A Comparison of Airway Responses During Desflurane and Sevoflurane Administration via a Laryngeal Mask Airway for Maintenance of Anesthesia

Desflurane and sevoflurane are the most widely used general anesthetics in North America. Each has particular advantages and disadvantages. Desflurane has a lower solubility in blood (1,2) and thus allows a more rapid awakening from anesthesia (3). Sevoflurane does not irritate the respiratory tract, whereas desflurane can do so at concentrations that exceed the minimum alveolar anesthetic concentration (MAC) (4), implying that sevoflurane would be preferable to desflurane when anesthesia is to be delivered via a laryngeal mask airway (LMA; Laryngeal Mask Company, Henley-on-Thames, United Kingdom). The importance of a difference lies in the widespread use of the LMA for delivery of anesthesia. The LMA is widely used because it supports the airway, minimizes dead space, may be inserted without the use of muscle relaxants, provides a partial seal of the airway, can allow effective use of positive-pressure ventilation, and is less stimulating than a tracheal tube (5). However, the LMA does not prevent laryngospasm and is more difficult to use with controlled ventilation than a tracheal tube. Mahmoud et al. (6) found no significant difference in respiratory responses between desflurane and sevoflurane given through a LMA. However, anesthesia duration equaled 18 min (10–31 min) (mean [range]) after the induction of anesthesia with propofol, and therefore, this may have been a limited test of any differences between desflurane and sevoflurane. In studies that compared desflurane with another anesthetic (not sevoflurane) (7,8) or sevoflurane with another anesthetic (not desflurane) (9), all delivered through a LMA, no differences in the incidence of measures of airway irritation (e.g., coughing) were found between anesthetics or studies. We hypothesized that the absence of a difference resulted from the use of all anesthetics at the smaller concentrations often used in clinical practice. We tested this hypothesis in the present study. In addition, we took the opportunity to make comparative measurements of clinically useful variables such as the rapidity and quality of awakening. Methods The study was approved by the respective IRBs at the University of California, San Francisco (UCSF) and Texas Tech University, Lubbock. ASA physical status I–III male (n = 62) and female (n = 65) patients aged 18–75 yr scheduled to have general anesthesia for surgical procedures at either institution were given a written description of the study at their visit to the presurgical clinic. Inclusion criteria included selection by the attending anesthesiologist of general anesthesia through a LMA. One patient who was randomized to receive desflurane (not counted in the above totals) was deleted from analysis because he also received an infusion of propofol throughout anesthesia. After obtaining informed consent from each patient on the day of surgery, a standard anesthetic consisting of the following was delivered: The attending anesthesiologist decided whether to premedicate each patient with midazolam or nothing. Anesthesia was induced with propofol at a dose defined by the attending anesthesiologist–a dose sufficient to allow insertion of a LMA. The attending anesthesiologist also could add fentanyl to the induction regimen. Once the LMA was positioned and spontaneous ventilation resumed, desflurane or sevoflurane (assigned randomly using a computer-generated scheme) was administered in a target background of approximately 50% nitrous oxide at a maintenance total gas flow of 1 L/min. Desflurane and sevoflurane concentrations to be used were at the discretion of the attending anesthesiologist, usually between 0.3 and 1.0 MAC, as revealed by end-tidal measurements of respired gases. MAC values were assumed to equal 6% desflurane and 1.85% sevoflurane (10). Additional propofol (boluses) and/or an opioid were prescribed as the attending anesthesiologist saw fit. Although the usual opioid used was fentanyl, occasional patients received morphine, Dilaudid, meperidine, or ketorolac. We assume that 10 mg of morphine, 1.5 mg of Dilaudid, 75 mg of meperidine, and 30 mg of ketorolac each were equivalent to 100 μg of fentanyl (11,12). A blinded observer recorded the patient’s vital statistics (age, weight, and height), significant diseases, smoking history, duration of anesthesia, the anesthetic concentrations and drug doses delivered, and the incidence of respiratory events (specifically coughing, breath holding, and laryngospasm). Events were graded for severity. Coughing was defined as 0 if no coughing occurred, 1 if a single cough occurred and Spo2 ≥95%, 2 if multiple coughs occurred and Spo2 ≥95%, 3 if multiple coughs occurred and Spo2 <95%, and 4 if multiple coughs occurred, Spo2 <95%, and if coughing required an administration of IV medication. Breath holding was defined as 0 if no breath holding occurred, 1 if no breath holding occurred for 10–20 s, 2 if no breath