Awakening Propofol Concentration with and without Blood-Effect Site Equilibration after Short-term and Long-term Administration of Propofol and Fentanyl Anesthesia 

After approval from the District Ethics Committee of Hamamatsu University Hospital, 36 patients classified as American Society of Anesthesiologists physical status I and II who were aged 20–56 yr and scheduled for elective plastic surgery expected to last <30 min were selected for short-term propofol infusion (group 1). Sixty-four patients classified as American Society of Anesthesiologists physical status I and II who were aged 20–59 yr and scheduled for elective surgery of the extremities, neck, or body expected to last 3–4 h were selected for long-term propofol infusion (group 2). Patients with known cardiac, pulmonary, liver, renal, or metabolic disease, and those who were significantly obese (body mass index > 30) or were receiving any form of medication were excluded from the study. The patients provided written informed consent after the study was explained to them. No preanesthetic medication was given. When the patient arrived in the operating room, during local anesthesia a cannula was inserted into a peripheral vein to infuse propofol and fentanyl and to replace fluids (6–8 ml [center dot] kg sup -1 [center dot] h sup -1 lactated Ringer's solution). An arterial catheter was inserted to measure blood pressure and to sample blood. Heart rate, blood pressure, electric activity of the heart, end-tidal carbon dioxide, rectal temperature, and oxyhemoglobin saturation were monitored continuously during the surgery. Body temperature was maintained at 36–37 [degree sign] Celsius using a forced-air warming blanket (Bear Hugger; Augustine Medical, Eden Prairie, MN).

Propofol and fentanyl were administered with target-controlled infusion devices. In both groups, anesthesia was induced with a target propofol concentration of 5 micro gram/ml and a fentanyl concentration of 3 ng/ml, and with intubation facilitated by 0.1 mg/kg vecuronium. Respiration was controlled to maintain normocarbia as measured by capnography. Vecuronium, in 1–3 mg intravenous bolus doses, was given as required by surgical conditions. At least one or two twitch responses of train-of-four stimulation were always present, as measured by a peripheral neuromuscular function monitor. At the end of the surgical procedure, 2.5 mg neostigmine and 1 mg atropine were given intravenously. Propofol and fentanyl target concentrations, which were maintained in ranges of 2.5–5 micro gram/ml and 2–3 ng/ml, respectively, during most of the surgery, were rigidly controlled at 3 micro gram/ml and 1 ng/ml, respectively, for at least the final 20–30 min of surgery. To ensure equivalent concentration sampling, target propofol and fentanyl concentrations were kept steady for at least 30 min after surgery before blood samples were obtained. The concentration-hour of propofol (integration of predicted propofol concentration) was calculated in each patient. The patients were further divided into two subgroups within each group: subgroup A in which abrupt discontinuation of propofol resulted in disequilibration between the blood and effect site concentrations; and subgroup B in which 0.2 micro gram/ml descending sequences of propofol concentrations for 15-min intervals at each concentration were used to equilibrate blood to the effect site. Arterial blood for propofol and fentanyl concentrations was sampled at discontinuation of propofol in the A subgroups and just before the descending step of propofol in the B subgroups (3 micro gram/ml of propofol and 1 ng of fentanyl).

In the A subgroups, the response to verbal command was evaluated every 30 s, and arterial blood samples for propofol and fentanyl concentration were taken when patients responded positively. The time from termination of propofol to a positive response to verbal command was also recorded.

In the B subgroups, each descending step of the target concentration was maintained for 15 min. Two blood samples of each step were drawn to measure propofol concentrations 10 and 15 min into each step (Figure 1). We calculated the mean value of the two samples as a step propofol concentration. The midpoint concentration between the step concentration at the first positive response and the step just before a positive response was called the awakening concentration equilibrated with the effect site for each patient. A blood sample for fentanyl concentration was also taken at the step of positive response.

To maintain the blood concentration of propofol and fentanyl, we used a pharmacokinetic model-driven infusion device, such as those designed for computer-assisted continuous infusion, which consisted of a microcomputer (NEC 9821; Tokyo, Japan) interfaced to a syringe pump (model 1235; ATOM Co., Tokyo, Japan) and using a three-compartment model with central elimination. The rate of drug-mass change in the three compartments can be expressed as three simultaneous differential equations. Table 1lists the pharmacokinetic parameters used in computer-assisted continuous infusion for fentanyl and propofol. [2,7]We solved these equations using the pharmacokinetic microconstants using the Runge-Kutta [8]method at the recursive period of 12 s. We programmed the control software ourselves in Turbo Pascal (Borland International, Scotts Valley, CA).

