose is shown for morphine as this gave a mean response on day 1 similar to that of tion of the day 1 response compared to morphine (P < 0.05). the morphine/clonidine combination None of the spinal cords showed evidence of neuronal dropout or demyelination suggestive of toxic inmean response was low at the 2.5 ~tg morphine dose jury. Spinal cords of untreated rats ("healthy controls") (mean response = 3250) but did not vary greatly over had essentially no inflammatory change; only the lower the dose range 3 7.5 ~tg (mean responses 8150 9800). section of one cord showed scattered mononuclear The influence of morphine dose on response appeared cells; this appearance was considered to be within norto be largely overshadowed by inter-individual vari- mal limits. Cords of animals which had been cannulated had a "connective tissue tunnel" in the top and ation in response among rats. The combination of morphine (1 ~tg) and clonidine middle sections, where the catheter had been (Fig. 2). (17.5 ~tg) had a marked antinociceptive effect in both Rats which had been treated with saline showed signifithe hot plate and the tail flick test. On day 1 this effect cantly more inflammation at all three levels of the cord was similar on average to that of 7.5 pg morphine in (median score = 2 at each level) than the healthy conthe hot plate (mean AUC = 8150 for combination, trols (median score = 0 at each level) (P < 0.01 at each 9800 for morphine 7.5 gg) and tail flick (AUC = 15500 level). Among the rats which had been cannulated, and 13500, respectively) tests. However, for both there were no significant differences between those nociceptive tests the variability of responses to the which had received morphine (all median scores = 2) 15
|
I
I
!
I
621 Table 2 Fraction of initial antinociceptive activity remaining after 3 or 5 days treatment with intrathecal morphine or intrathecal morphine + elonidine, hot plate test
Fraction of activity remaining Day 1 response AUC = Day 3b Morphine Morphine plus Clonidine Day 5c Morphine Morphine plus Clonidine
2000-6000 a
6000-10000
10000-18000
0.46 a (10,0.29 0.89) 0.44 (9,0.28 1.16)
0.46 (14,0.33-0.62) 0.52 (7,0.41 1.00)
0.48 (10,0.43-0.55) 0.60 (9,0.37 0.86)
0.60 (10, 0.38 0.89) 0.91 (9, 0.63-1.48)
0.35 (14, 0.07-0.50) 0.52 (7, 0.05 0.67)
0.21 (10, 0.15-0.31) 0.31 (9, 0.20-0.42)
a Rats were grouped into three approximately equal sized blocks on the basis of the day I AUC bNo significant difference on day 3 between fraction of activity remaining for morphine and morphine + clonidine, P > 0.05. cOn day 5 morphine + clonidine group had significantly greater fraction of activity remaining than morphine alone, P < 0.05. dAll entries are median ratio of response on day 3 or day 5 to response on day 1 (n, interquartile range). Table 3 Fraction of initial antinociceptive activity remaining after 3 or 5 clays treatment with intrathecal morphine or intrathecal morphine + clonidine, tail flick test Fraction of activity remaining Day 1 response AUC = Day 3b Morphine Morphine plus Clonidine Day 5° Morphine Morphine plus Clonidine
4000-11000 a 11000 16000
16000-23000
0.82 a 0.67 (14, 0.54-1.01)(18, 0.49-0.81) 0.52 0.68 (5, 0.31-1.39) (5, 0.51 0.85)
0.58 (4, 0.43-0.84) 0.58 (15, 0.33 0.62)
0.40 (14, 0.2~0.44) 1.31 (5, 0.50-1.72)
0.28 (4, 0.10-0.54) 0.43 (15, 0.27-0.69)
0.42 (18, 0.19 0.65) 0.54 (5, 0.35-0.67)
Rats were grouped into three approximately equal sized blocks on the basis of the day 1 AUC b No significant difference on day 3 between fraction of activity remaining for morphine and morphine + clonidine, P > 0.05. cOn day 5 morphine + clonidine group had significantly greater fraction of activity remaining than morphine alone, P < 0.05. aAll entries are median ratio of response on day 3 or day 5 to response on day 1 (n, interquartile range)
Fig. 2 Transverse section of spinal cord from a cannulated rat at the level of T10, showing a connective tissue tunnel (arrow).
Discussion and those which had not (median scores = 1, 2, 2 at top, middle and lower sections, respectively) or between those which had received clonidine (all median scores = 2) and those which had not (median scores 2, 1, 2 at top, middle and lower sections, respectively). There was also no significant correlation between dose of morphine received and inflammation score (r~ = 0.13, 0.16 and 0.25 for top, middle and lower sections, respectively, all P > 0.05). Inflammation scores did not differ significantly among the three levels of the cord examined (P = 0.83).
