Site(s) Mediating Sympathetic Activation with Desflurane

After institutional review board approval of the study protocols, written informed consent was obtained from healthy, male volunteers (aged 18–27 yr), who had no history of asthma, chronic obstructive pulmonary disease, smoking, coughing with exercise, or past pneumonia. Subjects were asked to refrain from coffee, alcohol, and oral intake for at least 12 h before the study. In each study, catheters were inserted into a peripheral vein and radial artery and monitoring included electrocardiogram, arterial BP, and direct recordings of sympathetic nerve activity to skeletal muscle blood vessels (muscle SNA) via percutaneous needle placement in the peroneal nerve (microneurography) as described elsewhere [10](see also Sympathetic Microneurography and Limitations sections). Calibrated infrared spectroscopy was employed to maintain normocarbia throughout each study and to measure inspired and end-tidal concentrations of desflurane and isoflurane.

Lines and monitors were placed in seven volunteers, after which a 20-min rest period was observed. Baseline measurements of HR, BP, and SNA were then recorded and averaged over 5 min. Mean arterial pressure (MAP) was calculated as diastolic pressure plus 1/3 pulse pressure (systolic minus diastolic pressures) for each beat, averaged over each measurement interval. Anesthesia was then induced with 2.5 mg/kg propofol and 0.15 mg/kg vecuronium and the lungs were manually ventilated with oxygen and desflurane at 5.4%(inspired) via a face mask until neuromuscular blockade was established.

A 39-French right-sided, double-lumen, endobronchial tube was positioned with the bronchial cuff just distal to the vocal cords to permit selective administration of desflurane or isoflurane to:(1) the mucosa at and above the vocal cords, including the larynx, pharynx, and nasal and oral mucosa (via the tracheal lumen); and (2) to both lungs, via the bronchial lumen. The position of the double-lumen tube was confirmed by direct visualization. The bronchial cuff was inflated with air to seal the trachea during continued positive airway pressure at 20 cm H²O pressure (confirmed by listening for a leak and by looking for air bubbles arising above the bronchial cuff), while the tracheal cuff was left deflated. Bilateral lung, mechanical ventilation was maintained during control, upper airway stimulation, and recovery periods with 0.75 minimum alveolar concentration (5.4%) desflurane/oxygen (6 l/min) at 10 breaths/min with tidal volumes adjusted to maintain normocarbia. The tracheal lumen of the endobronchial tube was attached to a second anesthesia machine and continuous flow of 6 l/min oxygen was delivered through this port. A scavenging system was established over the subject's nose and mouth using a standard oxygen mask sealed with tape and attached to wall suction.

After a 20-min accommodation period, 5 min of steady state neurocirculatory measurements were collected and averaged. Desflurane (14.5%) or isoflurane (2.4%)(order of first anesthetic randomly assigned by a coin toss: desflurane = 4; isoflurane = 3) was added to the oxygen flowing to the upper airway via the tracheal lumen. Data were recorded during 5 min of administration. End-tidal anesthetic concentrations were continuously monitored from the bronchial lumen to rule out any systemic absorption from the upper airway exposure. After 5 min, the anesthetic to the upper airway was discontinued and 100% oxygen was resumed. After a 20-min recovery, the second anesthetic gas (isoflurane or desflurane) was delivered to the upper airway.

After another 20-min recovery and repeat baseline recordings, desflurane (14.5%) or isoflurane (2.4%; ordered as earlier) was abruptly administered through the bronchial lumen to both lungs while oxygen was continued in the upper airway. Five minutes of data were recorded, followed, in order, by a 20-min recovery at 5.4% desflurane, repeat baseline recordings and rapid administration of the second anesthetic gas to the lungs. The peak changes in HR, BP, and SNA from the prestimulus, anesthetized baseline were determined and Student's t tests were employed for comparisons between treatments.

