The micro agonist opioid, remifentanil, has a very fast onset (half-life of equilibration between blood and brain = 1.3 minutes)  and an ultrashort duration of action [1-3]. The time required for a 50% reduction in blood concentration after discontinuation of an infusion that has attained steady state (context-sensitive half-time) is about three minutes and does not increase with duration of infusion [1-3]. These properties should confer ease of titration to changing intraoperative conditions and may be especially suitable for total intravenous anesthesia (TIVA). The purpose of this multicenter clinical trial was to examine the safety and efficacy of remifentanil in a TIVA regimen with propofol followed by remifentanil infusion for postoperative analgesia. This article reports the results of the postoperative phase. The results of the intraoperative phase, reported previously , suggest that TIVA with remifentanil and propofol is a useful anesthetic technique.
Emergence from remifentantil-based anesthesia should be swift and predictable. However, the quick offset of action, while desirable, may result in inadequate immediate postoperative analgesia. Thus, a transition must be made from remifentanil to some other longer-acting analgesic for surgeries that result in significant postoperative pain. Several methods have been proposed for making the transition. The method used in this clinical trial was to reduce the infusion of remifentanil from an anesthetic dose to an analgesic dose at the conclusion of surgery, with titration of the analgesic infusion in the recovery room for 30 minutes followed by a gradual transition to a longer-acting opioid.
This study was performed according to a detailed protocol that specified the doses and method of administration of all anesthetic drugs. The essential aspects of the protocol are given below.
Seven academic medical centers contributed patients undergoing spine, joint replacement, abdominal, or thoracic surgery. Each center studied patients undergoing only one type of operation. Thoracic surgery was studied at one center only, while each other type of surgery was studied at two centers. The rationale for the multicenter design was to quickly enroll a large number of patients undergoing diverse types of major surgical procedures. Male and female patients, 18 yr of age of older, ASA physical status I-IV, were enrolled in the study after giving institutionally approved, written informed consent. Otherwise eligible patients were excluded from the study based on the following criteria: (a) clinically significant and unstable medical conditions, (b) chronic use of benzodiaz-epines or clonidine or acute use within 12 h of surgery, (c) chronic opioid use or acute use within 12 h of surgery, (d) clinically significant drug abuse (e.g., alcohol or illicit drugs), (e) weight more than 100% of ideal body weight, (f) pregnant or breast-feeding women.
TIVA consisting of remifentanil, propofol, and vecuronium was administered according to the protocol summarized below:
- Two intravenous (IV) lines were used. Remifentanil was administered through one, all other drugs and fluids were administered through the other.
- Midazolam 0.025-0.050 mg/kg IV was given for premedication.
- Induction was with remifentanil, 1.0 micro g/kg IV, followed after 3 min by propofol, 0.5 mg/kg IV. If the patient was not unconscious 1 min after the propofol bolus, additional 20-mg boluses of propofol were administered every 30 s until the patient was unconscious.
- Remifentanil infusion at 0.5 micro g [centered dot] kg-1 [centered dot] min-1 ("smaller dose group") or 1.0 micro g [centered dot] kg-1 [centered dot] min-1 ("larger dose group"). Assignment of patients to smaller or larger infusion dose groups was randomized and double blind.
- Propofol was infused at 75 micro g [centered dot] kg-1 [centered dot] h-1.
- Vecuronium, 0.15 mg/kg IV, was administered to facilitate tracheal intubation. Six minutes after the remifentanil bolus, the trachea was intubated. Five minutes after tracheal intubation, the remifentanil infusion was reduced by 50%.
- No inhaled anesthetics were given.