holding occurred for 20–30 s, and 3 if breath holding exceeded 30 s. Laryngospasm was defined as 0 if no evidence of phonation or stridor was present, 1 if phonation or stridor appeared for <15 s and no therapy other than positive-pressure ventilation was required, 2 if phonation or stridor occurred for >15 s and no therapy was required other than positive-pressure ventilation, and 3 if phonation or stridor occurred for >15 s and IV mediation was required. Spo2 was measured as the smallest value observed during each 15-min epoch. The average, median, and largest vapor and nitrous oxide concentrations were measured. All variables were evaluated for each 15-min epoch during anesthesia and recorded for each epoch. A blinded observer made the observations for the first hour after discontinuation of the anesthetic administration. The observer determined the time from discontinuation of the anesthetic administration to first appropriate response to command and to orientation to time and place. Digit symbol substitution tests were performed at 15-min intervals after discontinuation of the anesthetic administration, and the results were compared with those obtained before anesthesia. A modified Aldrete score was obtained at 15-min intervals after discontinuing the anesthetic administration (13). For each postanesthetic 15-min epoch, we recorded the incidence of nausea, vomiting, or both. In addition, postoperative nausea and vomiting (PONV) were graded 0–3, where 0 equaled no nausea or vomiting, 1 = mild nausea without vomiting or medical intervention, 2 = nausea without vomiting but with administration of antiemetic therapy, and 3 = vomiting. Finally, time to be fit for discharge (as defined by the ability to sit up and to ambulate with assistance) was recorded. For ordinal data, unpaired two-tailed Student’s t-test were applied, and significance was accepted at P < 0.05 without correcting for multiple comparisons. χ2 analysis was applied as indicated. Results We excluded one patient given desflurane from our analysis who (unlike all other patients) received a propofol infusion and thereby deviated from the study protocol. This patient coughed during the first (but not later) 15-min interval of anesthesia at a time when the LMA was removed and then reinserted. Desflurane administration was discontinued during the reinsertion of the LMA, and the end-tidal concentration was 2.6% during this 15-min time interval. Therefore, a total of 63 patients received desflurane, and 64 received sevoflurane. Patients at UCSF and Texas Tech did not differ except that patients at UCSF were older and had longer anesthetic durations. Similarly, the demographics, including the types of surgery, for patients given desflurane did not differ from those for patients given sevoflurane (Table 1). Coughing, breath holding, laryngospasm, and Spo2 did not differ between anesthetics (Table 2). The largest sevoflurane MAC value applied (1.06) significantly (P < 0.01) exceeded that for desflurane (0.88); the sevoflurane average MAC value (0.85) did not significantly exceed the MAC value for desflurane (0.75), and the nitrous oxide values did not differ between groups. Twenty-seven percent of patients received desflurane concentrations exceeding 6%. The incidence of coughing or breath holding in these patients did not differ from the incidence in patients not given concentrations exceeding 6%. Forty-eight percent of patients received sevoflurane concentrations exceeding 1.85%. The incidence of coughing or breath holding did not differ from the incidence in patients not given concentrations exceeding 1.85%. Smoking did not significantly increase the incidence of coughing, breath holding, or laryngospasm.Table 1: Comparison of Desflurane Versus Sevoflurane DemographicsTable 2: Respiratory ObservationsAfter anesthesia, patients anesthetized with desflurane responded to command sooner, were oriented sooner, had a greater Digit Symbol Substitution Test result at 15 min (Fig. 1), and reached an Aldrete score of 13 sooner. However, time to fit for discharge did not differ between desflurane and sevoflurane (Table 3). Pain and sore throat scores did not differ between the anesthetics.Figure 1: The rapid return of the results of the Digit Symbol Substitution Test (a test that requires the subject to match a number to a symbol) for both desflurane and sevoflurane indicates the quickness of recovery found with both anesthetics, consistent with their relatively low solubilities. Recovery at 15 min is significantly increased with the less soluble (than sevoflurane) desflurane, but no difference is found for later times.Table 3: Recovery ValuesIntraoperative (i.e., prophylactic) antiemetic therapy was given to 37 of 63 (59%) patients anesthetized with desflurane and 43 of 64 (67%) patients anesthetized with sevoflurane (difference not significant). None of the 63 patients given desflurane had PONV scores >1, whereas 5 of 64 patients given sevoflurane had PONV scores >1 (P < 0.1). For the first hour after anesthesia, patients