Fourteen volunteers were monitored as described earlier and awake baseline data were obtained after a 20-min accommodation period. Induction of anesthesia and intubation were identical to those in study 1 with one exception: a left-sided double-lumen tube was placed in the usual position with the bronchial portion in the proximal left mainstem bronchus and the bronchial cuff placed just past the level of the carina. The double-lumen tube was pressure tested to 20 cm H²O (again checking for bubbles above the bronchial cuff), and the position was confirmed by fiberoptic visualization. Both lungs were mechanically ventilated (fresh gas flow = 6 l/min) at a rate of 10 breaths/min with tidal volume regulated to maintain normocarbia. After 20 min and achievement of stable, end-tidal desflurane concentrations at 5.4%, HR, BP, and SNA were recorded for 5 min and averaged.

The breathing system of a second anesthesia machine was purged with oxygen. The ventilator of this machine was synchronized to the same respiratory cycle as the first anesthesia machine. The second anesthesia machine was then attached to either the tracheal or bronchial lumen (randomly assigned as left or right by coin toss in each subject) while the original anesthesia machine was attached to the other lumen. Extreme care was taken to avoid tube movement, which might have stimulated a hemodynamic response. Tidal volumes for both machines were set at one half the original value. The inspired desflurane concentration of the first anesthesia machine was doubled from 5.4% to 11% desflurane while the second anesthesia machine ventilated the opposite lung with 0% desflurane in 100% oxygen. In this way, tidal volume, respiratory rate, and systemic concentrations of desflurane were held constant while the inspired concentration of desflurane delivered to one lung was doubled. Continuous neurocirculatory measurements were made over the next 5 min and averaged in consecutive 1-min epochs for further analysis.

The breathing system of the original anesthesia machine was then reconnected to both lumens of the double-lumen tube, the tidal volume was reset and the inspired desflurane was set to 5.4%. After a 20-min recovery period and repeat baseline measurements at 5.4% desflurane, a rapid increase in the inspired desflurane concentration to 11% was directed simultaneously to both lungs. A subset of four volunteers were studied on days separate from their study days to confirm the stability of systemic desflurane levels during one-lung stimulations. In these four subjects, arterial blood samples were drawn during the 5.4% baseline period, and at each minute for 10 min after desflurane increases to one lung.

Blood desflurane concentration was measured by analyzing the gas phase above the blood sample that is in equilibrium with it. The gas was analyzed in duplicate by head space analysis via gas chromatography. The assay is sensitive to 0.01 mM. The variance of known standards of the same order of magnitude as typical levels in this study is 1%.

Heart rate, MAP, and SNA responses to right versus left lung stimuli were compared using analysis of variance. There were no statistical differences between right and left lung responses and the data were combined for comparison to the two-lung condition using repeated-measures analysis of variance. Fisher's least significant difference test was used to make comparisons at equivalent time points when analysis of variance had shown significance between paired means. Significance was achieved if P was less than 0.05.

Twenty-four subjects were randomly assigned to one of three groups. Nine subjects received lidocaine anesthesia of the upper airway (airway lidocaine) consisting of pledgets soaked with 4% lidocaine and applied to the glossopharyngeal region along the glossopalatine arch and nebulized 4% lidocaine mist given while subjects breathed deeply through their mouth and nose. A 25-G needle was used to administer 1% lidocaine bilaterally to the glossopharyngeal (1 ml/side) and superior laryngeal nerves (2 ml/side). In addition, 4 ml 4% lidocaine were administered transtracheally through the cricothyroid membrane. The total dose of lidocaine administered by these combined routes was 520 mg given over approximately 15 min. A test of airway anesthesia by gentle direct laryngoscopy to view the epiglottis produced a negative gag reflex in all subjects. Eight other subjects received intravenous lidocaine followed by a continuous lidocaine infusion designed to give similar systemic lidocaine concentrations to those measured in the first phase of this study. (One subject in this group did not complete the study protocol because of bronchospasm shortly after increasing the inspired desflurane concentration. This subject experienced inspiratory and expiratory wheezing that required multiple doses of albuterol by metered dose inhaler, and terbutaline, 0.25 mg subcutaneously; the SpO²remained in the low 80's for several minutes until the bronchoconstriction was relieved. This subject's data were excluded from the group analysis.) A third group comprised seven subjects who served as time controls and received similar administration of saline to the airways.