- IV anesthetics were titrated upward to control responses to stimulation, as manifested by (a) an increase in systolic blood pressure > 15 mm Hg above the baseline (average of the two lowest systolic blood pressure measurements recorded within 14 days prior to surgery, in the preoperative holding area, and in the operating room immediately prior to induction of anesthesia) for >1 min; (b) heart rate >90 bpm for >1 min; and (c) clinical signs of light anesthesia. Upward titration consisted of a remifentanil IV push bolus of 1 micro g/kg, followed immediately by an increase in the infusion rate by 0.5 micro g [centered dot] kg-1 [centered dot] min-1 (smaller dose group) or 1.0 micro g [centered dot] kg-1 [centered dot] min (-1) (larger dose group). Up to five bolus doses and two infusion rate increases were permitted within a 5-min interval. If remifentanil given in this fashion did not control the response, propofol, 20 mg IV push bolus, was given every minute until the response was controlled; if three or more propofol boluses were required, the propofol infusion was increased to 125 micro g [centered dot] kg-1 [centered dot] min-1. The maximum allowable rate of the remifentanil infusion was 2 micro g [centered dot] kg-1 [centered dot] min-1 (smaller dose group) or 4 micro g [centered dot] kg-1 [centered dot] min-1 (larger dose group).
- IV anesthetics were titrated downward only in response to hypotension that was not responsive to replacement of fluid losses or bradycardia. Hypotension was arbitrarily defined as systolic pressure less than 80 mm Hg. Bradycardia was defined as heart rate < 40 bpm. Remifentanil was decreased by increments of 0.25 micro g [centered dot] kg-1 [centered dot] min-1 (smaller dose group) or 0.50 micro g [centered dot] kg-1 [centered dot] min-1 (larger dose group). If hypotension persisted after reduction of the remifentanil to 0.125 micro g [centered dot] kg-1 [centered dot] min-1 (smaller dose group) or 0.25 micro g [centered dot] kg-1 [centered dot] min-1 (larger dose group), then propofol was reduced to 50 micro g [centered dot] kg-1 [centered dot] min-1. Ephedrine and/or atropine was used if necessary when reduction of the anesthetics did not restore blood pressure or heart rate to an acceptable range.
- Vecuronium was titrated by intermittent boluses to maintain adequate surgical relaxation, with at least one twitch of the train-of-four maintained at all times.
- Fifteen minutes prior to the anticipated end of surgery, the propofol infusion was reduced by 50%. Vecuronium was reversed with neostigmine and glycopyrrolate. At the time of the last suture or skin staple, the propofol infusion was discontinued and the remifentanil infusion was reduced to 0.05 micro g [centered dot] kg-1 [centered dot] min-1 in both the smaller and larger dose groups and continued into the postoperative period.
- The trachea was extubated when adequate spontaneous ventilation and response to verbal commands were established.
The remifentanil postoperative infusion was continued for at least 45 min. During the first 30 min, the infusion was titrated with the goal of controlling pain at a level of 0-1 on a scale of 0-3 (0 = no pain; 1 = mild pain; 2 = moderate pain; 3 = severe pain). Titration was accomplished with boluses of 0.5 micro g/kg, accompanied by infusion rate increases of 0.05 or 0.1 micro g [centered dot] kg-1 [centered dot] min-1, with each change of dose separated by at least 5 min. The maximum allowable infusion rate was 0.5 micro g [centered dot] kg-1 [centered dot] min-1. If pain was moderate or severe at that rate, morphine was administered ("morphine rescue") as described below.
After 46 patients were treated, the protocol for postoperative remifentanil analgesia was modified because of unsatisfactory pain control immediately after emergence from anesthesia in some patients. The initial remifentanil analgesic infusion was increased from 0.05 to 0.10 micro g [centered dot] kg-1 [centered dot] min-1, and the method of adjusting subsequent doses was linked to the severity of pain. For mild pain, either the infusion was increased by 0.025 micro g [centered dot] kg-1 [centered dot] min-1 or a bolus dose of 0.25-0.5 micro g/kg was given. For moderate pain, the infusion was increased by 0.025-0.10 micro g [centered dot] kg-1 [centered dot] min-1, with no option for a bolus dose. For severe pain, the infusion was increased by 0.025-0.1 micro g [centered dot] kg-1 [centered dot] min-1, with the option of giving a bolus of 0.25-0.5 micro g/kg in conjunction with the increase in the infusion rate. One hundred eleven patients were treated under the modified postoperative infusion protocol.
