Although beta blockers have been used primarily to decrease unwanted perioperative hemodynamic responses, the sedative properties of these compounds might decrease anesthetic requirements. This study was designed to determine whether esmolol, a short-acting beta 1-receptor antagonist, could reduce the propofol concentration required to prevent movement at skin incision.
Sixty consenting patients were premedicated with morphine, and then propofol was delivered by computer-assisted continuous infusion along with 60% nitrous oxide. Patients were randomly divided into three groups, propofol alone, propofol plus low-dose esmolol (bolus of 0.5 mg/kg, then 50 micrograms.kg-1.min-1), and propofol plus high-dose esmolol (bolus of 1 mg/kg, then 250 micrograms.kg-1.min-1). Two venous blood samples were drawn at equilibrium. The serum propofol concentration that prevented movement to incision in 50% of patients (Cp50) was calculated by logistic regression.
The propofol Cp50 with nitrous oxide was 3.85 micrograms/ ml. High-dose esmolol infusion was associated with a significant reduction in the Cp50 to 2.80 micrograms/ml (P < 0.04). Propofol computer-assisted continuous infusion produced stable serum concentrations with a slight positive blas. Esmolol did not alter the serum propofol concentration. No intergroup differences in heart rate or blood pressure response to intubation or incision were found.
Esmolol significantly decreased the anesthetic requirement for skin incision. The components and mechanism of this interaction remain unclear. A simple pharmacokinetic interaction between esmolol and propofol does not explain the Cp50 reduction. These results demonstrate an anesthetic-sparing effect of a beta-adrenergic antagonist in humans under clinically relevant conditions.
Beta-Adrenergic receptor antagonists have been in clinical use for more than 30 yr. These drugs have central nervous system depressant activity in animals [1–4]and anxiolytic effects in humans. [5,6]More than 25 yr ago, Miller et al. suggested that drugs that affect central catecholamine release may alter anesthetic requirements. Sympathomimetic drugs that enter the central nervous system, such as ephedrine, mephentermine, and amphetamine, can increase halothane minimum alveolar concentration (MAC), whereas drugs that do not enter the central nervous system, such as isoproterenol, norepinephrine, and epinephrine, do not alter MAC directly. [8–10]Antagonism of catecholamines in the brain or spinal cord might decrease anesthetic requirements. Although beta-adrenergic receptor antagonists are used every day in the operating room, no published studies have examined the effect of these drugs on anesthetic requirements for skin incision in humans.
Minimum alveolar concentration of volatile agents is defined as suppression of movement to incision in 50% of patients. [11,12]A similar definition has been proposed for intravenous anesthetics: the minimum effective plasma concentration, or Cp50. Anesthetic drug interactions may be measured by reduction in MAC or Cp50of the primary anesthetic by a second drug. For example, intravenous opioids markedly decrease isoflurane and desflurane MAC. [14–16]Measurement of drug interactions between intravenous anesthetic agents has proved difficult due to several technical issues. Steady-state plasma concentrations must be maintained for each drug. In addition, the individual rates of equilibration between the plasma and effect site compartments must be considered. [17–21]Computer-assisted continuous infusion (CACI) devices can overcome several of these concerns by providing stable plasma levels of intravenous anesthetic agents that can be assumed to approach steady-state conditions after a reasonable interval. Recently, Smith et al. used CACI to examined the reduction in propofol Cp50by fentanyl.
Esmolol is an ultra-short-acting, cardioselective beta1-adrenergic receptor antagonist [23,24]that is effective in blunting adrenergic responses to several perioperative stimuli, including laryngoscopy with intubation, [25–27]intraoperative events, emergence, and extubation. Its use has been promoted for minimizing the deleterious effects of intraoperative hypertension and tachycardia on myocardial oxygen consumption. [25,30]Esmolol has been proposed as an alternative to alfentanil during propofol-nitrous oxide anesthetics in patients receiving neuromuscular blocking agents. 
This study was designed to determine whether esmolol infusions could decrease the Cp50of propofol under conditions approximating steady state.
