Comparison of Desflurane with Isoflurane or Propofol in Spontaneously Breathing Ambulatory Patients

Abstract

Desflurane is a potentially useful anesthetic for ambulatory surgery, but it has had limited evaluation in spontaneously breathing patients.After the induction of anesthesia with propofol and laryngeal mask insertion, 90 patients were randomized to receive isoflurane (0.25%-1%), propofol (50-200 [micro sign]g [center dot] kg (^-1) [center dot] min^-1), or desflurane (1.4%-6%) for anesthetic maintenance. Respiratory complications were uncommon; only six patients coughed (three who received isoflurane, one who received propofol, and two who received desflurane), and no anesthetic produced significant respiratory depression. Purposeful movement was significantly more common with propofol (19 patients; 63%) compared with isoflurane (7 patients; 23%) or desflurane (2 patients; 6.7%), but no patient had recall. Emergence times were similar in the isoflurane, propofol, and desflurane groups (5.1 +/- 2.3, 5.6 +/- 3.1, and 4.4 +/- 1.4 min, respectively). Later recovery end points and pain and sedation visual analog scale scores did not differ among groups. Overall, 85 patients (94%) were free from postoperative nausea and vomiting. Desflurane produced few respiratory complications in spontaneously breathing ambulatory patients but offered no improvement in emergence or recovery compared with isoflurane. Propofol also did not reduce recovery times or side effects; however, it was more difficult to maintain an adequate depth of anesthesia. We conclude that neither desflurane nor propofol offered any major advantages over the older anesthetic, isoflurane, under the conditions of our study. Implications: The new inhaled anesthetic desflurane is acceptable in spontaneously breathing outpatients despite its known ability to irritate the airway. The IV anesthetic propofol was associated with more patient movement (without awareness) during surgery. Neither anesthetic conferred any clinically significant advantages over the older inhaled drug, isoflurane.

(Anesth Analg 1998;87:312-8)

Desflurane is a volatile anesthetic with properties that, theoretically, should allow rapid emergence and recovery after anesthesia. This would be of particular advantage in ambulatory surgery. Investigations have demonstrated that emergence from desflurane anesthesia is more rapid compared with isoflurane ^[1-3] and sevoflurane ^[4] and is comparable to propofol ^[5,6], although subsequent recovery and discharge times have not differed significantly from controls ^[1-7]. These findings may be influenced, in very short cases, by the choice of induction drug, as well as by factors other than the rapidity of recovery, which can affect discharge time. In addition, most investigations have examined desflurane in association with neuromuscular blocking drugs and controlled ventilation via a tracheal tube. In contrast, increasing use of the laryngeal mask airway (LMA) during ambulatory surgery has resulted in more patients breathing spontaneously ^[8]. The ability of patients to move and/or cough in the absence of neuromuscular blocking drugs may alter anesthetic requirements and therefore modify recovery times.

Desflurane produces airway irritation ^[9], which is particularly troublesome during the induction of anesthesia in adults ^[5] and children ^[10]. Airway irritation by desflurane is also thought to be partly responsible for the sympathetic stimulation that occurs when the delivered concentration is increased rapidly ^[11,12]. However, airway irritation was not a problem during emergence when patients' tracheas were extubated under deep desflurane anesthesia ^[13,14]. The use of desflurane in spontaneously breathing ambulatory patients seemed to warrant investigation to determine whether irritation of the airway would be a problem during the maintenance phase, and whether emergence would still be rapid after a depth of anesthesia sufficient to prevent intraoperative movement and coughing in the absence of neuromuscular block. We therefore designed this study to compare the use of desflurane with the use of two standard anesthetics administered under clinically relevant conditions to spontaneously breathing patients undergoing superficial ambulatory surgery.

Methods

The study was approved by the local research ethics committee, and written, informed consent was obtained from all patients. Ninety patients, aged 18-70 yr, ASA physical status I or II, undergoing ambulatory body surface surgery requiring general anesthesia were studied. Patients were excluded if they suffered from chronic respiratory diseases, had a personal or family history of malignant hyperpyrexia, had documented allergy or previous unusual responses to anesthetic drugs, or had a significant history of alcohol or recreational drug abuse.

