Applying the principles of thermal, mechanical, chemical, or electrical stimulus, a variety of methods have been developed to assess the analgesic activities of opioids in animals [1-5]. Although each of these techniques can detect the analgesic property of drugs similar to that of morphine, they vary considerably in their sensitivity [2,3,6,7]. A change in the reaction time of animals exposed to a thermal stimulus, such as in the tail-flick or hot-plate test, is probably the most widely used method of determining analgesic activity [3,4,6]. However, the sensitivities of these tests have been questioned by the fact that they have not consistently demonstrated the antinociceptive activities of several mixed agonist-antagonist analgesics, such as buprenorphine and nalbuphine [2-4,6-10].
Since most thermal tests of nociception use heat as a stimulus, it might therefore be of importance to have a test where cold sensitive nociceptors are stimulated. A cold water tail-flick test in the rat (-10 degrees C) was reported  to determine the antinociceptions of opioid agonists and mixed agonist-antagonists without a false-positive reaction to nonopioids. This was a great step in the detection of the antinociceptions of opioids. However, the test is not sensitive enough. The efficacies of agonist-antagonists in this test were too low. It is very difficult to differentiate the potency between small and large doses of these drugs.
In an attempt to design a novel method which can recognize the antinociceptive actions of opioid analgesics, such as agonists and agonist-antagonists, yet satisfy the criteria of sensitivity, specificity, simplicity, and reproducibility, we developed the cold ethanol tail-flick test (CET) in the rat.
Male Sprague-Dawley rats weighing between 175 and 225 g were used. They were housed in groups of five to six at least 1 wk in a climate-controlled room maintained at 21 degrees C with approximately 50% relative humidity. Lighting was on a 12-h light/dark cycle (lights on at 6:00 AM), with food and water available ad libitum up to the time of testing. All tests were performed in accordance with the recommendations and policies of the International Association for the Study of Pain and the National Institutes of Health guidelines for the handling and use of experimental animals.
The apparatus used was a custom-designed circulation system (BL-110; Yie-der, Taipei, Taiwan) with a temperature range adjustable from 100 to -35 degrees C (+/- 0.1 degrees C precision) and a bath solution of water or 75% ethanol. Animals were held firmly over the opening of the bath and their tails submerged approximately half way into the bath. The nociceptive threshold was taken as the latency until the rat removed or flicked its tail from the bath. The time from immersion to tail removal was measured to the nearest tenth of a second with a laboratory timer.
Temperatures warmer than 0 degrees C have no value in detecting antinociception of opioid agonist-antagonists [4,12]. Thus, temperatures between -5 and -30 degrees C were applied. To obtain an optimal temperature range, -5, -10, -15, -20, -25, and -30 degrees C were screened. Male Sprague-Dawley rats (n = 12) without medication were studied. At 2-h intervals, each animal was exposed to each one of the above-mentioned cold temperatures. During testing, a rat which was held more than 60 s might move its body or tail and possibly interfere with the test results. Therefore, a baseline tail-flick latency greater than 60 s, or group mean values of the baseline tail-flick latency greater than 30 s, was not practical. According to these points, a screened temperature that had either one of the positive results was considered not strong enough to act as a noxious stimulus and was unreliable for practical use.
After screening, temperatures between -5 and -15 degrees C were thought to be not strong enough to act as noxious stimuli and were unreliable for practical use. The other temperature range between -20 and -30 degrees C was the candidate range of testing, and one of the temperatures (-20 degrees C) was chosen for testing.
Before medication, all animals (n = 6, each group) were tested at 35, 25, and 15 min. To minimize the variabilities, the first datum was discarded and the remaining two were averaged to obtain a baseline reading. In this way, each animal served as its own control: experimental values were compared with the predrug baseline values of each rat. To minimize damage to the animal's tail, a predetermined cutoff time of 40 s was used, and was considered to be the maximum latency. Under this condition, neither frostbite nor skin color changes was found on the tail throughout the experiment.
Measurements of the antinociceptive thresholds of opioids were done 20 min apart or even longer after intramuscular (IM) administration of different doses of an opioid agonist (morphine) and two agonist-antagonists (nalbuphine, buprenorphine). After testing, the dose-response time curves were constructed.