Materials and Methods
Sixty consenting male and female patients, ages 18 to 70 yr, who were classified as American Society of Anesthesiologists physical status 1 to 3 and scheduled for elective surgery were studied. Exclusion criteria included a history of an allergic reaction to any study medication; advanced hepatic, renal, or cardiac dysfunction; long-term opioid, ethanol, sedative, or beta-blocker use; poorly controlled asthma, diabetes, or hypertension; or weight more than 150% of ideal body weight.
After premedication with morphine sulfate (0.1 mg/kg to a maximum of 10 mg), patients were randomly assigned to one of three groups: propofol, propofol plus low-dose esmolol (bolus 0.5 mg/kg, then 50 micro gram [center dot] kg-1[center dot] min-1), or propofol plus high-dose esmolol (bolus 1 mg/kg, then 250 micro gram [center dot] kg-1[center dot] min-1). On arriving in the operating room, awake physiologic variables were noted. Continuous esmolol infusions by infusion pump (Baxter Model AS40a; Baxter Health Care, Deer-field, IL) were started before induction and continued at the given rate until the study was complete. After preoxygenation, anesthesia was induced by intravenous infusion of propofol using a CACI pump set at an initial effect-site target of 5.5 micro gram/ml. The CACI device used a three-compartment pharmacokinetic model with the kinetic data set of Smith et al. When patients lost consciousness, succinylcholine (1.5 mg/kg) was administered to facilitate endotracheal intubation. Mechanical ventilation to normocapnia with nitrous oxide (60%) was initiated. After intubation, the new predetermined target plasma propofol concentration was entered into CACI. Body temperature was maintained at more than 35.5 degrees Celsius.
Target propofol concentrations were assigned using a modification of Dixon's method. Within each group, a concentration of propofol was estimated for the first patient. If that patient moved in response to skin incision, then the next patient in that group received a 10% increase in the target propofol concentration. If that patient did not move, then the next patient in that group received a 10% reduction in target propofol concentration. Using this method, the observer was not blinded to treatment group. Esmolol bolus and infusion rates were chosen to rapidly attain fivefold different plasma concentrations representing the upper and lower limits of the suggested therapeutic range.
Steady-state conditions were approximated by allowing at least 9 min to elapse before incision after the computer-predicted target propofol concentration had been reached. Full reversal of neuromuscular blockade was confirmed with a peripheral nerve stimulator. A skin incision of at least 2 inches was made. A positive response at incision was defined as movement of limbs, head, or body within 60 s. Coughing, chewing, or swallowing were not considered movement. Movement was not assessed until steady-state conditions had been met. However, spontaneous movements after equilibration were also considered positive.
Venous blood samples were drawn from the arm contralateral to the drug infusions 5 min after the predicted target propofol concentration had been reached and again at incision. Samples were immediately placed on ice, centrifuged as soon as possible, and frozen at -80 degrees C until assay. Propofol plasma concentrations were determined by high-performance liquid chromatography, as described by Plummer, except that pentane was used as the extraction solvent. The coefficient of variation for analysis of known propofol concentrations was 6.2% in the range of 100 to 5,000 eta g/ml. The lower limit of sensitivity of the assay was 2 eta g/ml.
Data are represented as means +/- SD, or as mean +/- 95% confidence interval (CI). Statistical significance was set at a probability value less than 0.05. The Cp50of propofol was calculated by two methods. First, the maximum likelihood solution to a logistic regression model (see appendix Table 3) was used. Second, the Cp (50) was calculated from a limited data set consisting of only independent crossover pairs within each group by the up-down method of Dixon. The performance error, median performance error, and median absolute performance error of the CACI infusion were calculated using the serum propofol concentration at skin incision, as previously described. [22,34]One-way analysis of variance (ANOVA) was used to compare serum propofol concentrations and median performance error. Two-way repeated measures ANOVA was used to evaluate sequential physiologic data. When indicated, individual group means from ANOVA were compared with a post hoc Tukey's Honestly Significant Difference test. A Kruskal-Wallis one-way ANOVA was used to analyze the median absolute performance error of the CACI device, which was not normally distributed.
Of the sixty patients studied, 41 were women and 19 were men, with an average age of 42 +/- 13 yr (range, 21–72 yr) and an average weight of 77 +/- 17 kg (range, 48–120 kg). Ninety-five percent of patients were ASA physical status 1 or 2. There were no significant differences between groups with regard to demographic or surgical data. Most of the procedures were performed by general surgery (60%), with gynecological cases accounting for approximately 25%.