After instrumentation and before lidocaine or saline pretreatment, a 20-min rest period was allowed. Baseline measurements of HR, BP, and SNA were then recorded and averaged over 5 min. Recordings were repeated 5 min after lidocaine/saline administration and an arterial blood sample was obtained to determine the plasma concentration of lidocaine. Anesthesia was induced as in study 1 except that ventilation was maintained via a face mask with 100% Oxygen²(6 l/min) for 2 min. The desflurane anesthetic vaporizer was then activated at 11% and ventilation was continued for the next 10 min. Neurocirculatory parameters were recorded and averaged for each minute after the increase in desflurane concentration with the exception of minutes 5–7 (to permit computer buffer downloading).

Serum lidocaine levels were measured using a homogeneous enzyme immunoassay technique. The assay is accurate to 0.1 micro gram/ml with a coefficient of variance of 0.08 micro gram/ml. Changes in HR, BP, and SNA were plotted over time after advancement of the desflurane vaporizer. Significance was assessed by analysis of variance. Fisher's least significant difference was employed to compare measurements at equivalent time points, when significant differences of the means were found with analysis of variance. Significance was set at the 0.05 level.

Sympathetic neural recordings are made from peripheral, multifiber nerves using percutaneously positioned microelectrodes. Neural recordings do not contain efferent motor or afferent stretch receptor input when the leg is relaxed in awake subjects, or after neuromuscular blockade. Skin afferents also may be recorded with this technique, however, the electrodes are meticulously adjusted to omit skin afferent signals as tested by the absence of activity in response to lightly stroking the skin and the observation of purely motor responses during electrical stimulation. Sympathetic recordings from skin sites demonstrate responses to emotion and temperature changes and the absence of pulse synchrony.

The muscle SNA recording is further qualified by determining that the nerve activity is pulse synchronous, occurring approximately 1.0–1.5 s after the arterial pressure pulse, and that the bursts are symmetrical, and are stimulated by spontaneous decreases in diastolic pressure, painful stimulation, and brief elevations in arterial carbon dioxide levels during voluntary, prolonged apnea. Further adjustments in the microelectrode position are made before initiating the experimental protocol, until burst amplitudes of muscle SNA are at least three times the background noise. A custom-written, semiautomated, computerized analysis system identifies single sympathetic bursts and quantifies their amplitude, frequency, and duration. Each computer-derived measurement is visually confirmed using a list of objective criteria developed to distinguish bursts from background noise.

In study 1, neurocirculatory variables did not differ during consecutive baseline recording periods at 5.4% desflurane before desflurane or isoflurane administration to upper or lower airways (Figure 1). Administration of 14.5% desflurane to the upper airway via a continuous fresh gas flow system led to significant increases above baseline in HR, MAP, and SNA that averaged 26 +/- 13, 23 +/- 9, and 92 +/- 36%, respectively. The administration of desflurane to the lower airway (lungs) resulted in significantly larger increases in HR, MAP, and SNA, which averaged 112 +/- 15, 48 +/- 5, and 266 +/- 92%, respectively (Figure 1). The administration of isoflurane to upper or lower airways did not lead to significant increases in any measured variable except HR, which increased 4.5 +/- 1.1 beats/min when isoflurane was administered to the lungs. A recording from one volunteer receiving desflurane to the upper airway is displayed in Figure 2.

