| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the *Department of Anesthesia and Perioperative Care, University of California, San Francisco, California; and Department of Epidemiology and Biostatistics; University of California San Francisco, San Francisco, California.
Address correspondence and reprint requests to Dr McKay, Department of Anesthesia, C-450, University of California, San Francisco, CA 94143-0464. Address e-mail to eshimar{at}anesthesia.ucsf.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMETHODSRESULTSCOMBINED RESULTSDISCUSSIONREFERENCES |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMETHODSRESULTSCOMBINED RESULTSDISCUSSIONREFERENCES |
|---|
Despite the potentially greater airway irritant properties of desflurane, studies comparing desflurane with sevoflurane administered via an LMA find no difference in the incidence of coughing, breath holding, or laryngospasm (4,5). We believe the absence of any difference results from the use of both anesthetics at concentrations that are not irritating to the airway. That is, at the lower concentrations commonly used in clinical practice, neither desflurane nor sevoflurane irritates the airway, and both may be given without evidence of increased airway irritation via an LMA. The use of small doses of opioids would further minimize any differences in the response to the two anesthetics (6).
However, previous studies have not focused on the contribution of smoking to desflurane’s potential to elicit a greater incidence of untoward effects. In the study by Eshima et al. (4) comparing airway responses during desflurane versus sevoflurane anesthesia, only 22% (27 of 127) of the patients smoked. Twenty-two percent (6/27) of these smoking patients coughed or held their breath at some moment during delivery of anesthesia via an LMA, while only 6% (6/100) of the nonsmokers coughed or held their breath. These differences were not statistically significant, but the assessment of significance would have been limited by the small size of the group of patients who smoked. The preceding observations prompted the present study of untoward airway responses during desflurane versus sevoflurane anesthesia given via an LMA in patients with an immediate history of smoking.
|
METHODS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSCOMBINED RESULTSDISCUSSIONREFERENCES |
|---|
Our "standard" anesthetic consisted of the following. The attending anesthesiologist decided whether to premedicate each patient with up to 2 mg of midazolam or not. Anesthesia was induced with propofol at a dose defined by the attending anesthesiologist, a dose sufficient to allow insertion of an LMA. The attending anesthesiologist also could add fentanyl (up to 100 µg) and/or lidocaine (up to 100 mg) 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 1–2 L/min. Desflurane and sevoflurane concentrations used were determined by the attending anesthesiologist, and usually varied between 0.3–1.0 MAC as revealed by end-tidal measurements of respired gases. A Datex S5 module (Datex-Ohmeda, Madison, WI) using infrared analysis recorded end-tidal anesthetic concentration, and the anesthesiologist manually entered those values into the anesthesia record. MAC values were assumed to equal 6% desflurane and 1.85% sevoflurane (7). Additional propofol (boluses) and/or fentanyl were prescribed as the attending anesthesiologist saw fit.
A blinded observer recorded data before surgery, including the patient’s vital statistics (age/weight/height), the presence of chronic or acute lung disease, smoking history, history of opioid use, and baseline oxygen saturation. Before induction of anesthesia, a panel was placed in front of the anesthesia vaporizers and end-tidal concentration monitors so that the observer was not able to see which anesthetic (desflurane or sevoflurane) was administered. In the operating room, the observer recorded the duration of anesthesia, lowest oxyhemoglobin saturation (SpO2) and incidence of respiratory events (specifically coughing, breath holding, laryngospasm) during each 15-minute epoch, and the time from discontinuation of inhaled anesthetic until the patient first followed commands. To maintain blinding, anesthetic values and opioid and propofol doses were obtained by a later review of the written record. Events were graded for severity as follows:
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%; 4 if multiple coughs occurred, Spo2 < 95%, and coughing required administration of IV medication.
Breath holding was defined as 0 if no breath holding occurred; 1 if breath holding occurred for 10–20 s; 2 if breath holding occurred for 20–30 s; 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; 3 if phonation or stridor occurred for >15 s and IV mediation was required.
Spo2 was measured on a Nellcor Oximax® N-600TM unit (Nellcor, Pleasanton, CA), and the observer recorded the lowest saturation value displayed during each 15-min epoch of anesthesia. End-tidal anesthetic concentrations were obtained from review of the written anesthetic record after completion of data collection. The average, median, and highest vapor and nitrous oxide concentrations were calculated from end-tidal values recorded by the anesthesiologist for each 15-min epoch.
