Induction, Recovery, and Safety Characteristics of Sevoflurane in Children Undergoing Ambulatory Surgery: A Comparison with Halothane

With approval from the federal regulatory agencies in the United States and Canada and the ethics committees of all five participating institutions, written informed consent was obtained from the parents of 375 ASA physical status 1 and 2 children, aged 1–12 yr, who were scheduled for elective ambulatory surgery (including genitourinary, superficial lower abdominal, and plastic or superficial orthopedic surgery) lasting 0.5–2 h. Seventy-five children were enrolled in each center. Within each center, the children were randomly assigned to receive either sevoflurane or halothane in a 2:1 ratio using a computer-generated randomization schedule after consent was obtained.

All children were fasting and unpremedicated. After application of standard monitors, including an electrocardiogram, pulse oximeter, and noninvasive blood pressure, anesthesia was induced with 60% N²O and 40% O²followed by stepwise increases in the inspired concentrations of either sevoflurane (1.5–2.0% increments) or halothane (0.5–1.0% increments) every three to four breaths. The maximum inspired concentration of sevoflurane was 7% and that of halothane was 4.3%. Sevoflurane was delivered using an Ohmeda SevoTec-3 vaporizer (Maruishi, Osaka, Japan). Anesthesia was delivered via a nonscented mask through a Mapleson D or F circuit (Bain or Ayre's t piece) in four centers and a circle system with a carbon dioxide absorber filled with soda lime in the fifth center. Fresh gas flows during induction of anesthesia were 7–10 l *symbol* min sup -1 for all children.

Inspired and exhaled gases were sampled through a 2-inch catheter positioned either at the entrance to one of the nares or in the oropharynx. [10]The anesthetic concentrations and carbon dioxide partial pressure in the sample gas were analyzed using a Datex Capnomac airway gas monitor (Helsinki, Finland). The gas monitor was calibrated using a commercial standard gas before each study.

Immediately after induction of anesthesia, intravenous access was established. Whole blood (2 ml) was collected in plastic syringes containing heparin and transferred to plastic tubes. The contents were centrifuged within 30 min of collection, and the plasma supernatant was pipetted into plastic tubes that were sealed and stored at -20 degrees Celsius until analysis for the inorganic fluoride concentration.

Lactated Ringer's solution was administered during anesthesia at a maintenance rate appropriate for the child's age and fasting interval. Anticholinergic medication was administered only for the treatment of bradycardia (defined as a heart rate < 100 beats/min).

A single-dose caudal block was administered after induction of anesthesia to those children in whom it was clinically indicated. Bupivacaine (without epinephrine) in a concentration and volume appropriate for the level of desired blockade (maximum dose of 2.5 mg *symbol* kg sup -1) was used. Wound infiltration with local anesthetics was permitted as an alternative.

During the maintenance period, the children breathed spontaneously through a face mask and an anesthetic circuit. Fresh gas flows were adjusted to 3–6 l *symbol* min sup -1 for children aged 1–7 yr and 6–10 l *symbol* min sup -1 for those aged 8–12 yr. The inspired anesthetic concentration was adjusted to produce an end-tidal concentration of 1.3 x MAC of the inhalational anesthetic [4,11]in 60% N²O and oxygen. A heat and moisture exchanger was interposed between the elbow of the circuit and the fresh gas sleeve in the Ayre's t piece or at the distal end of the Bain circuit at the discretion of the staff.

The algorithm for the control of the depth of anesthesia follows: In the presence of light anesthesia (heart rate or systolic blood pressure > 30% of baseline, movement, swallowing, or tearing), the inspired concentration of the anesthetic was increased appropriately. If these increases resulted in clinically unacceptable cardiovascular or respiratory depression, the inspired concentration was decreased to an acceptable value, and only then was fentanyl administered. Fentanyl was not given within 30 min of the conclusion of surgery. If vital signs were within the 20% of baseline limits, the inspired anesthetic concentration was adjusted to produce an end-tidal concentration of 1.0 x MAC of the inhalational anesthetic (excluding nitrous oxide) during the maintenance period.