The remifentanil analgesic infusion was reduced in the event of respiratory depression, which was defined as 1) oxygen saturation by pulse oximetry (SpO2) < 90% or respiratory rate < 12/min, in which case the remifentanil analgesic infusion was decreased midway between the current and previous rate; or 2) respiratory rate < 8 breaths/min, in which case the remifentanil analgesic infusion was temporarily discontinued and then restarted, once the patient had achieved a respiratory rate of > 12/min, at a rate 0.05 micro g [centered dot] kg-1 [centered dot] min-1 less than the last infusion rate. Naloxone was administered when deemed clinically necessary by individual investigators.
Thirty minutes after initiating the remifentanil postoperative infusion, morphine, 0.05 mg/kg IV, was given. If pain was moderate or severe at the maximum allowable remifentanil infusion rate (0.5 micro g [centered dot] kg-1 [centered dot] min-1), morphine administration was initiated prior to the 30-min time point. The remifentanil postoperative infusion was tapered by 50%, 15 min after the first dose of IV morphine, and was discontinued within 60 min. After the first dose of morphine, 0.05 mg/kg, investigators were free to continue postoperative analgesia with any method of their choosing.
The Aldrete postanesthesia recovery score  (Appendix) was determined at 2 and 5 min after entry into the recovery room and subsequently at 5-min intervals until a score >or=to 9 was obtained. Each occurrence of nausea and/or vomiting was recorded. Patients were queried for the presence of nausea within 15 min after extubation, at the time of discontinuation of the remifentanil analgesic infusion, and at the time of discharge from the recovery room.
The following end points were examined: remifentanil dose requirements for postoperative analgesia, efficacy of postoperative analgesia, respiratory and cardiovascular side effects, nausea, vomiting, and time to Aldrete score >or=to 9.
There were two groups of patients for the postoperative remifentanil infusion: the first 46 patients, treated under the original protocol, and the subsequent 111 patients, treated under the modified protocol. The initial remifentanil postoperative infusion rate and the method of titrating the infusion were different for these two groups, which therefore will be compared, when relevant, in the subsequent data analysis. Data from the different centers and types of surgeries are combined except where important differences were found. Statistical comparisons were performed with the t-test and chi squared test.
One hundred fifty-seven patients were studied. These patients were a subset of the 178 patients enrolled in the intraoperative portion of the trial ; excluded from analysis in this report are patients evaluated for safety but not efficacy and patients who received remifentanil intraoperatively but not postoperatively. Age, weight, gender, and the percentage of patients enrolled in the modified protocol (in distinction to the original protocol) are shown in Table 1, according to the type of surgery. The percentage of patients enrolled in the modified protocol varied between sites, depending upon the number of patients that had already been enrolled in the original protocol at the time the modified protocol was introduced. There was a substantial imbalance in the gender of patients enrolled in the centers performing abdominal surgery because of the prevalence of gynecologic surgery.
The postoperative remifentanil analgesic infusion required no adjustment in 48% of patients in the original protocol group and 47% in the modified protocol group. The distribution of remifentanil infusion doses at the time of morphine administration is shown in Figure 1. The majority of infusion rates (78%) were in the range of 0.05-0.15 micro g [centered dot] kg-1 [centered dot] min-1, while 5% were <0.05 micro g [centered dot] kg-1 [centered dot] min-1 and 17% were >0.15 micro g [centered dot] kg-1 [centered dot] min-1. "Morphine rescue" for inadequate pain control at the maximum allowable dose of remifentanil was required in 13 patients (8%). The mean total remifentanil dose during the postoperative infusion period was larger in the modified protocol group (4.7 micro g/kg [SD = 2.2]) than in the original protocol group (3.9 micro g/kg [SD = 2.7]) (t-test; P < 0.05).
At the end of the 30-min remifentanil analgesic infusion titration period, immediately prior to morphine administration, 59% of the patients in the original protocol group and 67% of the patients in the modified protocol group (or 64% of all patients) had pain scores of 0 or 1 (chi squared; P = 0.70). The corresponding mean remifentanil infusion rate was 0.064 micro g [centered dot] kg-1 [centered dot] min-1 (SD = 0.049) for the original protocol group and 0.086 micro g [centered dot] kg-1 [centered dot] min-1 (SD = 0.064) for the modified protocol group (t-test; P = 0.021).
Blood pressure and heart rate were stable during the infusion of remifentanil, as shown in Figure 2.