The results described here are in agreement with earlier findings by ourselves (Plummer et al. 1992) as well as by others that antinociceptive effects of intrathecal morphine and clonidine are supraadditive, in that relatively large doses of morphine alone were required to produce an effect similar to that of a small dose of morphine combined with a dose of clonidine which, given alone, had only a weak antinociceptive effect. Furthermore, reduction in the antinociceptive effect when 9 doses were given over 5 days was less with the combination compared to morphine at doses giving similar initial responses. This observation is consistent
622
with results of Yaksh and Reddy (1981), who found that daily intrathecal doses of morphine alone or ST-91 (2-(2,6-diethylphenylamino)-2-imidazoline, like clonidine, an a2-adrenoceptor agonist) alone to macaque monkeys resulted in tolerance developing to the antinociceptive effects, whereas when the drugs were combined the effect was maintained. Antinociceptive effects of spinal ~2-adrenoceptor agonists are believed not to involve opioid receptors (Yaksh and Reddy 1981; Loomis et al. 1988). By addition of an ~2-adrenoceptor agonist to morphine, antinociceptive effects can be maintained with smaller doses of morphine and less activation of opioid receptor systems. Yaksh and Reddy (1981) suggested that this lesser activation may be the reason for retarded development of tolerance. No significant difference in retention of effect was found between morphine and the morphine/clonidine on the third day of treatment. While this could reflect different time-profiles of development of tolerance to the treatments, it may simply be that sensitivity of our method to detect differences is low when only a slight degree of tolerance has developed. On day 3 most of the animals still retained over 40% of their day 1 response in the hot plate test, and over 50% in the tail flick test (Tables 2,3). No evidence of histological changes which could be ascribed to morphine, clonidine or their combination was found on examination of the spinal cords. "Connective tissue tunnels" which presumably surrounded the catheter were observed in rats which had been cannulated. These tunnels have also been reported in humans (Sj6berg et al. 1992). Mild inflammatory changes were observed which could be ascribed to the presence of the cannula. This lack of toxicity is in agreement with previous reports. No evidence of toxicity was observed after multiple intrathecal doses of clonidine in rats (Gordh et al. 1986), dogs (Gordh et al. 1984), sheep (Eisenach et al. 1987) or man (Coombs et al. 1986). There is also no clear evidence for neurotoxicity of intrathecal morphine, despite its extensive use. In one study, 2 of 7 patients who received prolonged intrathecal morphine therapy had posterior column degeneration, but these patients had not received the highest morphine doses and possible causes other than morphine were present (Coombs et al. 1985). A separate study of 15 patients who had been treated with intrathecal morphine and bupivacaine revealed no pathological changes which could be ascribed to morphine (Sj6berg et al. 1992). Although results of this and other animal studies (Yaksh and Collins 1989; Plummer et al. 1992) suggest that combinations of intrathecal morphine and clonidine might be clinically useful, studies of the combination given intrathecally (Coombs et al. 1987) or epidurally (Motsch et al. 1990; Van Essen et al. 1991) to postoperative patients have yielded mixed results. This may be due to use of non-optimal doses. Furthermore,
it is likely that the most immediate applications of such combinations will be in cancer patients who are already receiving high doses of morphine intrathecally, rather than in opioid-naive patients. Acknowledgements This work was supported by a grant from the National Health and Medical Research Council of Australia. Clonidine was kindly donated by Boehringer Ingelheim Pty. Limited, NSW, Australia.