For a minimum of 10 min before the conclusion of surgery, the inspired anesthetic concentration was adjusted to an end-tidal concentration of 1.0 x MAC of the inhalational anesthetic (excluding nitrous oxide). [4,11]At the time of placement of the last suture, all anesthetic agents were discontinued, and 100% Oxygen²was administered for 1–2 min. In the recovery room, vital signs (including heart rate, blood pressure, transcutaneous hemoglobin oxygen saturation, respiratory rate, and temperature) were assessed as described below. All recovery and discharge data were collected by a registered nurse who was blinded to the treatment assignment. Sixty minutes after discontinuation of anesthesia, a second whole blood sample (2 ml) was obtained from the indwelling intravenous cannula, and the plasma was stored as described above. Discharge criteria from the recovery room included restoration of stable vital signs, a modified Aldrete recovery score greater or equal to 8, absence of pain or bleeding (objective pain/discomfort less or equal to 6), orientation to name and place if appropriate for age, and tolerance of oral clear fluids.

During the recovery period, pain was managed with rectal acetaminophen (10–25 mg *symbol* kg sup -1), intramuscular codeine (1–1.5 mg *symbol* kg sup -1), or intravenous morphine sulfate (0.05–0.10 mg *symbol* kg sup -1) as needed. Fentanyl (1–2 micro gram *symbol* kg sup -1) was administered intravenously if required postoperatively.

Demographic data, including age, weight, and type of surgery, were recorded for all subjects. Vital signs, including heart rate, systolic and diastolic blood pressures, and respiratory rate were recorded at the time of consent and 1 min before induction of anesthesia (baseline). Vital signs, inspired and expired concentrations, end-tidal carbon dioxide tension, hemoglobin oxygen saturation, and temperature were recorded every minute until incision, at 1-min intervals for the first 5 min after incision, and every 5 min until the conclusion of surgery.

Induction and emergence characteristics, including coughing, laryngospasm, breathholding, nausea, vomiting, secretions, bronchospasm, excitement, and any other unanticipated events, were recorded.

The induction interval (time from application of the face mask until loss of the eyelash reflex), duration of anesthesia (defined as the interval from application of the mask until discontinuation of the anesthetics), and duration of surgery (defined as the interval from skin incision until the final skin suture was inserted) were recorded. The times from the discontinuation of anesthesia until the child responded appropriately to commands or demonstrated purposeful movement and until the first postsurgical analgesic was administered were recorded by a nurse who was blinded to the treatment administered.

All postoperative vital signs and discharge criteria were recorded by a nurse who was blinded to the treatment administered. Vital signs, a modified Aldrete recovery score (Table 5appendix 1), and the objective pain discomfort scale (Table 6appendix 2) were recorded on arrival in the recovery room and then every 10 min for 60 min. Before discharge from hospital, each child was questioned about any recall of events during the intraoperative period.

Each patient and/or parent was telephoned 24 h after discharge from hospital by a research nurse or assistant who was unaware of the anesthetic administered. The parents were questioned regarding the presence of fever, nausea and vomiting, coughing, dizziness, and drowsiness after discharge from hospital. Depending on the child's age, either the child or a parent was asked the following:(1) whether they would request the same anesthetic again, (2) whether they recalled how they were put to sleep, (3) whether they remembered going to sleep, and (4) whether the anesthetic had an odor (If the answer was yes, they were asked whether the odor was mild, strong, or very strong.).

The MAC *symbol* h for each child were calculated based on the area under the end-tidal anesthetic concentration versus time curve (using the trapezoidal rule) and were normalized using age-adjusted MAC values. [4,11].

All analyses were performed with the SAS (Statistical Analysis System) procedures GLM and FREQ. Descriptive statistics were used to summarize age, weight, and gender. Continuous data were compared using a two-way analysis of variance model with effects for the investigator, anesthetic, and interaction between investigator and anesthetic. Nominal data were compared using Cochrane-Mantel-Haenszel test or Fisher's exact test. Data are mean+/-SD except where indicated. Statistical significance was accepted for P < 0.05.

Data from 374 of the patients were analyzed; one child who received halothane was eliminated from the analysis because of severe airway obstruction and hypoxia caused by a mucous plug. Demographic data, including the duration of anesthesia and surgery, were similar for the two treatment groups (Table 1). Demographic data did not differ significantly among the five centers. The distributions of race and gender were similar in the two treatment groups.