The sensitivity of CET to opioids was compared with assays using heat (radiant heat and hot water) morphine 5 mg/kg, and nalbuphine 0.5 and 100 mg/kg. The radiant heat tail-flick test [3,4] was thermally evoked by laying the tail of the rat over a slit through which the light of a 300-W quartz bulb was focused. The latency from the time of stimulus and tail-flick was assigned as the response latency. To prevent tissue damage, the trials were terminated after 8 s. The hot water tail-flick test, described previously by Cowan et al. , is a rat-tail immersion test with a water bath maintained at 50 degrees C. The cutoff time was set at 5 s.
The specificity of the CET was challenged by a specific opioid antagonist, nonopioid analgesics, and tranquilizers. In the antagonist study, animals (n = 6, each group) were pretreated (5 min before) with saline or naloxone 0.4 mg/kg IM, and then each received one of the test opioids (buprenorphine, nalbuphine, or morphine) IM in the contralateral hind legs. In the nonopioid study (n = 6 for each drug), several nonopioid analgesics (aspirin, acetaminophen, indomethacin, and ketoprofen), tranquilizers (diazepam and droperidol), and the vehicles used (0.9% saline or 10% ethanol) were injected IM.
The following drugs and doses were used: buprenorphine hydrochloride 100, 10, 5, 0.5, 0.1, and 0.01 micro gram/kg; nalbuphine hydrochloride 10, 5, 1, 0.5, 0.1, 0.05, and 0.01 mg/kg; morphine hydrochloride 10, 5, 1, 0.5, 0.1, 0.05, and 0.01 mg/kg; naloxone hydrochloride 0.4 mg/kg; droperidol 0.3 mg/kg; diazepam 3 mg/kg; acetaminophen 50 mg/kg; aspirin 100 mg/kg; indomethacin 0.3 mg/kg; and ketoprofen 10 mg/kg. Drugs, such as morphine, buprenorphine, nalbuphine, naloxone, and droperidol, were dissolved in 0.9% saline. All remaining drugs were dissolved in 10% (vol/vol) ethanol/0.9% saline. The solvents were also tested as controls.
In the study, the antinociceptive effects of drugs were transformed to be a percentage of maximum possible analgesia (%MPA) using the formula: Equation 1
Mean +/- SEM value of the maximum latency was calculated for each treatment group. The AD50 values (dose of agonist that produced 50% analgesia) and 95% confidence intervals were derived by log-linear interpolation using points between 16% and 84% of MPA on the dose-response curve . Statistical significance of data was determined by the following tests: 1) One-way analysis of variance with Bonferroni t-test when comparing different treatment groups; and 2) Student's t-test when testing the differences between groups. P value less than 0.05 was considered significant.
As shown in Figure 1, at a test temperature of -5 degrees C, 4 of 12 rats had a baseline tail-flick latency beyond 60 s. Also, at temperatures of -10 and -15 degrees C, 2 and 1 of 12 rats, respectively, had baseline tail-flick latencies far beyond 60 s. The mean values of the baseline tail-flick latency at temperatures of -5 and -10 degrees C were greater than 30 s. However, there was no prolonged latency occurring for -20 degrees C or lower. Therefore, a temperature range from -5 to -15 degrees C was thought not to be strong enough to act as a noxious stimulus and was unreliable for practical use.
An IM injection of morphine produced a dose-related antinociception to the cold (-20 degrees C) stimulus in the dose range from 0.1 to 10 mg/kg. All animals in the 1 mg/kg group showed maximum analgesia. Buprenorphine and nalbuphine also produced dose-related antinociception in the dose range from 0.5 to 100 micro gram/kg and 0.1 to 100 mg/kg, respectively.
(Figure 2) shows the dose-response curves of buprenorphine, morphine, and nalbuphine comparing the %MPA of each in CET. Buprenorphine was very potent, having the strongest antinociceptive actions, while morphine and nalbuphine showed similar potencies. The AD50 values with 95% confidence intervals determined by CET for buprenorphine, morphine, and nalbuphine were 0.22 (0.09-0.51) micro gram/kg, 0.16 (0.007-0.38) mg/kg, and 0.19 (0.08-0.49) mg/kg, respectively.