An average of 30 min elapsed between premedication and induction of anesthesia. The mean time from induction to incision or movement was 28 +/- 11 min (range, 9–73 min). The mean time between the two blood samples was 11 +/- 1 min (range, 3–39 min). Eight patients moved before incision after at least 9 min at the computer-predicted target propofol concentration. These patients were considered positive for movement. Incision was made at the first blood sample in three patients. With these 11 persons, the second blood sample was collected within 3 to 5 min. No significant differences in time from induction to incision, in time between blood samples, or in number of persons moving before incision were found among the three treatment groups.
The esmolol infusion rate and the natural log of serum propofol concentration at incision were significant factors in the logistic regression model for predicting movement to incision (Figure 1). Age, height, weight, ASA physical status, and time from induction to incision were not significant factors. Based on the maximum likelihood solution to a logistic regression model, the Cp50of propofol for skin incision was 3.85 micro gram/ml in the presence of 60% nitrous oxide and morphine premedication (Table 1). Continuous high-dose esmolol infusion significantly reduced the propofol Cp50by 26%, to 2.85 micro gram/ml (P < 0.04) by logistic regression analysis. No significant change in propofol (Cp50(3.50 micro gram/ml) was seen after adding a low-dose esmolol infusion. Pre- and postincision heart rates were significant predictors of movement to incision in the logistic regression model. However, when mean heart rate in each group at each time point were included in the model, no significant change in the 50% response line was observed. Therefore, these two factors were eliminated from the logistic regression model and equation (Appendix 1 Table 3).
Propofol Cp50can also be computed by the method of Dixon using only independent crossover pairs within each group (Table 1). By this method, an identical Cp50was found using either measured serum concentration or the computer-predicted target propofol concentration at incision. Both low- and high-dose esmolol infusions significantly reduced the propofol Cp50in a dose-dependent manner. Based on the CACI-target propofol concentrations, a 50% reduction in Cp (50) to 1.95 micro gram/ml was predicted in the presence of an infusion of 250 micro gram [center dot] kg-1[center dot] min-1esmolol.
The average percentage difference between the two plasma propofol samples in each patient was 0 +/- 15%(range, +32% to -33%). The overall bias or median performance error of the CACI device was positive at 14%(mean, 17%; 95% CI, 8–27). The overall accuracy or median absolute performance error was 20%(mean, 27%; 95% CI, 20–34). Adding esmolol did not significantly alter the bias or the accuracy of the computer-controlled infusion (Figure 2, Table 2). In this study, a power of 80% was achieved to detect a difference of 25% in the median performance error and 20% in median absolute performance error between control- and esmolol-treated groups. No significant correlation was observed between the bias or accuracy of CACI and movement within individual patients. Within the low-dose esmolol group, a statistically significant positive correlation was found between the bias and serum propofol concentration (y = 54x - 149, r = 0.86, P < 0.001) in patients who moved at incision (Figure 2(b)). No age-specific differences in the bias or accuracy of the CACI infusion were found by ANOVA with esmolol.
(Figure 3) shows the physiologic response to endotracheal intubation and skin incision. Heart rate and mean arterial pressure at four specific times during surgery were normalized to each patient's awake control value. A significant increase in heart rate (18–24%; P < 0.01) and mean arterial pressure (17–23%; P < 0.01) was found within each group after endotracheal intubation but not after incision. After equilibration at the target serum propofol concentration, mean arterial blood pressure decreased by 14% to 17%(P < 0.05) within each group. However, no change in baseline heart rate was seen under these conditions. Two-way repeated-measures ANOVA did not detect a significant difference among the three treatment groups with regard to any measured physiologic variable at any of the four time points.
Transient wheezing developed in one patient at incision, but it resolved spontaneously. Four patients received ephedrine within 10 min of incision for mean arterial pressures less than 70% of awake baseline. Ephedrine use was evenly distributed across all three groups and was not correlated with movement to incision. One outpatient in the high-dose esmolol group was admitted to the hospital overnight due to persistent low blood pressure in the recovery room. Although the patient was asymptomatic and had a blood pressure within 20% of preoperative baseline, he was admitted for logistical and institution-specific concerns. This patient was discharged from the hospital the next morning with no further postoperative sequelae. No patient reported any intraoperative recall during the postoperative interview.