The time to loss of the eyelash reflex during sevoflurane anesthesia was minimally but significantly faster than during halothane anesthesia (Table 2). The time to loss of the eyelash reflex did not vary among the five centers, nor did it depend on the type of circuit used. During the interval between application of the mask and skin incision, the maximum end-tidal sevoflurane concentration (normalized for MAC), 1.7+/-0.58, was significantly less than the maximum halothane concentration (normalized for MAC), 2.7+/-0.57 (P < 0.001). The average end-tidal concentration of sevoflurane (normalized to its MAC) during this interval, 1.10+/-0.24, was also significantly less than the average halothane concentration (normalized for MAC), 1.4+/-0.24 (P < 0.001).

The total mean (+/-SD) MAC *symbol* h exposure to sevoflurane was significantly less than the exposure to halothane (P < 0.013)(Table 1). The duration of surgery was less than 10 min in 4%, or 10 of 250, children who received sevoflurane and 6%, or 8 of 125, children who received halothane (P = NS). Although the duration of anesthesia was less than stipulated in the protocol, these data were included in the analysis on the basis of "intent to treat."

The times to emergence and early recovery after sevoflurane anesthesia were significantly more rapid than after halothane anesthesia (Table 2). Recovery milestones, including a modified Aldrete score greater or equal to 8 (Figure 1) and time to orientation after sevoflurane, were reached more rapidly than after halothane (P < 0.001;Table 2). Orientation could be evaluated only in a subset of children, because some were too young to cooperate (Table 2). Suitability for discharge after intake of clear fluids could be evaluated only in a subset of children, because some of the children refused to ingest fluids (Table 2). Objective pain/discomfort scores at 1, 10, 20, and 30 min after arrival in the recovery room after sevoflurane anesthesia were significantly greater than those after halothane (P < 0.001) but were similar for the two groups at all times thereafter.

A regional block was provided with bupivacaine in 92% of patients in the sevoflurane group and 90% in the halothane group. Lidocaine was used for regional blockade in nine (3.6%) in the sevoflurane group and four (3.2%) in the halothane group. Fentanyl (1.6 micro gram/kg) was administered during anesthesia to four children (1.6%) in the sevoflurane group and three children (2.4%) in the halothane group more than 30 min before the end of surgery. In the recovery room, supplementary analgesia in the form of acetaminophen alone was administered to 20% of the children in the sevoflurane group and 31% in the halothane group. Intravenous morphine (with or without supplemental acetaminophen per rectum) was administered to 30% of those in the sevoflurane group and 18.4% of those in the halothane group (P < 0.018). Codeine and fentanyl were administered to fewer than 1.6% of the children in each of the two treatment groups.

Side effects, including airway reflex responses, vomiting, and excitement, were similar during both induction of and emergence from anesthesia with sevoflurane and halothane (Table 3).

All study-related side effects listed in Table 4were considered mild or moderate in intensity by the investigators. Cardiovascular events, including bradycardia, occurred significantly less frequently with sevoflurane than with halothane anesthesia (Table 4). Seven had abnormal electrocardiogram results during the study that were considered by the investigators to be either possibly or probably related to the anesthetic: < 1%, or 2 of 250, in the sevoflurane group and 4%, or 5 of 125, in the halothane group (P < 0.04). The electrocardiographic abnormalities in the sevoflurane group consisted of premature ventricular contractions intermittently for 1 h, 40 min during maintenance anesthesia without evidence of hemodynamic instability in one child and 30 s of a nodal rhythm in a second child. In both instances, the arrhythmias resolved spontaneously. The abnormalities in the halothane group consisted of three children with premature ventricular contractions, one with bigeminy for 45 s and one with a nodal rhythm that lasted 2 min. None of the arrhythmias were associated with hypotension. Similarly, gastrointestinal side effects, including vomiting, occurred less frequently with sevoflurane than with halothane (Table 4). Agitation, defined as involuntary movement of one or more extremities during induction of or recovery from sevoflurane anesthesia, occurred 2.5 times more frequently than that during or after halothane anesthesia (P < 0.004). Six additional experiences (in five patients) were considered serious, although only two were considered study drug-related. One 3-yr-old boy in the sevoflurane group experienced apnea for 3 min after induction of anesthesia. The apnea occurred after a caudal block was administered but before skin incision. The lungs were ventilated manually for 3 min, after which time spontaneous ventilation resumed and the child completed the study. A 1-yr-old boy in the halothane group experienced severe obstruction of the airway that was thought to be due to thick secretions that became inspissated in a bronchus during induction of anesthesia. This event was associated with severe hypoxia, which required pharyngeal suctioning, tracheal intubation, and assisted manual ventilation. The data from this child were omitted from the analysis.