given desflurane did not significantly differ in the incidence of PONV from those given sevoflurane (overall values for the first hour after anesthesia were 6 of 63 [10%] patients given desflurane and 13 of 64 [20%] given sevoflurane], but fewer patients given desflurane received antiemetic therapy (2 of 63 [3%]) than patients given sevoflurane (13 of 64 [20%];P < 0.01). Discussion We found that both desflurane and sevoflurane could be delivered through a LMA for maintenance of anesthesia with minimal or no evidence of airway irritation and no significant difference in airway irritation between desflurane and sevoflurane. These results are consistent with the results obtained by Mahmoud et al. (6) who studied respiratory responses for brief (18 minutes of anesthesia) gynecological procedures. The brevity of the procedures after the induction of anesthesia with propofol limits the interpretation of their findings (i.e., the propofol might have compromised the ability to assess a difference between the anesthetics). The present results support our hypothesis that the irritant effects of desflurane are not clinically evident at the concentrations used for maintenance of anesthesia. Those concentrations usually were less than MAC (i.e., less than an irritating concentration). However, 27% of patients received desflurane concentrations in excess of MAC, and the incidence of coughing or breath holding in these patients was not increased, nor was the incidence decreased in patients given sevoflurane concentrations in excess of 1.85% (52% of patients). The conditions used in this study may have influenced the finding of no difference between desflurane and sevoflurane. Anesthesia was induced with propofol and fentanyl, and the placement of the LMA proceeded in this context. Both inhaled anesthetics were delivered after insertion of the LMA, and the initial concentrations applied were less than MAC. Other studies have demonstrated that concentrations of desflurane <6% do not elicit untoward respiratory responses such as coughing or laryngospasm (14). However, as noted above, several patients received concentrations exceeding MAC for desflurane (6%) and did not have an increased incidence of coughing or breath holding. Administration of fentanyl provided another factor that may have abolished a difference between the responses to desflurane versus sevoflurane. Administration of 1 μg/kg of fentanyl as premedication decreases the incidence of coughing on the induction of anesthesia with desflurane by 80%(15). Smoking did not significantly increase the incidence of coughing, breath holding, or laryngospasm. However, the number of smokers in our sample was too small to provide a robust test of whether smoking would affect these variables. As anticipated from their respective solubilities, early recovery was quicker with desflurane than sevoflurane. However, this difference also might have resulted from the slightly larger (MAC) concentration of sevoflurane used for anesthesia (Table 1). Late recovery did not differ between the two anesthetics. The finding of increased use of postanesthesia care unit (rescue) antiemetics to treat PONV with sevoflurane than with desflurane was not expected. Most studies find no difference in PONV, and our result may be spurious. However, with one exception, it does not seem to be explained by increased administration of drugs that predispose to nausea (e.g., fentanyl or nitrous oxide; see Table 1), nor is it explained by a smaller administration of antiemetic drugs (including propofol) to patients given sevoflurane. The exception that might explain the more frequent incidence with sevoflurane is the trend to use a larger concentration of sevoflurane for anesthesia (Table 1). Deficiencies in study design may limit interpretation of our results. As noted above, we did not impose a rigid dose scheme regarding the drugs or doses used (e.g., concentrations of the inhaled anesthetics, doses of opioids, and the choice of opioids), and bias of the anesthesiologists may have altered some of the results. This might explain, for example, the finding of increased use of postanesthesia care unit antiemetics to treat PONV with sevoflurane than with desflurane. The anesthesiologist was not blinded to the anesthetics and drugs delivered (although the person recording the data was blinded). In summary, we find that the respiratory complications that arise during maintenance delivery of anesthesia through a LMA are minor, their incidence is small, and the incidence does not differ for desflurane versus sevoflurane. Initial recovery may be more rapid with desflurane, but time to be fit for discharge does not differ. Nausea may be slightly increased after anesthesia with sevoflurane, but this result may be because of a slightly larger MAC of sevoflurane used to maintain anesthesia rather than any inherent difference between the two anesthetics. The authors appreciate the several suggestions made by Dr Edmond I Eger II, MD, who is a paid consultant to Baxter Healthcare Corp.