An open-label, prospective study design was used, involving random allocation to one of three parallel groups for the maintenance of anesthesia (isoflurane, propofol, or desflurane) with blind assessment of postoperative recovery. Before surgery, patients completed baseline 100-mm visual analog scales (VAS) for anxiety, pain, and sedation (with 0 = minimal severity to 100 = maximal intensity), and nausea was assessed using a 4-point scale (none, mild, moderate, severe). Each patient then received slow-release ibuprofen 1600 mg orally approximately 1 h before the induction of anesthesia, as prophylactic postoperative analgesia. No other form of premedication was administered to any patient.

A standard technique for the induction of anesthesia was used in all three groups, and all anesthetics were administered by a single anesthesiologist (IS) with considerable experience with each technique. Preoxygenation was not performed. After siting an IV cannula, fentanyl 1 [micro sign]g/kg was administered, followed after 1 min by lidocaine 20 mg and propofol, administered at a rate of 20-40 mg every 10 s, in a dose sufficient to induce loss of consciousness, defined as a loss of verbal contact and absence of the eyelash reflex. An appropriately sized LMA was inserted after adequate jaw relaxation. If the LMA could not be satisfactorily placed at the first attempt, additional bolus doses of propofol (20-40 mg) were administered until satisfactory conditions were achieved. Ventilation was assisted using nitrous oxide (N₂ O) in oxygen after loss of consciousness and before LMA insertion. After successful placement of the LMA, the maintenance anesthetic was introduced according to the randomization schedule.

Patients received either isoflurane at an end-tidal concentration of 0.6% initially, then 0.25%-1%; a continuous infusion of propofol, initially 160 [micro sign]g [center dot] kg^-1 [center dot] min^-1 and subsequently 50-200 [micro sign]g [center dot] kg^-1 [center dot] min^-1; or desflurane at an end-tidal concentration of 3.6% initially, then 1.4%-6% as required. In addition, all patients breathed N₂ O 67% in oxygen at a fresh gas flow of 6 L/min for 10 min and subsequently 2 L/min from a circle absorber system. The propofol infusion rate or the inspired concentration of isoflurane or desflurane was adjusted as clinically indicated ^[15], the objective being to maintain the heart rate and blood pressure within 30% of the preinduction baseline value. Episodes of clinically inadequate anesthesia were treated by increasing the infusion rate (with a bolus administration of propofol if required) or inspired concentration of inhaled anesthetic in proportion to the severity of the disturbance being treated. If adequate anesthesia, indicated by the cessation or absence of patient movement, could not be achieved by these means, then fentanyl was administered in 25-[micro sign]g increments. Ventilation was manually assisted during the period of apnea resulting from the fentanyl-propofol induction sequence, after which patients breathed spontaneously.

Mean arterial blood pressure (MAP), heart rate (HR), hemoglobin oxygen saturation (SPO₂), and respiratory rate (RR) were measured noninvasively and recorded immediately before the induction of anesthesia. These measurements, as well as ETCO₂ and expired anesthetic concentration or propofol infusion rate, were repeated at 1-min intervals until 5 min after skin incision, and subsequently at 5-min intervals until the end of anesthesia. The incidence of coughing, excessive secretions, hiccups, stridor, laryngospasm, other respiratory complications, and any other events that delayed or disrupted surgery were recorded.

Just before insertion of the last skin suture, the wound edges were infiltrated with 0.5% bupivacaine for additional postoperative analgesia. After surgery, the maintenance anesthetic and N₂ O were discontinued simultaneously. Patients subsequently breathed 100% oxygen at the rate of 10 L/min. HR, MAP, SPO₂, and RR were recorded at 2-min intervals until removal of the LMA. The LMA was removed when patients made a purposeful attempt to remove it themselves. Patients then received oxygen via a face mask and were monitored as clinically indicated.