The plasma inorganic fluoride concentration measured 1 h after discontinuation of sevoflurane anesthesia was 10.3+/-3.5 micro Meter (maximum 23.2 micro Meter) compared with 2.1+/-1.7 micro Meter (maximum 12.9 micro Meter) measured 1 h after discontinuation of halothane.

When questioned before discharge from the recovery room, none of the children in either group experienced intraoperative recall.

In the 24-h followup questionnaire, the percentage of children or parents who would request the same anesthetic again was 96.3% in the sevoflurane-treated group and 95.6% of the halothane-treated group (P = NS). The percentage of children who remembered the induction technique was 64% in the sevoflurane group and 63.2% in the halothane group. A similar proportion of the children remembered an odor associated with the anesthetic: 61% in the sevoflurane group and 75% in the halothane group.

The results of this multicenter study indicate that the clinical characteristics of sevoflurane are similar to those of halothane in children undergoing ambulatory surgery. Early recovery after sevoflurane is more rapid than that after halothane, although late recovery and discharge time from the hospital were similar for the two anesthetics. Both anesthetics appeared safe, with a low incidence of side effects.

In the design of this study, we included three strategies in an attempt to deliver equipotent end-tidal concentrations of sevoflurane and halothane during induction of anesthesia. First, because the maximum deliverable inspired concentration of sevoflurane was 7%, we limited the maximum inspired concentration of halothane to 4.3%. This maximum inspired concentration for halothane was based in part on the relative MAC values of the two anesthetics, [4,11]pharmacokinetic data for sevoflurane and halothane in adults, [12]and an assumption that MAC multiples of these anesthetics were equipotent. When these data are taken into consideration, a 7% inspired concentration of sevoflurane and a 4.3% inspired concentration of halothane should yield equipotent end-tidal anesthetic concentrations during induction of anesthesia. Second, the stepwise increase in the inspired concentration of sevoflurane was 2–3 times that of halothane, corresponding to a 250% difference in their MAC values. [4,9]These stepwise increases in concentrations should have increased the depth of anesthesia similarly for the two anesthetics. Third, the range of fresh gas flows used facilitated rapid changes in the inspired concentrations of both anesthetics with the Mapleson D, F, and circle circuits.

Notwithstanding the above, recent preliminary data suggest that the difference between the rate of increase in end-tidal to inspired concentrations of sevoflurane and halothane in children may be less than that reported in adults. [12,13]In assuming the pharmacokinetics of these two anesthetics in children are similar to those in adults, it is possible that we biased our study design in favor of a more rapid equilibration of end-tidal to inspired concentrations of halothane than of sevoflurane.

Despite these strategies, induction of anesthesia as measured by the time to loss of the eyelash reflex was statistically but not clinically more rapid with sevoflurane than with halothane. Four possible explanations could account for this surprising observation. First, the rate of increase in the partial pressure of sevoflurane within that portion of the central nervous system that controls the eyelash reflex was more rapid than that of halothane. Second, the assumption that a MAC multiple of one inhalational anesthetic is equipotent to the same MAC multiple of another anesthetic may be erroneous. Third, anticipation that an inhalational induction with sevoflurane would be free of airway reflex responses may have prompted the investigators to increase the inspired concentration of sevoflurane more rapidly than halothane. Fourth, it is possible that the halothane vaporizers did not deliver the dialed concentration at the large fresh gas flows used, whereas the newer vaporizers for sevoflurane did. As a consequence, the delivered concentration of halothane may have been less than the dialed concentration, and induction with this agent may have been delayed.

The rate of emergence from inhalational anesthesia depends on the solubility of the anesthetic, the duration of exposure and concentration of anesthetic during the maintenance period, and the metabolism of the anesthetic. We found that emergence and early recovery were 25–50% more rapid with sevoflurane than they were with halothane. This is consistent with their relative blood solubilities and published recovery data. [6–8,14]Rapid emergence should parallel a rapid elimination of the anesthetic, provided the anesthetic exposure and rate of metabolism are comparable between the two anesthetics. We controlled the anesthetic exposure by restricting the anesthetic concentration during the maintenance period to [nearly equal] 1.0 X MAC when possible but, at the least, during the final 10 min of surgery in all children. However, the total exposure to sevoflurane in this study was 13% less than the exposure to halothane. This difference favors a more rapid recovery after sevoflurane but does not completely explain the magnitude of the differences in emergence and early recovery between sevoflurane and halothane.