Blood samples were kept on ice and stored at 5 [degree sign] Celsius until extraction and assay. Plasma concentrations of propofol were determined using high-performance liquid chromatography with fluorescence detection at 310 nm after excitation at 276 nm (CTO-10A, RF550, and C-R7A; Shimadsu, Kyoto, Japan). [9]For each batch of blood samples (from a single patient), a standard curve was computed by adding pure propofol liquid to drug-free human plasma to make up concentrations of 1, 5, 10, and 25 micro gram/ml. Linear regression (the method of least squares) was used with serum propofol concentration as the dependent variable. Propofol concentrations in this study were calculated using the obtained regression equation. The lower limit of detection was 17 ng/ml, and the coefficient of variation was 8.4%. The plasma for fentanyl concentration was separated and frozen at -70 [degree sign] Celsius until assay. The plasma concentration of fentanyl was measured by a gas chromatograph mass-spectrometer (model 5989; Hewlett Packard, CA) in an outside laboratory. The lower limit of detection was 0.2 ng/ml.

Probit analysis was used to define the probability of no response to verbal command. The EC⁵⁰and EC⁵⁰eq were obtained from this analysis. All data are presented as means +/- SD if not stated otherwise. Statistical analysis included analysis of variance for principal effects and Student's t test. Probability values < 0.05 were considered significant.

The demographic data profiles of the patients did not differ significantly between groups (Table 2). Anesthesia induction and maintenance were smooth in all cases, and no excitatory movements were observed. The concentration-hour did not differ significantly between the subgroups (Table 2). Recovery from anesthesia was similarly uneventful.

Fentanyl concentrations during the measurement of response to verbal command were stable and did not differ significantly between the groups (Table 2). Propofol concentrations at discontinuation of the A subgroups and just before the first descending step of the B subgroups were similar (Table 2). In the B subgroups, propofol blood concentrations at just one step before positive response to verbal command and the step at positive response to verbal command confirmed a stable concentration (Figure 2).

(Table 2) shows the EC⁵⁰and EC⁵⁰eq values of each group, and Figure 3shows the dose-response curves. The EC⁵⁰after the abrupt discontinuation of propofol (group 1A) was significantly lower than that after the long-term decreasing administration (group 2A). However, the EC⁵⁰eq value obtained under the condition of equilibration between plasma and the effect site propofol concentration in group 1B was quite similar to that in group 2B. In groups 1 and 2, the EC⁵⁰values were lower than those for EC sub 50 eq. The time required to respond to a verbal command was 9.3 +/- 3.5 min for group 2A, which was significantly longer than the 5.2 +/- 1.8 min required for group 1A (Table 2).

We observed that the EC⁵⁰of awakening concentration after long-term propofol infusion was greater than that after short-term infusion and entailed a longer awakening time. The blood propofol awakening concentration after short-term infusion was consistent with previously reported values obtained after repeated bolus administration (0.9–1 micro gram/ml). [3]Shafer et al. [1]reported that the predicted blood propofol concentration at which 50% of patients awoke after surgery was 1.07 micro gram/ml. They found that this value was independent of patient age, sex, weight, or type of surgery. But the patients who had major surgery took a longer time to wake (18 min compared with 9 min). [1]The EC⁵⁰value in our study was influenced by the length of propofol infusion, and the EC⁵⁰after long-term infusion of propofol in our study was higher than that reported by Shafer et al. Possible explanations for the observed higher EC⁵⁰in our study include a different infusion regimen, a different blood sampling site, [10]and different opiate administration. Schuttler et al. [4]reported that awakening time and propofol concentration were 7.9 min and 1.59 micro gram/ml after 131 min of propofol infusion, which are similar to those of our EC⁵⁰of long-term propofol infusion. Wessen et al. [5]reported a 0.7 micro gram/ml propofol concentration for awakening and 1.8 micro gram/ml for arousal state after propofol infusion with an epidural block. Our criteria that the patients open their eyes after having their names called in a normal voice was comparable to Wessen's reported arousal state.

To eliminate the confounding effects of pharmacokinetic and pharmacodynamic variability in the response of patients to a stimulus, constant blood concentrations and blood to effect site equilibration are required. In our study, we determined the awakening concentration after blood-effect site equilibration, EC⁵⁰eq. The EC⁵⁰eq was independent of the duration of propofol infusion. This is in contrast to the EC⁵⁰, which was significantly influenced by infusion duration. To obtain the EC⁵⁰eq, we used descending sequences of propofol concentrations for 15-min intervals. To simulate the effect site concentration, we added a pseudocompartment of effect site to our pharmacokinetic model. The concentration time course was derived from the plasma concentration by using Ke0 (the rate constant for elimination from the effect site) by solving a first-order differential equation, which was first proposed by Sheiner et al. [11]We used the Ke0 of propofol previously published by Schuttler et al.* In Figure 1, we show target plasma concentration, predicted plasma concentration, and the predicted effect site concentration of our study design with descending sequences of propofol after long infusions. The time interval that we used was long enough to allow equilibration between blood and effect site when the target propofol concentration was more than 1 micro gram/ml.

