Opioid Tolerance Blunts the Reduction in the Sevoflurane Minimum Alveolar Concentration Produced by Remifentanil in the Rat

After obtaining Institutional Animal Care Committee approval (Madrid, Spain), the reduction in the sevoflurane MAC in response to remifentanil given in continuous infusion was evaluated in rats. Sevoflurane was obtained from Abbott (Sevorane; Abbott Laboratories, Madrid, Spain), and remifentanil from Glaxo-Wellcome (Ultiva; Glaxo-Wellcome Laboratories, Madrid, Spain).

Fifty-two male Wistar rats (Charles River Laboratories. Barcelona, Spain) with an average weight of 391 g (SD, 67 g), were housed in groups of 4 to 6 with a 12-h light, 12-h dark cycle; a relative humidity of 50 to 70%; and 20 ± 2°C ambient temperature. Food (B&K Universal, Grimston, England) and water were provided ad libitum . The animals were allowed to acclimatize for at least 1 week. All the studies were performed during the morning (starting at 9:00 am).

The rats were placed in an induction chamber into which 8% sevoflurane in a continuous oxygen flow of 3 l/min was directed (Sevoflurane Vaporizer, Sevorane Dräger Vapor 2000, Lübeck, Germany). After 2 to 3 min, the inhaled sevoflurane concentration was reduced to 3 to 5%. Tracheal intubation was performed using a 14-gauge polyethylene catheter (Terumo Surflo IV Catheter, Terumo Europe NV, Leuven, Belgium) with the animal positioned in sternal recumbency. A flexible, blunt-tip wire guide was inserted into the trachea with an otoscope and used to direct the endotracheal catheter. After positioning of the catheter was ascertained, it was connected to a small T piece breathing system with minimum dead space. Fresh gas flow to the T piece was adjusted to 1l/min oxygen, and sevoflurane concentration was adjusted as necessary. Remifentanil was administered with an infusion pump (Syringe pump, model Sep11S, Ascor S.A. Medical Equipment, Warsaw, Poland) by means of a 22-gauge polyethylene catheter inserted in a tail vein.

Heart and respiratory rates were continuously monitored. Arterial blood pressure, pulseoxymetry, and heart and respiratory rates were recorded immediately before each MAC step, and after 30 min of drug administration. The carotid artery was catheterized (Venocath-18, Venisystems, Abbott, Sligo, Ireland) via  surgical cutdown. This access allowed arterial blood sampling and blood pressure measurement via  a calibrated pressure transducer (Transpac IV; Abbott Laboratories, Abbott Park, IL). Systolic, diastolic, and mean arterial blood pressures were recorded continuously (RGB Medical Devices, Madrid, Spain). Arterial blood (1 ml) was collected for blood gas analysis (Rapidlab 860; Bayer AG, Leverkusen, Germany) at the end of the study to ensure that the values (at that time point) were within normal limits of pH (7.34–7.44), and oxygen (Pao²; > 90 mmHg), and carbon dioxide arterial partial pressures below 55 mmHg (Paco²; 35–55 mmHg). Rectal temperature was also monitored and maintained between 37.0°C and 38.5°C by means of a water-circulating warming blanket (Heat Therapy Pump, Model TP-220; Gaymar, Orchard Park, NY). A heating light was also used when necessary to maintain body temperature above 37.0°C.

Intratracheal gas sampling was used to measure anesthetic gas concentration and determine the MAC. This method has been described in detail previously.18In brief, a fine catheter with an 0.9-mm external diameter was inserted through the endotracheal catheter, with the fine catheter tip located at the level of the carina. The proximal end of the catheter was connected to a 10-ml gas-tight glass syringe (Hamilton Syringe, 1000 Series Gastight, model 26211-U; Sigma-Aldrich, St. Louis, MO). Samples were obtained by withdrawing 10 ml of gas over 5 min using an infusion pump (Mod. 55-2226; Harvard Apparatus, Millis, MA). Gas sampling actually mixed both alveolar and inspired gases, which were obtained consecutively in duplicate to ensure a constant alveolar concentration, and the final value was the mean, at every sevoflurane concentration step. The fine catheter was withdrawn between samples. After every step change in anesthetic concentration delivered by the anesthetic circuit, at least 10 min were allowed for equilibration before tail clamping. The samples were assayed using a side-stream infrared analyzer (Capnomac Ultima; Datex-Ohmeda, Hertfordshire, England).