Emergence and recovery times were assessed by a blind observer (JA) who entered the operating room after discontinuation of the maintenance anesthetic. Emergence was assessed by recording the time from the end of anesthesia until patients could open their eyes, respond to verbal commands, and correctly state their date of birth. Subsequent recovery was assessed as the time required for patients to sit unsupported, tolerate oral fluids, walk unaided, and be assessed as fit for discharge. Patients were encouraged to recover at their own rate, without imposition of predetermined constraints. Patients were considered fit for discharge when they had achieved the preceding recovery end points, had stable vital signs, had no significant new symptoms, and were free from significant pain or nausea. VAS scores for anxiety, pain, and sedation were obtained at 15-min intervals during the first hour of recovery and subsequently at 30-min intervals. Nausea was assessed at the same time points using the 4-point rating scale, and any episodes of retching or vomiting were recorded. Two to three weeks after the operation, the patients were contacted by telephone to elicit any pain, nausea, headache, or somnolence that had occurred after discharge and to enquire as to the patients' overall satisfaction with their anesthetic.

Data were analyzed using analysis of variance for continuous variables (applying post hoc tests and Bonferroni's correction for multiple comparisons among groups, as required), the Kruskal-Wallis test for VAS scores, and chi squared test for discrete variables. In all cases, a P value <0.05 was considered significant.

Results

The patient characteristics for the three groups were generally comparable (Table 1), although there were significantly more patients of ASA physical status I in the isoflurane group. There was no significant difference among the groups for history of smoking. The average induction dose of propofol (194 +/- 21 mg; 3.0 +/- 0.5 mg/kg) did not differ among the three groups (Table 1).

Table 1: Demographic Characteristics of the Three Anesthetic Groups

The overall incidence of respiratory complications was low (Table 2) and comparable among the groups. Most complications occurred in association with the induction of anesthesia, before the effects of the maintenance anesthetic became established. Apnea after the induction of anesthesia occurred in all but four patients (two each in the isoflurane and propofol groups). The mean duration of apnea did not differ among the isoflurane, propofol, and desflurane groups (4.2 +/- 1.9, 5.9 +/- 3.3, and 5.3 +/- 2.5 min, respectively). Hemoglobin oxygen desaturation (<95%) occurred after the induction of anesthesia in 10, 15, and 14 of the patients in the isoflurane, propofol, and desflurane groups, respectively. The duration of the period of desaturation was brief, however, and resolved with assisted ventilation.

Table 2: Respiratory Complications in the Three Anesthetic Groups

Two patients each in the isoflurane and desflurane groups and four patients in the propofol group developed hiccups immediately after LMA insertion. The duration of hiccuping was brief (<3 min) in the inhaled anesthetic groups, although hiccuping continued for more than 10 min in three of four patients in the propofol group. Coughing occurred in six patients, and episodes of laryngospasm were recorded in seven patients (Table 2). All but one episode of laryngospasm (propofol group) occurred during the induction of anesthesia in association with LMA insertion. Four episodes (two in the propofol group and one in each of the other groups) were associated with oxygen desaturation. All episodes resolved with assisted ventilation, although an additional bolus of propofol was required in two patients (one each in the isoflurane and propofol groups). After reestablishing spontaneous ventilation, no patient in the isoflurane group experienced hemoglobin oxygen desaturation (<95%). Significantly more episodes of desaturation occurred in the propofol and desflurane groups (five and seven patients, respectively), although only one patient had a saturation level <90% (desflurane group; 89%).

The mean RRs and ETCO₂ values for the three anesthetic groups are shown in Figure 1. From the onset of regular spontaneous ventilation, the RR tended to be relatively lower in the propofol group and higher in the isoflurane group compared with the desflurane group. Only the difference between propofol and isoflurane achieved statistical significance, however. The higher respiratory rate in the isoflurane group was reflected in slightly lower ETCO₂ values compared with the other groups (Figure 1B), although ETCO₂ values for desflurane and propofol were almost identical and most differences among the three groups were not statistically (or clinically) significant. There were relatively few hemodynamic differences among the three groups (Figure 2). The HR in the desflurane group tended to be lower than that in the other groups, although this difference only occasionally reached statistical significance. The MAP (Figure 2B) did not differ among the three groups.

Figure 1: Respiratory rate (RR) (A) and end-tidal CO₂ values (B) at the indicated times after skin incision when anesthesia was maintained with isoflurane ([large circle]), propofol ([square]), or desflurane ([black circle]). Values are mean +/- SEM. * P < 0.05 versus the isoflurane group.