References Bancroft JD, Stevens A (1982) Theory and practice of histological techniques. Churchill Livingstone, Edinburgh, pp. 306 Coombs DW, Fratkin JD, Meier FA, Nierenberg DW, Saunders RL (1985) Neuropathologic lesions and CSF morphine concentrations during chronic continuous intraspinal morphine infusion. A clinical and post-mortem study. Pain 22:337-351 Coombs DW, Saunders RL, Fratkin JD, Jensen LE Murphy CA (1986) Continuous intrathecal hydromorphone and clonidine for intractable cancer pain. J Neurosurg 64:890 894 Coombs DW, Jensen LB, Murphy C (1987) Microdose intrathecal clonidine and morphine for postoperative analgesia. Anesthesiology 67:A238 Cousins M J, Cherry DA, Gourlay GK (1988) Acute and chronic pain: use of spinal opioids. In: Cousins M J, Bridenbaugh P (eds) Clinical anesthesia and management of pain, 2nd edn. Lippincott, Philadelphia, pp 955 1029 DeConno F, Caraceni A, Martino C, Spoldi E, Salvetti M, Ventafridda E (1991) Hyperalgesia and myoclonus with intrathecal infusion of high-dose morphine. Pain 47: 337-339. Eisenach JC, Dewan DM, Rose JC, Angelo JM (1987) Epidural clonidine produces antinociception, but not hypotension, in sheep. Anesthesiology 66:496 501 Gordh TE, Ekman S, Lagerstedt A-S (1984) Evaluation of possible spinal neurotoxicity of clonidine, Upsala J Med Sci 89:266-273 Gordh T, Post C, Olsson Y (1986) Evaluation of the toxicity of subarachnoid clonidine, guanfacine, and a substance P antagonist on rat spinal cord and nerve roots. Anaesth Analg 65:1303 1311 Hogan Q, Haddox JD, Abram S, Weissman D, Taylor ML, Janjan N (1991) Epidural opiates and local anaesthetics for the management of cancer pain. Pain 46:271 279. Kitahata LM (1989) Spinal analgesia with morphine and clonidine. Anesth Analg 68:191-193 Lehmann EL (1975) Nonparametrics. Statistical methods based on ranks. McGraw-Hill, New York, pp. 134-137 Loomis CW, Milne B, Cervenko FW (1988) A study of the interaction between clonidine and morphine on analgesia and blood pressure during continuous intrathecal infusion in the rat. Neuropharmacology 27:191 199 McLean CF, Mather LE, Odontiadis J, Sloan PA (1990) Improved method for morphine determination in biological fluids and tissues: rapid, sensitive and selective. J Pharm Pharmacol 42:669-671 Motsch J, Gr/iber E, Ludwig K (1990) Addition of clonidine enhances postoperative analgesia from epidural morphine: A double blind study. Anesthesiology 73:1067-1073 Ossipov MH, Suarez L J, Spaulding TC (1989) Antinociceptive interactions between alpha-2 adrenergic and opiate agonists at the spinal level in rodents. Anesth Analg 68:194 200 Plummer JL, Cmielewski PL, Reynolds GD, Gourlay GK, Cherry DA (1990) Influence of polarity on dose-response relationships of intrathecal opioids in rats. Pain 40:339-347 Plummer JL, Cmielewski PL, Gourlay GK, Owen H, Cousins MJ (1991) Assessment of antinociceptive drug effects in the presence of impaired motor performance, J Pharmacol Methods 26:79-87
623 Plummer JL, Cmielewski PL, Gourlay GK, Owen H, Cousins MJ (1992) Antinociceptive and motor effects of intrathecal morphine combined with intrathecal clonidine, noradrenaline, carbachol or midazolam in rats. Pain 49:145-152 Sj6berg M, Karlsson P-A, Norborg D, Wallgren A, Nitescu P, Appelgren L, Linder L-E, Curelaru I (1992) Neuropathologic findings after long-term intrathecal infusion of morphine and bupivacaine for pain treatment in cancer patients. Anesthesiology 76:173-186 Solomon RE, Gebhart G F (1988) Intrathecal morphine and clonidine: antinociceptive tolerance and cross-tolerance and effects on blood pressure. J Pharmacol Exp Ther 245:444 454 Takano Y, Yaksh TL (1993) Chronic spinal infusion of dexmedetomidine, ST-91 and clonidine: spinal alpha 2 adrenoceptor subtypes and intrinsic activity. J Pharmacol Exp Ther 264:327 335. VanEssen E J, Bovill JG, Ploeger E (1991) Extradural clonidine does not potentiate analgesia produced by extradural morphine after meniscectomy. Br J Anaesth 66:237-241
Ventafridda V, Tamburini M, Caraceni A, DeConno F, Naldi F (1987) A validation study of the WHO method for cancer pain relief. Cancer 59:850-856 Wilcox GL, Carlsson K-H, Jochim A, Jurna I (1987) Mutual potentiation of antinociceptive effects of morphine and clonidine on motor and sensory responses in rat spinal cord. Brain Res 405:84 93 Wilczynska-Wojtulewicz I, Sadtij-Sosnowska N (1986) Determination of clonidine hydrochloride in pharmaceutical preparations by high-performance liquid chromatography. J Chromatogr 367:434 437 Yaksh TL, Collins JG (1989) Studies in animals should precede human use of spinally administered drugs. Anesthesiology 70:4 6 Yaksh TL Reddy SVR (1981) Studies in the primate on the analgetic effects associated with intrathecal actions of opiates, ct-adrenergic agonists and baclofen. Anesthesiology 54:451-467 Yaksh TL, Noueihed RY, Durant PCA (1986) Studies of the pharmacology and pathology of intrathecally administered 4anilinopiperidine analogues and morphine in the rat and cat. Anesthesiology 64:54-66