In the antagonist study, antinociception produced by morphine (0.5 mg/kg), buprenorphine (0.5 micro gram/kg), and nalbuphine (0.5 mg/kg) were all completely blocked by prior administration of naloxone (0.4 mg/kg, IM; Figure 3).
In studies of nonopioid drugs, as shown in Table 1, %MPA +/- SEM in the control groups, such as 0.9% saline and 10% (vol/vol) ethanol, showed only 2% +/- 2% and 3% +/- 2% changes. IM administration of diazepam and droperidol caused marked sedation, but did not demonstrate significant antinociception when compared with the saline group. Acetaminophen, a nonopioid analgesic, and several nonsteroid antiinflammatory drugs, such as aspirin, indomethacin, and ketoprofen, did not show any significant antinociception in the CET when compared with the control group of 10% (vol/vol) ethanol vehicle. Using radiant heat and hot water tail-flick tests (50 degrees C), nalbuphine (0.5 and 100 mg/kg) was not an effective antinociceptive drug, whereas morphine (5 mg/kg) showed a significant analgesic effect Table 2.
Agonists at both mu- and kappa-opioid receptor sites are antinociceptive in animals [3,4,6]. Buprenorphine, a mixed agonist-antagonist with high affinity to both mu- and kappa-opioid receptors, demonstrates a partial agonist activity at the mu receptor and an antagonist activity at the kappa receptor [14,15]. In contrast, nalbuphine shows a partial agonist activity at the kappa receptor and an antagonist activity at the mu receptor [9,16]. Both drugs are potent analgesics in clinical use [2,8,10,14,16]. However, assays using heat as the noxious stimulus do not reliably detect the analgesic activity of these drugs [4,6,17]. Schmauss and Yaksh  reported that intrathecal administration of mu (morphine) and delta (enkephalin) agonist but not kappa agonist (U50488H, etc.) or agonist-antagonists (nalbuphine and buprenorphine) produced a dose-dependent inhibition of all cutaneous thermal responses (hot-plate and radiant heat tail-flick) in the rat. Shaw et al.  and other authors [6,17] also reported similar results.
The CET has properties of testing analgesic efficacies of opioids. As expected in any valid test for antinociception, morphine produced a dose-related effect that reached maximum levels. Its analgesic action was blocked by naloxone. However, it is quite different from assays using heat or cold water (-10 degrees C). Mixed agonist-antagonist opioids (buprenorphine and nalbuphine) also gave dose-related antinociceptive effects that reached a maximum and would be blocked by naloxone.
Since many other nonopioid analgesics have not been used in the test, we will conduct further study to challenge the specificity of the CET. According to the preliminary study results, as shown in Table 3, commonly used assays using a thermal, mechanical, chemical, or electrical stimulus can detect the antinociceptive activities of many nonopioid compounds. Unlike most other assays, the CET is more specific to opioids. The specificity of the CET has been challenged preliminarily using two tranquilizers (droperidol and diazepam), one nonopioid analgesic (acetaminophen), and three nonsteroid antiinflammatory drugs (aspirin, indomethacin, and ketoprofen). The results showed that the CET ignored the antinociceptive actions of nonopioids.
On the screening of the optimal temperature range, tail-flick response was sharper and faster in the cooler temperature range (-20 to -30 degrees C). This suggested that the noxious cold stimulus produced within the temperature range of -20 to -30 degrees C is stronger and more reliable for practical use. Pizziketti et al.  reported a cold water tail-flick test in rats using a water/ethylene glycol mixture maintained at -10 degrees C. Their study showed that the cold water tail-flick test could recognize and differentiate the antinociceptive actions between opioid agonists (morphine and methadone) and mixed agonist-antagonists (pentazocine, nalorphine, and nalbuphine) without false-positive results with central nervous system depressants (chlorpromazine, etc.) or nonopioid analgesics (acetaminophen, etc.). This is a major step in detecting the antinociceptive action of opioids using cold stimulus. However, some unresolved problems remain, such as: 1) The low efficacies of narcotic agonist-antagonists make it difficult for differentiating the antinociception among different doses. 2) Is the temperature of -10 degrees C an optimal temperature to act as an noxious cold stimulus? What is the optimal temperature range? 3) Is the bath solution of ethylene glycol and water an ideal antifreeze agent for the cooling system? What is the ideal bath solution? In our study, a wide temperature range was explored. The optimal temperature range was suggested to be between -20 degrees C and -30 degrees C. In dose-response curves (drawn at -20 degrees C), the antinociceptive action of opioid agonist-antagonists can reach the maximum analgesic effect. This is quite different from the cold water tail-flick test reported by Pizziketti et al. . Also, the bath solution was changed to 75% ethanol, and no solid debris was found with temperatures cooler than -30 degrees C.