Figure 2: Heart rate (HR) (A) and mean arterial blood pressure (MAP) (B) before the induction of anesthesia (bl) and at the indicated times after the induction of anesthesia and skin incision when anesthesia was maintained with isoflurane ([large circle]), propofol ([square]), or desflurane ([black circle]). Values are mean +/- SEM. * P < 0.05 versus the isoflurane group. [dagger] P < 0.05 versus the propofol group.

Episodes of clinically inadequate anesthesia, as indicated by purposeful patient movements, occurred significantly more frequently in the propofol group compared with the inhaled anesthesia groups (Table 3). Episodes of purposeful movements in the inhaled anesthetic groups were terminated promptly by increasing the inspired anesthetic concentration. Increasing the propofol infusion rate only abolished patient movement in 3 of 19 patients-the remaining 16 patients required at least one bolus dose of propofol (Table 3). In addition, seven patients in the propofol group required one or more bolus doses of fentanyl to prevent movement. The average delay to surgery caused by these disruptive episodes is shown in Table 3. Purposeful movements were not predictable by previous changes in RR or hemodynamic variables in any of the groups. No patient subsequently had recall of any intraoperative event on direct postoperative questioning.

Table 3: Episodes of Purposeful Movement in the Three Anesthetic Groups

The time-weighted average anesthetic concentrations or infusion rates during the maintenance period were: isoflurane 0.8% +/- 0.05% (0.71 +/- 0.04 minimum alveolar anesthetic concentration [MAC]), propofol 115.5 +/- 5.6 [micro sign]g [center dot] kg^-1 [center dot] min^-1, and desflurane 3.91% +/- 0.14% (0.65 +/- 0.02 MAC). The corresponding anesthetic concentrations and infusion rates at the end of anesthesia in the three groups were 0.6% +/- 0.2% (0.5 +/- 0.1 MAC), 70 +/- 16 [micro sign]g [center dot] kg^-1 [center dot] min^-1, and 3.1% +/- 0.7% (0.5 +/- 0.1 MAC). There were no statistically significant differences among the inhaled anesthetic groups in average or final concentrations as a multiple of MAC. There were no statistically significant differences observed among the three groups for emergence and recovery times (Table 4). Similarly, there were no significant differences in hemodynamic variables in the early recovery phase or in VAS scores for pain, anxiety, or sedation at any point during recovery. Postoperative analgesic requirements were low and did not differ among the groups. The incidence of postoperative nausea and vomiting were also low (Table 4) and similar in all groups. Only one patient (in the isoflurane group) required an antiemetic. Most patients said they would be happy to receive the same anesthetic again, with no differences detected among the groups. Telephone follow-up revealed a low incidence of postdischarge complications, with no differences among the groups.

Table 4: Emergence and Recovery Variables in the Three Anesthetic Groups

Discussion

Desflurane seemed to be a satisfactory maintenance anesthetic in spontaneously breathing ambulatory patients. There was no evidence of significant respiratory side effects with desflurane during the maintenance or emergence phase (even in patients with a history of smoking), as respiratory complications during this time were no more common than in the other two anesthetic groups. Clinically significant respiratory depression was not observed with any of the maintenance anesthetics. Most respiratory complications were related to the induction sequence and insertion of the LMA. The incidence of complications that we observed was probably higher than that encountered in routine clinical practice, in which it is common to introduce the maintenance anesthetic immediately after loss of consciousness. In contrast, our protocol only allowed the maintenance anesthetic to be administered after LMA insertion, although ventilation was assisted with N₂ O and oxygen. Of interest, the introduction of a volatile anesthetic for the maintenance of anesthesia seemed to rapidly abolish any hiccups that occurred after LMA insertion. In contrast, this minor side effect was prolonged in those patients receiving a propofol infusion, although it did not significantly interfere with the conduct of anesthesia or surgery.

No major differences in cardiovascular variables were observed among the three anesthetic groups. The HR was somewhat lower at several time points in the desflurane group compared with the other two groups. This may have reflected minor differences in the depth of anesthesia, as desflurane is known to have an effect on HR similar to isoflurane ^[16]. We did not observe any evidence of sympathetic stimulation with desflurane, although no patient required an abrupt increase in the delivered desflurane concentration. The increases required represented only a small percent change in the inspired concentration.