We used the pharmacokinetic/pharmacodynamic models to perform computer simulations of plasma and effect site concentrations versus infusion time (Figure 4) and probability of no response to verbal command versus infusion time (Figure 5) after abrupt discontinuation after various durations of propofol infusion maintained at 3 micro gram/ml. We assumed that the probability of no response to verbal command had a sigmoid relation with the effect site concentration, which was numerically expressed as a Hill equation. [12]We used the data from groups 1B and 2B to estimate the pharmacodynamic response of propofol from the effect site concentration for the constants in the equation, which were C⁵⁰(effect site concentration at 50% of maximum response) and r (steepness factor = EC⁹⁵/EC⁵⁰).

These simulations show that the difference between plasma and effect site concentration is significant, especially 3–10 min after termination of propofol infusion, and that the shorter the duration of propofol infusion, the faster the decrease in effect site propofol concentration; the longer the duration of propofol, the longer the awakening time. However, the time for 50% of patients to awaken is less than 10 min, even after a 10-h infusion at a 3 micro gram/ml propofol concentration.

Although the EC⁵⁰eq was independent of the duration of the propofol infusion, the EC⁵⁰was significantly influenced by the infusion duration in our study. Patients undergoing prolonged infusion required longer emergence times than those undergoing short infusion. Furthermore, our patients awoke at significantly higher blood propofol concentrations with prolonged infusion than with short infusion. In those who had abrupt discontinuation, the relation between the time and blood propofol concentration at emergence was influenced by the duration of propofol infusion. We assumed that the effect site propofol concentration at awakening would be the same EC⁵⁰eq value (2.2 micro gram/ml) whether the abrupt discontinuation happened after short or long propofol infusion. According to this assumption, we predicted each ke0 after short or long propofol infusion using the least-square method by applying the following data: awakening time, awakening propofol concentration, mean propofol infusion time, mean maintenance target propofol concentration, and mean propofol blood concentration at abrupt discontinuation. The predicted ke0 values were 0.26 after short infusion and 0.21 after long infusion. Considering the use of the maintenance propofol concentration as the predicted one, the use of fentanyl with propofol during surgery, and individual pharmacokinetic deviation, these predicted Ke0 values were similar to the value of ke0 = 0.239 min - 1 (t1/2e0 = 2.9 +/- 2.2 min) calculated by Schuttler et al.* using the electroencephalogram median frequency.

The propofol blood concentrations at which 50% of patients did not respond to verbal command (Cp⁵⁰awake) determined by Smith et al. [13]and Chortkoff et al. [14]were 3.3 and 2.69 micro gram/ml, respectively. Vuyk et al. [15]reported that the propofol concentrations at which 50% of the patients showed loss of eyelash reflex and loss of consciousness were 2.07 and 3.4 micro gram/ml. In those studies, the propofol infusion scheme was designed to allow blood-brain equilibration at each step of the target propofol concentration (10–15 min of steady state predicted concentration with computer-assisted continuous infusion). Our EC⁵⁰eq was lower than that in their reports, which is consistent with differences in our study design:(1) a different endpoint for awakening, (2) a different method for defining awakening concentration, (3) effect of fentanyl, and (4) postoperative pain. Our response criterion was the patients opening their eyes after hearing their names called twice in a normal voice, a definition of arousal state commonly used in clinical anesthesia. Although the concentration of fentanyl after surgery was < 1 ng/ml, the fentanyl concentration might decrease the awakening propofol concentration. According to the report by Smith et al., [13]1 ng/ml of fentanyl concentration decreased the EC⁵⁰eq value by approximately 1 micro gram/ml compared with the Cp⁵⁰awake value for propofol alone. In our study, the mean fentanyl concentration of each group during all measurements was 0.88–0.98 ng/ml, and there were no significant differences between groups. Our EC⁵⁰eq value in the presence of fentanyl thus was similar to that reported by Smith et al. [13]Therefore, we surmise that as the effect of fentanyl on awakening concentration was equivalent across all groups, it was not responsible for any differences seen in the EC⁵⁰or EC⁵⁰eq values between the groups.

In conclusion, we confirmed the EC⁵⁰after abrupt discontinuation of propofol at the end of surgery and the EC⁵⁰eq value obtained after allowing blood-brain equilibration at each decreasing target propofol concentration. Although the EC⁵⁰after short-term propofol infusion was less than that after long-term infusion, EC⁵⁰eq value was independent of duration of propofol infusion time.

*Schuttler J, Schwilden H, Stoeckel H: Pharmacokinetic-dynamic modeling of Diprivan [Abstract]. Anesthesiology 1986; 65:A549