Concerning the variables throughout the study, Hole and Tjolsen  reported that rat skin temperatures may play a role in modulating the latency of the thermal (heat) tail-flick test and formalin tests. Different background skin temperatures may cause different response latencies especially at the ambient temperature of 24-25 degrees C, more so than 21-22 degrees C. However, Lichtman et al.  reported that altering tail-flick latencies to account for changes in tail temperature when using the tail-flick test is unwarranted, and monitoring tail-skin or core temperature is unnecessary. In our study, even though we did not monitor tail-skin or core temperature, we did consider the temperature effect. We kept the ambient temperature at 21 degrees C, a relatively stable circumstance. We also attempted to eliminate the stress factors, as rats were housed in relatively comfortable conditions 1 wk before testing so that they might be habituated to the testing environment. Also, during the testing procedure, gentle manipulations were requested. All of these efforts were performed to minimize the variability of the test.
Since most thermal tests of nociception use heat as a stimulus, it might be of potential importance to use cold as a noxious stimulus. Although the mechanisms responsible for the differences seen between noxious heat and cold are unknown, the opiate mixed agonist-antagonists may play an important role as full agonists in modulating nociception produced by cold.
In conclusion, the rat cold ethanol tail-flick test is simple to perform and requires no specialized equipment or extensive training. It permits the testing of many animals per day and allows several readings in an individual animal to observe the time course of drug effects. The CET provides a sensitive, reproducible, and opioid-specific test for antinociceptive actions of opioid analgesics.
The authors thank Professor Shih-Hsun Ngai for his kind assistance in preparation of the manuscript. Generous gifts of buprenorphine hydrochloride (Macfarlan Smith) and nalbuphine hydrochloride (Du Pont Merck) are gratefully acknowledged.
1. Taber RI. Predictive value of analgesic assay in mice and rats. In: Braude MC, Harris LC, May EL, et al., eds. Narcotic antagonists: advances in biochemical psychopharmacology. New York: Raven Press, 1974;8:191.
2. Cowan A, Lewis JW, Macfarlane IR. Agonist and antagonist properties of buprenorphine, a new antinociceptive agent. Br J Pharmacol 1977;60:537-45.
3. Hayes AG, Sheehan MJ, Tyers MB. Differential sensitivity of models of antinociception in the rat, mouse and guinea-pig to mu- and kappa-opioid receptor agonists. Br J Pharmacol 1987;91:823-32.
4. Shaw JS, Rourke JD, Burns KM. Differential sensitivity of antinociceptive tests to opioid agonists and partial agonists. Br J Pharmacol 1988;95:578-84.
5. Bloss JL, Hammond DL. Intermittent shock titration task in the Rhesus monkey and its validity as an analgesiometric procedure. Brain Res 1989;499:344-56.
6. Tyers MB. A classification of opiate receptors that mediate antinociception in animals. Br J Pharmacol 1980;69:503-12.
7. Dykstra LA, Massie CA. Antagonism of the analgesic effects of mu- and kappa-opioid agonists in the squirrel monkey. J Pharmacol Exp Ther 1988;246:813-21.
8. Lewis JW. Buprenorphine. Drug Alcohol Depend 1985;14:363-72.
9. Schmidt WK, Tam SW, Schotzberger GS, et al. Nalbuphine. Drug Alcohol Depend 1985;14:339-62.
10. Schmauss C, Yaksh TL. In vivo studies on spinal opiate receptor systems mediating antinociception. II. Pharmacological profiles suggesting a differential association of mu-, delta- and kappa-receptors with visceral, chemical and cutaneous thermal stimuli in the rat. J Pharmacol Exp Ther 1984;228:1-12.
11. Pizziketti RJ, Pressman NS, Geller EB, et al. Rat cold water tail-flick: a novel analgesic test that distinguishes opioid agonists from mixed agonist-antagonists. Eur J Pharmacol 1985;119:23-9.