Nausea occurred in 35% of patients at some time during the study. Fifteen minutes after extubation, 2.5% of patients were nauseated, while at 30 min, just prior to the administration of morphine, 11.2% of patients were nauseated. At the time of discharge from the recovery room, 27.3% of patients were nauseated. Emesis occurred in 8% of patients at some time during the study.
The median time in the recovery room to reach an Aldrete score >or=to9 was 20 min for the original protocol group and 25 min for the modified protocol group (P > 0.05).
Respiratory adverse events, including SpO2 < 90, respiratory rate < 12 breaths/min, and apnea, occurred in 45 patients (29%). Two patients received naloxone to treat either low SpO2 or apnea in the recovery room. Both of these patients had received remifentanil boluses in addition to the postoperative infusion. Apnea occurred in 11 patients, 4 (8.7%) in the original protocol group and 7 (6.3%) in the modified protocol group. Apnea was treated by ventilation with a bag and mask and downward titration of the remifentanil infusion. Naloxone was given to one of the apneic patients. Of the 11 apneic patients, nine had received boluses of remifentanil prior to the onset of apnea. The incidence of respiratory adverse events was not evenly distributed between centers, as shown in Figure 3. Of the patients with respiratory adverse events, 73% were in two centers (thoracic and spine surgery) that included 38% of the total study population.
Infusion of remifentanil in analgesic doses produced effective pain relief in the immediate postoperative period after TIVA. The remifentanil infusion dose required for adequate analgesia was 0.05-0.15 micro g [centered dot] kg-1 [centered dot] min-1 for 78% of patients. The surgical procedures studied are capable of producing severe postoperative pain. Therefore, the range of 0.05-0.15 micro g [centered dot] kg-1 [centered dot] min-1 appears to be the approximate effective dosage range for treating major postoperative pain. However, some patients will require doses well outside this range. Morphine was administered to 8% of patients prior to the end of the 30-minute remifentanil infusion titration period because of inadequate pain control at the maximum remifentanil dose allowed by the protocol, 0.5 micro g [centered dot] kg-1 [centered dot] min-1. These patients could have received additional remifentanil in an ordinary clinical setting not otherwise constrained by a research protocol.
Respiratory adverse events, including respiratory depression (SpO2 < 90% or respiratory rate < 12 breaths/min) and apnea, were a notable problem, affecting 29% of patients. While respiratory rates < 12 breaths/min are not necessarily of clinical significance, apnea occurred in 7.0% of patients. Some degree of respiratory depression is not surprising in the course of treating pain with opioid analgesics; however, the incidence of apnea in this study would concern most anesthesiologists. The administration of a bolus of remifentanil preceded the onset of respiratory depression in 19 of 45 cases. A bolus of remifentanil preceded the administration of naloxone in 2 of 2 cases, and preceded the onset of apnea in 9 of 11 cases. These data suggest that remifentanil boluses plus an infusion are particularly likely to produce clinically significant respiratory adverse events. This may be related to the rapid onset of remifentanil , a property that may tend to produce pronounced peak effects, relative to the onset of drugs such as morphine that enter the brain more slowly.
The incidence of respiratory depression ranged from 0% to 75% depending upon the center. Of the patients with respiratory depression, 73% were in two centers (thoracic and spine surgery) that included 38% of the total study population. The reasons for the differences in outcome between centers are not known. Anesthesiologists have long recognized that subtle differences in the technique of administering anesthetic drugs can produce quite different results. However, it is seldom possible to identify the particular differences in technique that explain the different results obtained by various anesthesiologists. Differences in patient characteristics between centers, and the different types of operations performed, probably had an influence on the incidence and severity of respiratory depression. The thoracic surgery patients in particular may be more vulnerable. However, it is more difficult to explain the marked difference in the outcomes at the two centers performing spine surgery on this basis, as the operations and patient demographics appear to be similar.