Observations continued to be made by a blinded observer for the first hour after discontinuation of anesthetic administration. The observer determined the time from discontinuation of administration to first appropriate response to command. Spo2 was measured on arrival in the postanesthesia care unit (PACU) while the patient breathed oxygen delivered at 10 L/min from a facemask. Spo2 was measured along with oxygen flows delivered by nasal cannula or facemask, at four subsequent 15 min intervals. Digit symbol substitution tests were performed at 15 min intervals after discontinuation of anesthetic administration, and the results were compared with those obtained before anesthesia.
Nausea was graded on a 0–3 scale, where 0 equaled no nausea; 1, mild nausea; 2, severe nausea without retching or emesis; and 3, nausea accompanied by retching and/or emesis. Vomiting was graded on a 0–2 scale, where 0 equaled no vomiting; 1, retching without vomiting; 2, vomiting. Pain was assessed on a verbal analog scale of 0–10. The time and dose of opioids and antiemetic drugs were noted.
All variables were evaluated and recorded at 4 consecutive 15-min intervals during the first hour in the PACU after the time from which the patient first followed commands.
Twenty-four hours after anesthesia, patients were contacted by telephone (outpatients) or in their hospital rooms and questioned regarding their opioid and antiemetic use, worst pain and nausea scores experienced (0–10 scale), and the presence, if any, of vomiting (0–2 score). Patients were asked to estimate the percent of self-care activities they had resumed compared to their preanesthetic levels of such activities. A subset of patients was specifically asked if they had gotten out of bed, dressed, and left the house.
After review and analysis of data from the present study (intraoperative data from 110 smokers and postoperative data from 108 smokers), we combined the data with data obtained from an earlier study of patients evaluated under similar protocol (100 nonsmokers and 27 smokers). Methods differed only in that no data were collected past the first hour of recovery in the PACU in the earlier study, and selection was not based upon smoking status. Combining the groups was done in an attempt to examine the impact of smoking status on the frequency and severity of intraoperative respiratory complications, and to strengthen analyses of the effect of age, gender, length of surgery, MAC hours of anesthetic and opioid dose on postoperative cognition, pain, nausea, and vomiting during first hour after emergence.
Unpaired two-tailed t-tests or Mann–Whitney tests were applied where indicated for continuous or ordinal data, and significance was accepted at P < 0.05 without correcting for multiple comparisons. Fisher’s exact tests were used for comparison of respiratory events between patients receiving desflurane versus sevoflurane, and for other categorical comparisons. Logistic regression modeling was used to study the effect of smoking on dichotomized nausea and pain scores. The sample size of smokers in the study (n = 110), when combined with 27 smokers from the previous study, permits detection of a difference in incidence of coughing of 0.05 in the sevoflurane group versus 0.21 in the desflurane group with an 
= 0.05 and a power = 0.80.
|
RESULTS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSCOMBINED RESULTSDISCUSSIONREFERENCES |
|---|
The demographics, including the types of surgery, for patients given desflurane did not differ from those for patients given sevoflurane, except that patients receiving desflurane had more extensive smoking histories (Table 1). The average and highest end-tidal MAC fractions did not differ significantly between anesthetics. Consistent with the greater solubility and metabolism of sevoflurane, the highest sevoflurane MAC fraction concentration delivered from the anesthetic machine exceeded that for desflurane. Average oxygen saturation and lowest oxygen saturation measured during anesthesia did not differ between patients receiving sevoflurane or desflurane. After anesthesia, patients anesthetized with desflurane responded to command sooner than patients anesthetized with sevoflurane (284 ± 167 s vs. 355 ± 156 s, P = 0.03).
|
Coughing, breath holding, and laryngospasm did not differ between anesthetics (Table 2 and Fig. 1), nor did average and lowest Spo2. In 43% of the patients (n = 23) given desflurane, concentrations exceeded 6%. The incidence of coughing or breath holding (13%, n = 3) in these patients did not differ from the incidence in patients given concentrations under 6% (10%, n = 4). Digit Symbol Substitution Test (DSST) results and scores for pain, nausea, and vomiting did not differ between the anesthetics. However, more patients receiving sevoflurane were unable to take the DSST at 15 min after anesthesia as a result of drowsiness compared with those who had received desflurane (28% vs. 9%, P = 0.03). The overall number of patients receiving sevoflurane or desflurane unable to take the DSST because of nausea, pain, or unavailability (urination etc) did not differ significantly.