The MAC of sevoflurane (MAC^SEV) and the MAC of sevoflurane plus remifentanil at a constant rate of infusion were established according to the method described elsewhere.19A painful noxious stimulus was applied with a long hemostat (8-inch Rochester Dean Hemostatic Forceps, Martin, Tuttlingen, Germany) clamped to the first ratchet lock onto the tail for 60 s immediately after the gas sample was obtained from the trachea. The tail was always stimulated proximal to a previous test site when the previous response was negative or distal if response was positive, starting 6 cm distal to the tail base. A positive response was considered to be a gross purposeful movement of the head, extremities, or body. A negative response was considered to be the lack of movement or grimacing, swallowing, chewing, or tail flick. Where a negative response was seen, the sevoflurane concentration was then reduced in decrements of 0.2% until the negative response became positive. Similarly, where a positive response was seen, the sevoflurane concentration was then increased until the positive response became negative. The MAC was considered to be the concentration midway between the highest concentration that permitted movement in response to the stimulus and the lowest concentration that prevented such movement. Determination of the MAC was performed in a laboratory 600 m above sea level.

The MAC^SEVwas determined four times in every animal. Once the animals were anesthetized and instrumented, a baseline MAC^SEVwas determined, and each animal acted as its own control. Remifentanil (RMF) was given intravenously by continuous infusion in the tail vein with no loading dose and the MAC was determined then (RMF-1), and again approximately 90 min later (RMF-2). Finally, the remifentanil infusion was stopped and the MAC redetermined (PostRMF). Thirty minutes were allowed between MAC determinations, and 1 h was usually necessary to determine the MAC value. Overall, every experiment lasted between 6 and 7 h. No attempt was made to determine the existence of an opioid-induced hyperalgesia effect when baseline MAC^SEVwas regained. In a further experiment, the dose of remifentanil was increased once an acute opioid tolerance effect was observed (RMF-3), and in these rats the MAC^SEVwas determined five times.

Baseline MAC^SEVwas determined in all animals before starting the opioid infusion. The animals were randomly given 1 of 2 doses of remifentanil at a constant rate of infusion, 120 μg · kg⁻¹· h⁻¹(n = 14) and 240 μg · kg⁻¹· h⁻¹(n = 14). The animals were euthanized with potassium chloride given intravenously while still deeply anesthetized. In a further group (n = 19), the rats were given 120 μg · kg⁻¹· h⁻¹and, once the MAC values determination indicated an acute opioid tolerance effect, the remifentanil dose was increased to 240 μg · kg⁻¹· h⁻¹and the MAC value reassessed (RMF-3).

A pilot study (n = 5) was performed to ensure that the MAC value was not modified over time. Therefore, the MAC^SEVwas determined at the four different studied times of the first experiment, but while administering saline instead of remifentanil.

Sample size calculation (N Query Advisor 2.0; Statistical Solutions; Saugus, MA) indicated an n value of 14 to determine differences as a result of an opioid tolerance effect with both remifentanil constant infusion rate doses. However, a larger n value (n = 19) was necessary to establish these differences in the second experiment.

A two-way ANOVA (dose × time) for repeated measures was performed, with the intrasubject factor being time and the intersubject factor being dose. Since an interaction between dose and time was found (P = 0.004, Greenhouse-Geisser correction) a one-way ANOVA for repeated measures (intrasubject factor being time) was performed for dose = 120 μg · kg⁻¹· h⁻¹and for dose = 240 μg · kg⁻¹· h⁻¹. Student’s t  test was employed using dose (120 μg · kg⁻¹· h⁻¹and 240 μg · kg⁻¹· h⁻¹) as factor to determine differences between the two doses at each study time. A P  value of < 0.05 was set to indicate statistical significance. All analyses were performed using the SPSS statistical package (v. 15 for Windows; SPSS Inc., Chicago, IL). Results are presented as mean ± SD.

The MAC^SEVranged from 1.7 to 2.8%, with an average value of 2.3% ± 0.3%.