One possible factor that may have affected the results in this trial was the mechanical details of the remifentanil infusion. Remifentanil was administered from a syringe pump containing a concentration of 50 micro g/mL. Infusion rates for a 70-kg patient typically ranged from 4.2 to 12.6 mL/h (0.05-0.15 micro g [centered dot] kg-1 [centered dot] min-1). The remifentanil infusion was "piggybacked" into a standard IV infusion line. The "dead space" of the main IV tubing between the port where the remifentanil was inserted into the main IV tubing and the patient's vein could have varied from approximately 1-5 mL, depending upon the type of tubing used and the site of insertion of the remifentanil infusion. Thus, changes in the rate of flow in the main IV tubing could have a substantial impact on the moment-to-moment delivery of remifentanil. The rates of flow in the main IV tubing were not specified by the protocol and were not required to be controlled by a pump. If the flow in the main IV tubing were suddenly increased substantially, the remifentanil contained in the dead space of the IV tubing would be suddenly flushed into the patient. Such a bolus could easily produce respiratory depression, including apnea. There was a particularly dramatic example of this phenomenon in a study patient who was returned to the stepdown unit and became unconscious and cyanotic, with a respiratory rate of 4 breaths/min, immediately after an IV line was flushed. The IV line had been used for remifentanil administration in the recovery room, and apparently a substantial amount of remifentanil remained in the IV tubing. Thus, the attention paid to the mechanical details of the remifentanil infusion may explain some of the variability of results.
Whatever the reasons, there were substantial differences in the results obtained by different investigators with respect to respiratory side effects, despite the use of an extremely detailed protocol. Anesthetic practice may not be easily reduced to a protocol, and there are intangible factors that may have an important effect on outcome.
All micro agonist opioids can produce serious respiratory depression. The administration of potent, rapidly acting opioids by continuous infusion for postoperative analgesia must be performed with meticulous attention to detail and constant vigilance. Extreme caution should be exercised in the postoperative administration of bolus doses of remifentanil, as nine of 11 cases of apnea in this study were preceded by remifentanil boluses. We recommend that the remifentanil infusion be inserted into the main IV line as close as possible to the IV catheter to minimize dead space, and the rate of the main IV infusion should be controlled at a rate that is high enough to continuously flush the remifentanil from the tubing. A more dilute remifentanil solution that would be run at higher rates (on a volume per time basis) would help to minimize the effect of variations in flow rate of the main IV tubing on delivery of remifentanil to the patient.
The incidence of nausea in this study was comparable to that found in other studies of opioid-based anesthetic techniques . The peak incidence of nausea occurred after administration of morphine, during which time the remifentanil was tapered and discontinued. Thus, it is unclear whether this nausea was due to morphine or the combination of remifentanil and morphine. Residual propofol from the TIVA regimen may also have had some influence on the incidence and/or timing of nausea .
The Aldrete postanesthetic recovery score is widely used clinically as a guide to readiness for recovery room discharge (Appendix Table 2). Scores of 9-10 are considered optimal for discharge. There have been few prospective studies using Aldrete scores to compare recovery from different anesthetic regimens, and this study was not designed primarily to assess the speed of recovery from remifentanil-based TIVA. The time to reach an Aldrete score >or=to 9 in this study (median = 20 or 25 min) was similar to the times reported by Kortilla et al.  for brief ambulatory surgical procedures performed under propofol/nitrous oxide (mean = 23 min) or isoflurane/nitrous oxide (mean = 23 min) anesthesia.
In conclusion, this open, dose-ranging study demonstrated the use of remifentanil for early postoperative analgesia after TIVA with remifentanil and propofol, and an approximate dosing range was established. However, the technique as described is probably not practical for routine clinical use because of the occurrence of apnea in 7% of patients. Additional studies are needed to investigate the transition from remifentanil to longer-lasting analgesics and to refine analgesic and antiemetic strategies that minimize respiratory depression and nausea while optimizing pain control.
The authors acknowledge the valuable participation of the following co-workers: Kathy Cox, BS, Deryck Duncalf, MD, Kathy A. East, MS, Barbara Gaylord, RN, Ingrid Hollinger, MD, Karen F. Kessler, RN, Evan D. Kharasch, MD, PhD, George Kirvassilis, MD, and William Nichols, MD.
The Aldrete Score is a three-point postanesthesia recovery score that assesses five areas: activity, blood pressure, color, level of consciousness, and respiration . The scores for each area are added to obtain the total score. A total score >or=to 9 is considered to be indicative of satisfactory recovery from anesthesia.