This study shows that esmolol, a beta1-adrenergic receptor antagonist, can reduce the Cp50of propofol for skin incision in the presence of 60% nitrous oxide and morphine premedication. A 26% reduction in the Cp50of propofol was demonstrated by logistic regression analysis at a clinically relevant esmolol infusion rate of 250 micro gram [center dot] kg-1[center dot] min-1. Logistic regression was selected as the primary method of analysis because all available data were used. Previous studies [25–31]investigating control of adrenergic responses to perioperative stimuli have overlooked the anesthetic-sparing activity of esmolol. Many of these studies incorporated neuromuscular blocking agents into their study design, preventing detection of movement. A single bolus or noncontinuous drug administration also predominated throughout these studies.
No significant reduction in propofol Cp50was found at the lower esmolol infusion rate. The lack of significance at the lower esmolol infusion rate probably reflects the increased variability within this group and subsequent loss of statistical power. A larger sample size may have identified a significant difference. When the Cp50was calculated by the Dixon method with a more limited data set, esmolol produced a dose-dependent reduction that was significant at both infusion rates (Table 1). The control Cp50was constant regardless of the method of analysis.
In this study, the Cp50of propofol in tracheally intubated patients with 60% nitrous oxide after premedication with morphine was 3.85 micro gram/ml. This result corresponds closely with that of Shafer et al., who found a Cp50from venous samples of 3.40 micro gram/ml for propofol with nitrous oxide (70%) after meperidine premedication (1 mg/kg). Davidson et al. reported the propofol Cp (50) at 4.5 micro gram/ml using CACI with nitrous oxide (67%) in nontracheally intubated patients after temazepam premedication. A wide discrepancy exists in reported Cp50values for propofol/nitrous oxide anesthetic, with values ranging from 1.55 to 5.36 micro gram/ml. [35–39]Comparing our results with other studies in the literature is difficult because of differences in anesthetic technique, differences in endotracheal intubation, lack of computer-controlled drug infusions, arterial versus venous sampling, and measured response to skin incision. For example, significantly lower Cp50values, 1.66 and 2.5 micro gram/ml, were reported in nontracheally intubated patients receiving a continuous propofol infusion and nitrous oxide (67%), either with or without a opioid premedication, respectively. Opioid premedication and nitrous oxide were included in the experimental design to replicate the clinical conditions under which this interaction was first observed (Johansen JW, unpublished observations). The Cp50of propofol as the sole anesthetic for skin incision has been reported to be 8.1 micro gram/ml in nontracheally intubated patients after a benzodiazepine premedication and 15.2 micro gram/ml micro gram/ml in tracheally intubated, nonpremedicated patients. 
Computer-assisted continuous drug infusions provide stable plasma drug concentrations so that intravenous drug interactions can be explored. Equilibration between the plasma and effect-site approximating steady-state conditions can be assumed after a reasonable interval. [20,22,40–42]Under these conditions, a plasma drug concentration should be equal to that in the effect site or may represent a constant, stable fraction of that concentration. The half-time (t1/2keo) for equilibration between the plasma and effect site for propofol has been reported as 2.9 min. * In each patient, the two venous blood samples at equilibrium differed by 0 +/- 15% from their average. This suggests that stable propofol plasma concentrations were achieved under our experimental conditions. The first blood sample was taken at least 5 min after the computer-predicted target serum concentration had been reached. An average of 11 min or at least three time constants elapsed between the two blood samples, suggesting that more than 94% equilibration between plasma and effect site had occurred in most patients.
The overall CACI precision or median absolute performance error was 20%, well within the 10–40% accuracy reported by others using similar population pharmacokinetic estimates. [22,34,36,43,44]Adding esmolol did not affect the precision or bias of propofol delivery by the CACI system (Table 2). Our study was designed to examine a pharmacodynamic interaction between esmolol and propofol, not a pharmacokinetic interaction. However, this study had adequate power (80%) to detect a difference of 25% in bias and 20% in accuracy between control and esmolol-treated groups using measured serum propofol concentrations. If esmolol altered the bias or accuracy of propofol delivery by CACI and plasma concentrations were stable, no change in Cp50would have been found by our methods because only measured serum propofol concentrations were used in the Cp50calculation.