We found that emergence and early recovery after sevoflurane anesthesia were significantly more rapid than after halothane, although the recovery scores 30 min or more after arrival in the recovery room and the times to discharge from the recovery room were similar. These data are consistent with published studies of sevoflurane and desflurane. [5,15,16]Our failure to recognize a difference between sevoflurane and halothane in the times to discharge is unexpected because we designed this study with strategies to minimize the variability in the anesthetic management during emergence among the five centers, including a strictly regimented anesthetic protocol during the final 10 min of the surgery, a nurse who was unaware of the treatment assignment to monitor recovery, identical criteria for discharge with both treatments, and discharge times both with and without tolerance for clear fluids. If sevoflurane facilitates a clinically significant more rapid discharge from the recovery room than halothane, our sample size should have been large enough and the study design adequate to detect such a difference.

The incidence of agitation and excitement during emergence from sevoflurane was almost threefold greater than the incidence after halothane anesthesia. The reason for the greater incidence of agitation after sevoflurane is not clear at this time. At least two hypotheses have been proposed. The first is that agitation is a central nervous system effect of this ether anesthetic. To date, there is little evidence to support increased cell membrane excitability during sevoflurane anesthesia, although electroencephalographic evidence of seizurelike activity has been reported in children. [17]The second is that agitation and excitement during emergence are manifestations of acute pain and anxiety when the anesthetic is rapidly eliminated. In this study, recovery from sevoflurane anesthesia was rapid, and the analgesia requirements in the recovery room after sevoflurane were greater than those anesthetized with halothane. Although these two observations are consistent with the second hypothesis, they do not establish a cause and effect relationship. To prove the second hypothesis, we would have to compare the incidence of agitation in two groups of children anesthetized with sevoflurane, one with and one without intraoperative analgesia. Further studies are warranted to clarify this issue and to determine whether this is an avoidable side effect of sevoflurane.

Sevoflurane is degraded in vivo by cytochrome P450 2E1 releasing inorganic fluoride. [18]Plasma inorganic fluoride concentrations in excess of 50 micro Meter are believed to predispose to nephrotoxicity after methoxyflurane, although this opinion has been questioned. [19]The plasma concentration of inorganic fluoride reaches a maximum value approximately 1 h after discontinuation of sevoflurane, [nearly equal] 28 micro Meter 1 h after 2.5 MAC *symbol* h sevoflurane in children. [20]In this study, the mean plasma concentration of inorganic fluoride recorded 1 h after 1.05 MAC *symbol* h sevoflurane was only 10.3+/-3.5 micro Meter (maximum value recorded was 23.2 micro Meter). These data indicate that inorganic fluoride is unlikely to pose a concern insofar as the use of sevoflurane in children undergoing ambulatory surgery is concerned.

Sevoflurane is adsorbed and degraded in the presence of soda lime (or Baralyme) to five compounds, the most common of which is compound A. [21]Compound A in large concentrations in rats has been associated with histologic changes in the kidney. [22]We chose not to measure the inspired concentration of compound A in this study for several reasons. First, degradation of sevoflurane in the presence of soda lime is a hydrolytic process that is time-dependent. Based on the anticipated brief duration of the anesthetic, it is unlikely that measurable concentrations of compound A would have been produced in the anesthetic circuit in this study. Second, the fresh gas flows used in our study were large (at least 3 l *symbol* min sup -1 during maintenance). This would have diluted the concentration of compound A in the inspiratory limb of the circle circuit. Third, the rate of production of compound A decreases in parallel with body surface area in children. [23] 

In summary, the results of this study indicate that the induction, recovery, and safety characteristics of sevoflurane in children undergoing ambulatory surgery are comparable to those of halothane.

The authors thank G. H. Besselaar & Associates, Princeton, New Jersey, for auditing the study, and the authors' research colleagues N. Sikich, B.Sc.N., R.N., K. Maloney, B.S.N., J. Norden, M.S.N., P. Schrader, R.N., M.N., and M. K. Nespeca, R.N., B.S.N., for their assistance in the conduct of this study.