Baseline sevoflurane MAC in rats was similar in both dose groups of remifentanil (P > 0.05, effect size = 0.297); i.e. , 2.4 ± 0.2 vol% (n = 14, mean ± SD; 120 μg · kg⁻¹· h⁻¹), and 2.3 ± 0.3 vol% (n = 14, 240 μg · kg⁻¹· h⁻¹). When a constant rate of infusion for remifentanil was used, it reduced the MAC values (RMF-1) in a dose-dependent fashion to 1.8 ± 0.2 vol% and 1.5 ± 0.3 vol%, at the low and high doses, respectively (ANOVA, P < 0.05; effect size = 1.004); i.e. , a percentage of reduction in the MAC^SEVby 24 ± 6 vol% and 36 ± 7 vol%, respectively. These results were obtained approximately 1.5 h after the baseline MAC^SEVdetermination; i.e. , at 90 min and 81 min (low and high doses of remifentanil, respectively) after starting the remifentanil constant rate of infusion (fig. 1).

Approximately 1 h and 30 min later, the MAC values were redetermined (RMF-2) and found to be higher than those previously observed (RMF-1), even though the same remifentanil constant rate of infusion had been maintained; i.e. , 2.1 ± 0.3 vol% and 1.9 ± 0.3 vol% at the low and high dose of remifentanil, respectively (ANOVA, P > 0.05; effect size = 0.721). These values actually only reduced the baseline MAC^SEVby 12 ± 7% and 17 ± 9%, at the low and high dose of remifentanil, respectively, with an increase over the earlier measurements, also with the remifentanil infusion (RMF-1), of 12 ± 5% and 9 ± 12%. These results were obtained 89 min and 76 min after the previous MAC determination (RMF-1), or 179 min and 157 min after the start of the remifentanil infusion, at the low and high doses, respectively.

After the second sevoflurane MAC determination with remifentanil (RMF-2), the remifentanil infusion was stopped. Then the sevoflurane MAC values for both infusion rate groups were determined again (PostRMF) and found to be higher than the previous MAC determination (RMF-2); i.e. , 2.4 ± 0.2 vol% for the low dose and 2.2 ± 0.3 vol% for the high dose (ANOVA, P > 0.05; effect size = 0.562). These results were obtained 4 to 5 h after starting the remifentanil infusions. No significant differences were found between the first and fourth MAC determinations; i.e. , MAC^SEVwithout remifentanil infusion.

All study times were significantly different when compared with the previous or following study time (P < 0.001, Greenhouse-Geisser correction) at both doses, 120 μg · kg⁻¹· h⁻¹and 240 μg · kg⁻¹· h⁻¹. These differences were found at every time interval: Baseline-RMF-1, P < 0.001 for both doses; RMF-1-RMF-2, P < 0.001 for both doses; RMF-2-postRMF P < 0.002 with dose = 120 μg · kg⁻¹· h⁻¹and P < 0.001 with dose = 240 μg · kg⁻¹· h⁻¹. A dose-dependent effect of remifentanil was found at RMF-1 (Student’s t  test, P = 0.006), but not at RMF-2 (P = 0.055). No differences between doses were found at baseline and postRMF (P = 0.442 and P = 0.140, respectively).

Results from the pilot experiment showed no differences over time, and variation was within the normal range for the MAC determination method (10–15%): 2.2 ± 0.1 vol%, 2.3 ± 0.2 vol%, 2.2 ± 0.1 vol%, and 2.2 ± 0.2 vol% at baseline, RMF-1, RMF-2, and postRMF times, respectively.

While the degree of reduction in the MAC^SEV(2.3 ± 0.3 vol%) produced by remifentanil (RMF-2, 1.8 ± 0.3 vol%) decreased over time (RMF-2, 2.1 ± 0.3 vol%), the initial reduction (RMF-1) could be reattained by increasing the remifentanil dose from 120 μg · kg⁻¹· h⁻¹to 240 μg · kg⁻¹· h⁻¹(RMF-3, 1.5 ± 0.3 vol%). This observation demonstrates an acute opioid tolerance effect since, if the MAC^SEVhad risen, it would have been a consequence of opioid-induced hyperalgesia. The remifentanil dose was not doubled from 240 to 480 μg · kg⁻¹· h⁻¹to counteract the induced remifentanil tolerance, because the latter dose produced severe respiratory depression in the rats. These results were obtained 5 to 6 h after starting the remifentanil infusion (fig. 2).