|
|
Patients received 10 L/min of oxygen by facemask en route to the PACU from the operating room. On arrival, mean oxygen saturation did not differ significantly between groups, although there was a trend toward more patients receiving sevoflurane with Spo2 < 95% (n = 7) compared with those receiving desflurane (n = 1, P = 0.07). During the first 15 min period in the PACU, patients in both groups had similar oxygen saturations, but those recovering from sevoflurane received more oxygen to attain that saturation (3.3 ± 1.7 vs 4.3 ± 3.0 L/min for patients getting desflurane, P = 0.03), and more frequently received oxygen delivered via facemask (versus nasal cannula) compared with those recovering from desflurane (n = 11 vs 3, P = 0.04).
Intraoperative (i.e., prophylactic) antiemetic therapy was given to 45/54 (83%) patients anesthetized with desflurane and 50/54 (93%) patients anesthetized with sevoflurane (difference not significant). The incidence of nausea and vomiting in the recovery room and during the 24-h recovery period, as well as the need for rescue antiemetics during both periods, did not differ. Thirteen patients given desflurane and 10 given sevoflurane received antiemetic medication in the recovery room. Pain scores, analgesic use, and activity level (getting out of bed, dressing in clothing, or leaving the house) did not differ between patients receiving desflurane versus sevoflurane.
|
COMBINED RESULTS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSCOMBINED RESULTSDISCUSSIONREFERENCES |
|---|
Age, gender, use of supplemental regional anesthesia, opioids use, type of surgery, duration of potent inhaled anesthetic delivery, average MAC-fraction, and MAC-hours did not differ for the combined data (smokers plus nonsmokers) between patients given desflurane versus sevoflurane. Patients receiving sevoflurane were given antiemetic prophylaxis more often than patients receiving desflurane (82% vs 70%, P = 0.03).
Smokers were more likely to receive adjunctive regional anesthesia (46% vs 2%; P < 0.001) were more often males (64% vs 45%; P < 0.005), and were more likely to receive prophylactic antiemetic therapy (82% vs 69%; P < 0.05). As would be predicted from the difference in gender distribution, smokers had less gynecologic surgery (11% vs 29%) but more orthopedic surgery (65% vs 41%). The differences in types of surgery were significant (P < 0.001). Smokers received more fentanyl (150 µg median dose vs 100 µg for nonsmokers, P < 0.005), had longer anesthetics (87 min vs 58 min, P < 0.001) and longer MAC-hours (0.99 MAC-hours vs 0.71 MAC-hours; P < 0.005), but did not receive larger average MAC-fractions of anesthetic as reflected by end-tidal concentration (0.69 MAC vs 0.71 MAC; P > 0.05).
In the combined group of smokers and nonsmokers, the incidence of coughing, breath holding, or laryngospasm did not differ between patients receiving desflurane versus sevoflurane. More smokers coughed during anesthesia than did nonsmokers (12% vs 1% of patients, P < 0.002) (Fig. 2; Table 3). Coughing occurred at times when levels of stimulation changed abruptly or when anesthetic concentrations decreased (Fig. 3). That is, more of the episodes of coughing occurred immediately after LMA placement and at emergence than during maintenance.
|
|
|
Pain scores did not differ between patients receiving desflurane versus sevoflurane, or between male versus female patients. Younger patients and those undergoing longer surgery had higher pain scores. Smokers underwent longer surgery and received more opioid medication, but did not differ in mean age from nonsmokers. Although smokers had significantly higher pain scores after surgery than did nonsmokers (median score for the first hour after surgery 5.2 vs 4.3; P < 0.05), multivariate modeling that also considered age and anesthetic duration eliminated the statistical significance (odds ratio for a smoker reporting a pain score 5 1.63, 95% CI 0.9–2.9, P = 0.1).