The significance of the positive correlation between an increasing bias and serum propofol concentration in the low-dose esmolol group, predominately in the patients who moved at incision, is unclear (Figure 2(b)). This trend was not found in either control, high-dose esmolol groups, or in patients who did not move within the low-dose esmolol group. An example of increasing CACI bias has been described during propofol and alfentanil infusions. 
Venous propofol samples were used in this study. These have been shown to be reliable after an initial equilibration period. After a 20- to 25-min infusion, differences between arterial and venous propofol concentrations were shown to be minimal. Increased variability in propofol measurements after venous sampling has been reported in one study. However, subclinical infusions of propofol were used in this study. Davidson et al. found indirect evidence of incomplete venous mixing or pulmonary sequestration during propofol CACI propofol. A reduction in CACI bias from + 21.4% to -2% was found in this study after stopping the propofol infusion for 90 s. Although arterial versus venous blood sampling remains controversial, our results show that reliable and consistent data can be obtained from venous propofol samples. It remains possible that the slight positive bias we observed in all groups may be an artifact of venous sampling.
Esmolol infusions did not produce a significant change at either infusion rate in heart rate or blood pressure. Laryngoscopy and endotracheal intubation, but not incision, caused a small increase in heart rate and mean arterial pressure in all groups of patients. Opioid premedication has been shown previously to eliminate any clinically significant reduction in heart rate or blood pressure associated with intubation when a single bolus dose of esmolol was used. In this study, no hemodynamically significant bradycardia was observed. Although one outpatient in the high-dose esmolol group was admitted overnight to the hospital, this was related to logistic and institution-specific concerns and not to medical necessity.
Twenty-seven years ago, animal studies with acute and chronic propranolol administration showed no change in halothane MAC in dogs and no change in brain content of biogenic amines in rats. Investigation of biogenic amines and anesthetic requirements stopped nearly 20 yr ago. In the current study, the mechanism by which esmolol decreased anesthetic requirements for skin incision is unknown. Our results show that esmolol does not alter plasma concentration of propofol by CACI, suggesting that a simple pharmacokinetic interaction between esmolol and propofol does not account for the Cp50reduction. However, the components of this interaction remain unclear. Esmolol could potentiate the effects of propofol, nitrous oxide, or opioid premedication. Esmolol has very little sedative effect, no analgesic activity, and, by itself, is not an anesthetic agent. [48–50]The hemodynamic effects of esmolol are thought to be mediated by blockade at peripheral, beta-adrenergic receptors. The low-potency, low-lipid solubility and rapid metabolism within the blood stream [48,49]do not exclude a central site for esmolol action. Some evidence for potentiation of the analgesic activity of opioids by beta-blockers exists in the literature. Stanley et al. found a 25% decrease in the fentanyl dose for loss of consciousness in patients receiving long-term propranolol treatment. The significance and clinical utility of this drug interaction will depend on maximizing the Cp50reduction. Optimal conditions should be established with respect to essential components, sequence of interaction, time course, and dose response. It will also be necessary to determine whether the anesthetic-sparing effects of esmolol result from competitive antagonism at the beta1, receptor or from a nonspecific drug effect.
We found that continuous esmolol infusions can significantly decrease anesthetic requirements for skin incision during balanced anesthesia with propofol, nitrous oxide, and morphine. Propofol CACI was accurate and stable, as measured by two venous blood samples. Esmolol did not affect the measured serum propofol concentration delivered by CACI, suggesting that esmolol does not significantly alter the pharmacokinetic properties of propofol. Adding esmolol to this balanced anesthetic technique resulted in no significant change in heart rate or blood pressure. Beta-adrenergic antagonists may represent a novel class of drugs that can modify anesthetic requirements in humans.
(*)Schuttler J, Schwilden H, Stoeckel H: Pharmacokinetic-dynamic modeling of diprivan (abstract). Anesthesiology 1986; 65:A549.