12. Clark SJ, Follenfant RL, Smith TW. Evaluation of opioid induced antinociceptive effects in anaesthetized and conscious animals. Br J Pharmacol 1988;95:275-83.
13. DeLean A, Munson PJ, Rodbard D. Simulataneous analysis of families of sigmoidal curves: application to bioassay radioligand assay and physiological dose-response curves. Am J Physiol 1978;235:E97-102.
14. Leander JD. Buprenorphine is a potent kappa-opioid receptor antagonist in pigeons and mice. Eur J Pharmacol 1988;151:457-61.
15. Negus SS, Dykstra LA. Kappa-antagonist properties of buprenorphine in the shock titration procedure. Eur J Pharmacol 1988;156:77-86.
16. Schmauss C, Doherty C, Yaksh TL. The analgetic effects of an intrathecally administered partial opiate agonist, nalbuphine hydrochloride. Eur J Pharmacol 1983;86:1-7.
17. Leighton GE, Rodriguez RE, Hill RG, Hughes J. Kappa-opioid agonists produce antinociception after i.v. and i.c.v. but not intrathecal administration in the rat. Br J Pharmacol 1988;93:553-60.
18. Ossipov MH, Suarez LJ, Spaulding TC. A comparison of the antinociceptive and behavioral effects of intrathecally administered opiates, alpha2-adrenergic
agonists and local anesthetics in mice and rats. Anesth Analg 1988;67:616-24.
19. Tiseo PJ, Geller EB, Adler MW. Antinociceptive action of intracerebroventricularly administered dynorphin and other peptides in the rat. J Pharmacol Exp Ther 1988;246:449-53.
20. Schmidt C, Peyroux J, Noble F, et al. Analgesic responses elicited by endogenous enkephalins (protected by mixed peptidase inhibitors) in a variety of morphine-sensitive noxious tests. Eur J Pharmacol 1991;192:253-62.
21. Dykstra LA. Effects of ketocyclazocine and ethylketocyclazocine on electric shock titration. Eur J Pharmacol 1983;94:19-26.
22. Dykstra LA. Effects of buprenorphine on shock titration in squirrel monkeys. J Pharmacol Exp Ther 1985;235:20-5.
23. Loomis CW, Penning J, Milne B. A study of the analgesic interaction between intrathecal morphine and subcutaneous nalbuphine in the rat. Anesthesiology 1989;71:704-10.
24. Hendershot LC, Forsaith J. Antagonism of the frequency of phenylquinone-induced writhing in the mouse by weak analgesics and nonanalgesics. J Pharmacol Exp Ther 1959;125:237-40.
25. Carlsson KH, Jurna I. Interaction of pentobarbital and morphine in the tail-flick test performed on rats: synergism at the spinal and antagonism at the supraspinal level. Neurosci Lett 1986;71:356-60.
26. Ellenbroek BA, Peeters BW, Honig WM, Cools AR. The paw test: a behavioral paradigm for differentiating between classical and atypical neuroleptic drugs. Psychopharmacology (Berlin), 1987;93:343-48.
27. Bloss JL, Hammond DL. Shock titration in Rhesus monkey: effects of opiate and nonopiate analgesics. J Pharmacol Exp Ther 1985;235:423-30.
28. Chojnacka-Wojcik E, Naparzewska A. Pharmacological properties of the antiinflammatory agent, 10-methylamine-11-h-indeno-[1,2-b]quinolin-11-one. Pol J Pharmacol Pharm 1983;35:327-32.
29. Malmberg AB, Yaksh TL. Antinociceptive actions of spinal nonsteroidal anti-inflammatory agents on the formalin test in the rat. J Pharmacol Exp Ther 1992;263:136-46.
30. Carter RB. Differentiating analgesic and non-analgesic drug activities on rat hot plate: effect of behavioral endpoint. Pain 1991;47:211-20.
31. Hole K, Tjolsen A. The tail-flick and formalin tests in rodents: changes in skin temperature as a confounding factor. Pain 1993;53:247-54.
32. Lichtman AH, Smith FL, Martin BR. Evidence that the antinociceptive tail-flick response is produced independently from changes in either tail-skin temperature or core temperature. Pain 1993;55:283-95.