Despite the greater use of antiemetic prophylaxis, smokers had higher nausea scores. A dichotomous analysis showed that smokers had more moderate to severe nausea than nonsmokers (defined by a score of 2 or more) at all times during the first hour in the PACU (P < 0.01–0.05 for the various time intervals). A multivariate model that included gender, opioid dose, age, anesthetic (sevoflurane versus desflurane), MAC hours and smoking status found that smoking status remained a statistically significant predictor of frequent nausea (Table 4).
|
In smokers, recovery of judgment and cognition, reflected by the percent of baseline DSST administered at 15, 30, 45, and 60 min after following commands, did not show statistically significant differences between patients given desflurane versus sevoflurane when individual performances were compared to baseline scores. However, patients receiving desflurane had significantly higher scores than patients receiving sevoflurane during the first hour after following commands when scores were measured as area under the curve (AUC, DSST score by min) at the 0–60, 15–60 and 30–60 min intervals (P = 0.03–0.04 by Mann–Whitney test). The combined data from this and the previous study comparing desflurane and sevoflurane revealed significantly higher scores at 15 min after following commands in patients given desflurane (average score 55% vs 41% of baseline effort; P = 0.02). Smokers had DSST scores more closely approaching baseline scores at 15 and 30 min after following commands than did nonsmokers (Fig. 4).
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMETHODSRESULTSCOMBINED RESULTSDISCUSSIONREFERENCES |
|---|
Our previous study enrolled patients who were mostly (73%) nonsmokers, and was not designed or adequately populated to detect differences in respiratory responses based upon smoking status (4). When subjects from the current and previous study were considered together, the incidence of coughing in the smokers (11.7%) significantly exceeded the incidence in nonsmokers (1.0%, P = 0.002; Fig. 2). That is, smokers were more likely to display evidence of airway irritation (Table 3), regardless of anesthetic choice (Fig. 2). In both smokers and nonsmokers who had evidence of airway irritation, the response scores usually were 1 or 2. Two patients (one desflurane, one sevoflurane) had Spo2 values <95%.
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 placement of the LMA proceeded in this context. Both inhaled anesthetics were delivered after insertion of the LMA, and the initial concentrations administered were less than MAC. Other studies have demonstrated that concentrations of desflurane <6% do not elicit untoward respiratory responses, such as coughing or laryngospasm (8). However, as noted above, 23 patients received concentrations exceeding MAC for desflurane, and did not have an increased incidence of coughing or breath holding. The innocuousness of clinically used anesthetic concentrations during maintenance is further suggested by the times at which coughing occurred (Fig. 3). Coughing occurred primarily during induction and recovery and not during maintenance. The anesthetic concentrations associated with recovery are clearly too small to provoke responses.
Administration of fentanyl also may have minimized differences between the responses to desflurane versus sevoflurane. Administration of 1 µg/kg of fentanyl as premedication decreases the incidence of coughing on induction of anesthesia with desflurane by 80% (6). Patients receiving sevoflurane and desflurane received similar doses of fentanyl and breathed spontaneously throughout surgery.
The recently published study of Arain et al. (9) found a similar lack of airway responsiveness in patients undergoing a standard anesthetic (propofol induction, air/oxygen and 1 µg/kg fentanyl), with desflurane versus sevoflurane given at 1 MAC. Subsequently, the anesthetic deviated from standard practice, wherein the LMA was forcefully manipulated in the pharynx and 2 MAC of inhaled anesthetic administered abruptly. Coughing and circulatory stimulation were seen in half of those subjects given desflurane, consistent with desflurane’s known pungency when administered acutely at a high concentration and without adjunctive medication (e.g., fentanyl). This approach to anesthetic delivery seems to be clinically unsatisfactory for both anesthetics, albeit worse with desflurane, producing unwanted responses with both sevoflurane (20% coughing) and desflurane (40% coughing) during emergence. Why might more patients who received desflurane cough on emergence? Our earlier study demonstrated faster return of airway reflexes when patients awaken after desflurane compared with sevoflurane anesthesia (10), and faster recovery of airway reflexes implies earlier rejection of the LMA when no longer needed for airway support. We suggest that to avoid coughing at emergence, the LMA should be removed before a patient begins to respond to its presence. Furthermore, one might expect that deliberate attempts to traumatize the pharynx during anesthesia would increase airway reactivity in a rapidly awakening patient with an LMA in situ.