The quality of control of anesthetic depth was superior in the inhaled anesthesia groups compared with the propofol group, which was associated with an unacceptable rate of purposeful movements that significantly delayed surgery. Similar problems have been reported previously with propofol ^[17,18] and are probably related to the complex relationship between the delivered concentration over time and the resultant blood levels, as well as the wide variability in patient responses at a given plasma concentration of propofol ^[19]. Pharmacokinetic modeling of our infusion regimen (PK-SIM; Specialized Data Systems, Jenkintown, PA) suggested that our bolus-infusion regimen should have achieved and maintained plasma propofol concentrations of approximately 5 [micro sign]g/mL during the period when most movements occurred. This blood level value is considered adequate for minor surgery in association with N₂ O and/or opioids ^[19], although plasma propofol concentrations were not actually measured in our patients. Alterations in the propofol infusion rate have only a small effect on plasma concentration, which explains the requirement for bolus administration in most of our patients.

Of particular concern was that gross purposeful movements occurred without warning and were not preceded by alterations in HR, MAP, or RR. These clinical signs are less reliable in titrating IV anesthetics than they are with volatile drugs ^[20]. In spontaneously breathing patients, inadequate anesthesia is manifested by purposeful movements, which allows corrective action to be taken promptly. As a result, awareness is unlikely, and none of our patients recalled intraoperative events. When muscle relaxants and controlled ventilation are used, however, awareness is a greater risk because inadequate anesthesia can occur without obvious changes in clinical signs. A further problem with propofol is interruption of drug delivery leading to inadequate anesthesia. This affected one of our patients due to extravasation of propofol. This particular patient moved on two occasions, followed by a third, especially severe movement. At this point, extravasation was suspected and confirmed by inspection of the indwelling cannula, which was immediately replaced. Because no problem was observed during induction, we assume that the cannula became displaced during one of the earlier episodes of movement. No harm came to the patient through either extravascular propofol administration or inadequate anesthesia, although awareness would have been a risk had neuromuscular blocking drugs been used.

Emergence from anesthesia did not differ among the three anesthetic groups. Our sample size had sufficient power to detect a significant difference of 3.5 min in awakening, as this was considered to be the smallest clinically relevant difference. Significant differences in emergence times between desflurane and propofol maintenance anesthetics have not been found in earlier studies involving controlled ventilation ^[5,6,21], although other investigators have detected significantly earlier emergence from desflurane compared with isoflurane ^[1-3,22]. In several of these studies, however, anesthesia was induced with thiopental ^[1-3], and emergence times in both anesthetic groups were longer than the average times recorded for any of our patients. Although Gupta and colleagues ^[22] detected statistically significant earlier emergence from desflurane anesthesia compared with isoflurane (4.25 vs 5.45 min), this difference was of little clinical significance. A recent meta-analysis concluded that differences in recovery times between desflurane and isoflurane or propofol are only of minor clinical importance ^[21]. Like most other investigators, we found no difference in late recovery events among the three groups of patients, and our recovery times were generally similar to those reported previously after a variety of ambulatory surgical procedures.

Patient satisfaction was equal with all three anesthetics, and the overall incidence of postoperative nausea was acceptably low (approximately 6%). Using VAS scores for sedation, we could not detect any subjective differences in recovery at any time point among the three groups. Postoperative pain was well controlled by the combination of nonsteroidal antiinflammatory drugs and local anesthetic infiltration; additional analgesia was rarely required. The severity of postoperative pain was not affected by the maintenance anesthetic used.

In summary, a continuous infusion of propofol at the doses used in this study resulted in a poor quality of anesthesia with an unacceptably high incidence of intraoperative movements and no obvious clinical benefits. In contrast, desflurane and isoflurane seemed to be acceptable anesthetics for spontaneously breathing ambulatory patients, allowing excellent control of the depth of anesthesia without obvious respiratory irritant problems. Nevertheless, the lower solubility of desflurane did not result in more rapid early or intermediate recovery compared with isoflurane under these clinical conditions.

We are grateful to Professor Paul F. White from the University of Texas Southwestern Medical Center at Dallas for his assistance during the initial planning of the protocol. We thank our surgeons, Mr. R. Kirby, Mr. J. Skilton, and Mr. T. Duffy for their tolerance during the conduct of this study. We also thank the nursing staff of the recovery rooms and the Adult Day Care Unit for their cooperation with this protocol.