Like other potent opioids, remifentanil reduces the sevoflurane MAC in the rat in a dose-dependent fashion.20However, this reduction is not constant and is rapidly blunted in the short term, since a significant reduction is observed some 1.5 h after beginning the remifentanil infusion. This finding is consistent with previous reports reflecting the rapid development of acute opioid tolerance to remifentanil in people in the immediate postoperative period,3,21although the opposite, i.e. , a lack of an acute opioid tolerance effect, has also been reported, mostly in postsurgical patients.14–17Similarly, an acute opioid tolerance effect13and opioid-induced hyperalgesia10have been observed in healthy volunteers receiving remifentanil, possibly reflecting the importance of the surgical stimulus in the development of opioid-induced tolerance. Since increasing the remifentanil infusion dose maintained the initial reduction in the MAC^SEV, an acute opioid tolerance effect is suggested where increased doses were required to maintain the same level of analgesic effect.

The most commonly used method to determine acute opioid tolerance to remifentanil during the postoperative period in the clinical setting is to measure the dose of opioid administered, as the amount of morphine required over 24 h postoperatively, usually through patient-controlled administration. This has proven to be a more consistent method since it is a relatively objective measurement, although large interindividual differences may occur, and this may reduce the reliability of the method. However, no attempts have been made as yet to demonstrate acute opioid tolerance in the intraoperative period when a reduction of the MAC under inhalation anesthesia would consistently reflect the analgesic effect of opioids.

When the acute opioid tolerance effect was observed, the reduction in the MAC of remifentanil was roughly only 50% of the previous measurement. A similar or even stronger acute opioid tolerance effect has been described in conscious human volunteers in whom the analgesic effect declined, despite the constant rate of infusion, after 3 h of infusion to only one-fourth of the peak value, which had been observed 1 to 1.5 h after initiating the infusion,13and a threefold increase in pain scores was found as a result of acute opioid tolerance. Given the rapid development of acute opioid tolerance to remifentanil, the MAC reduction determined immediately after starting the remifentanil infusion may not reflect the maximum level of reduction potentially achieved, since it usually takes 80 to 90 min to determine MAC in rats. Similarly, tolerance might have been increased by allowing a longer intermeasurement period between MAC determinations during the remifentanil infusion. The magnitude of the MAC change produced by remifentanil is high enough as to be of clinical relevance, since either the dose of remifentanil or the concentration of the inhalant anesthetic should be increased to maintain a similar level of anesthesia. Obviously this may lead to a potential increase in the appearance of side effects derived from higher doses of opioids and inhalant anesthetics. However, since contradictory findings have been reported in clinical trials,3,10–17,21caution should be taken when extrapolating these findings to the clinical setting, and further studies in humans are necessary to confirm the results observed in the rat.

Determination of the MAC usually provides more consistent results when a standardized supramaximal noxious stimulus is employed.22While surgical incision is usually employed in humans,23tail clamping is the most common and consistent noxious stimulus used in the rat.20MAC values remain stable for long periods of anesthesia and are typically reduced by opioids such as morphine, buprenorphine, fentanyl, or remifentanil18,24in a dose-related fashion. Typical reductions range from 15 to 70%, although the doses employed in the rat are higher than in humans when comparing the mg per kg doses.25Accordingly, the reduction in the MAC value should be 20 to 30% when remifentanil is administered at a dose of 120 μg · kg⁻¹· h⁻¹, and 40 to 60% when the remifentanil dose is doubled to 240 μg · kg⁻¹· h⁻¹.20MAC reductions of up to 90% can be obtained in humans26,27when large doses of opioids are administered.

Although there is evidence that different noxious stimuli may reflect a similar acute opioid tolerance response,13a differential response has been suggested; i.e. , temperature (heat or cold), mechanical or electrical, may affect responses so a modified response cannot be ruled out.11The noxious stimulus produced by surgery may actually elicit a different response, so the acute opioid tolerance effect produced by remifentanil should be confirmed in a clinical setting. Thus the conflicting reports on the existence of a tolerance effect might be a consequence of, among other causes, the nociceptive stimulus employed.28 

In conclusion, short term infusion, i.e. , 90 min or less, of remifentanil at doses capable of decreasing the sevoflurane MAC in the rat actually decreases the effectiveness of the analgesic effect of the opioid. The observed effect would probably be a consequence of acute opioid tolerance and, if confirmed in humans, this loss of analgesic efficacy would have clinically significant implications.

The authors thank Dr. Enrique De Miguel, M.D., Head of the Experimental Surgery Department, La Paz University Hospital, Madrid, Spain, and Santiago Cano, Ph.D. in Mathematics, Research Support Service, University Complutense, Madrid, Spain.