As anticipated from their respective solubilities, early recovery was quicker with desflurane than sevoflurane. Recovery beyond 1 h after response to commands did not differ between the two anesthetics.
In the present study of 110 patients, no difference in recovery of DSST performance compared with baseline score was seen among patients receiving desflurane versus sevoflurane. This finding contrasts with that of our earlier work (4), wherein DSST performance was closer to baseline 15 min after wake-up in patients who had received desflurane compared with those receiving sevoflurane. One factor in the present study that may have contributed to the apparent lack of difference is that more patients receiving sevoflurane were unable to complete the test at 15 min because of somnolence (15 vs 5, P = 0.03). In attempting to analyze what, if any, effect smoking itself might exert on recovery of judgment and cognition, we combined the DSST recovery data of our smokers in the present study with the 27 smokers from the 2003 study (4), and compared the results in this larger group of smokers to results from the 100 nonsmokers from the 2003 study. This comparison suggests that smokers recovered judgment and cognition more rapidly than nonsmokers at 15 and 30 min after wake-up compared with nonsmokers, when scores were considered as percent of baseline (Fig. 4). When all smokers and nonsmokers were considered together, recovery of DSST as a percent of baseline was more than 15 minutes after wake-up in patients who had received desflurane versus sevoflurane, and performance was higher in patients receiving desflurane compared with patients receiving sevoflurane when calculated as area under curve (AUC) at the 0–60, 15–60, and 30–60 min intervals after discontinuation of anesthetic administration.
The finding that smokers had higher pain scores compared with nonsmokers may be a consequence of lengthier surgery, but there may be a physiologic basis related to chronic nicotine exposure. The study of Marco et al. (11) on women undergoing cesarean delivery showed that patients who smoked had higher pain scores and opiate consumption for the first 24 h after surgery. Extrapolation from the findings of Marco et al. to patients in general must be made with caution, given the small number (10 patients) of subjects enrolled (11). In rats, while acute nicotine exposure causes antinociception, chronic exposure (14 d) produces the opposite effect, with tolerance to µ-opioid agonists and increased density of µ-opioid receptors (12).
The more frequent incidence of nausea in smokers compared with nonsmokers (Table 4) was an unexpected finding, and contradicts the results of previous studies designed to systematically evaluate risk factors for postoperative nausea and vomiting (13,14). Our results should be interpreted with caution, given the greater MAC hours of anesthesia in the smokers compared with nonsmokers, although multivariate modeling showed a persistent positive association between smoking and nausea when length of MAC hours was considered (Table 4).
We acknowledge certain limitations to this study, some of which are noted above. For example, the use of fentanyl surely limited coughing, although the doses administered did not differ between patients who received desflurane versus sevoflurane. Similarly we allowed the anesthesiologist to determine anesthetic administration including dose of induction drug (although drug choice was predetermined, and dose did not differ between groups). Finally, although the study design dictated inhaled anesthetic choice, the administered concentrations were chosen by the anesthesiologist. Again, despite allowing the concentration to be determined by the clinician, the MAC-multiples chosen did not differ between anesthetics.
In summary, in patients who smoke, the incidence and severity of respiratory complications during maintenance of anesthesia delivered via an LMA are modest, but occur more than in nonsmokers. For either smokers or nonsmokers, the incidence does not differ for desflurane versus sevoflurane. Initial recovery is more rapid with desflurane.
|
ACKNOWLEDGMENTS |
|---|
|
Footnotes |
|---|
Supported by Baxter Healthcare Corp.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMETHODSRESULTSCOMBINED RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. K. Ramachandran, P. Hans, H. Marechal, and V. Bonhomme Cough during emergence from anaesthesia Br. J. Anaesth., June 1, 2009; 102(6): 884 - 886. [Full Text] [PDF]
|
|
|
|
|
|
P. Hans, H. Marechal, and V. Bonhomme Effect of propofol and sevoflurane on coughing in smokers and non-smokers awakening from general anaesthesia at the end of a cervical spine surgery Br. J. Anaesth., November 1, 2008; 101(5): 731 - 737. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
2 analysis). Although on average more patients given desflurane coughed on induction (4.2%) of anesthesia than during emergence (1.7%), the difference was not significant. Although the incidence of coughing might be more frequent on average with desflurane or with sevoflurane for induction or for emergence, none of these differences was significant (P > 0.05).