In study 2, the selective increase in desflurane to either lung without altering systemic concentrations of desflurane was successfully achieved. Serial measurement of the blood concentrations of desflurane in four volunteers are shown in Figure 3and reveal no significant changes over time for either right or left lung stimulation. At 5.4% desflurane baseline, average HR, MAP, and SNA were 65 +/- 3 beats/min, 69 +/- 2 mmHg, and 147 +/- 36 units ([burst frequency x mean burst amplitude] per 100 heart beats), respectively (Figure 4). Increases in the desflurane concentration to either the right or left lung led to significant increases in HR and MAP but did not alter SNA (Figure 4). These changes did not differ between right and left lung desflurane stimulation periods. Heart rate increased an average of 44 +/- 8% when desflurane was increased in one lung and this increase was significantly increased to 90 +/- 6% when the desflurane concentration was simultaneously increased in both lungs. Similarly, MAP increased an average of 32 +/- 4% when desflurane was increased in either lung but this was nearly twofold larger (63 +/- 6%) when desflurane was simultaneously increased in both lungs (Figure 4). Although SNA was not significantly increased when desflurane was directed to a single lung, SNA was significantly increased 777 +/- 320% within 2 min after desflurane increases were directed to both lungs.

In study 3 there were no differences in baseline variables between the treatment groups (Figure 5). Saline administration (placebo) did not significantly alter neurocirculatory variables whereas the topical (airway) and systemic administration of lidocaine led to significant increases in HR and MAP without changes in SNA. The change in MAP was greater with systemic than with topical lidocaine administration (15 +/- 4 vs. 4 +/- 2%). The majority of subjects receiving lidocaine by either route reported feeling anxious and appeared restless. Anesthetic induction with propofol resulted in significant decreases in MAP and SNA and increases in HR in topical lidocaine and placebo groups but did not change HR or SNA in the systemic lidocaine group (Figure 5). Maximal neurocirculatory activation triggered by rapidly increasing the inspired concentration of desflurane anesthesia to 11% in the placebo pretreatment (control) group averaged 53 +/- 7%, 29 +/- 11%, and 1,584 +/- 702%, for HR, MAP, and SNA, respectively. There were no significant differences between treatment groups or placebo in the measured end-tidal desflurane concentrations or in the neurocirculatory activation over the course of the desflurane administration. Plasma concentrations of lidocaine measured immediately after the fifth minute of desflurane exposure averaged 3.6 +/- 0.5 micro gram/ml in the topical lidocaine group and 3.7 +/- 0.2 micro gram/ml in the systemic lidocaine group.

It is now clear that rapid increases in the concentration of desflurane are associated with sympathetic activation, tachycardia, and hypertension in humans, [3,5,11]and in patients with coronary artery disease these responses have been associated with myocardial ischemia. [7]A potential relationship between the pungency of desflurane, the incidence of untoward airway complications associated with administration of desflurane via mask and the neurocirculatory activation has been suggested but never explored. [3,5]The current research provides new information on the location of the sites involved in transduction of the desflurane stimulus. Study 1 employed selective stimulation of the upper and lower airway to demonstrate that both locations have sites that are activated by desflurane. The upper airway sites seem to contribute less to the total response. Study 2 indicates that pulmonary airway afferents are involved in the HR and MAP increases but suggests that a systemic site might be more importantly involved in the peripheral sympathetic response to desflurane. Study 3 indicates that when airway reflexes are presumably attenuated with lidocaine, neurocirculatory activation via desflurane is not inhibited.

One objective of this research was to localize the site at which desflurane initiates its response. The first approach was to identify whether upper airway receptors might be involved. Several studies have reported substantial airway irritability including coughing, breath holding, and laryngospasm during the administration of desflurane via mask. [9],** By sealing the bronchial lumen of a double lumen tube at a level just distal to the vocal cords, high concentrations of desflurane (or isoflurane) were selectively delivered to the region from the cords to the oropharynx and nasopharynx and exhausted via a face mask sealed around the nose and mouth. This approach revealed a small but significant neural and hemodynamic response to desflurane in the upper airways and no response to isoflurane (Figure 1and Figure 2). Systemic absorption or leakage of the anesthetic around the cuff to the lung was not involved because end-tidal anesthetic concentrations sampled from the bronchial lumen were unchanged despite administration of 14.5% desflurane or 2.4% isoflurane to the upper airway. The concentration of desflurane and isoflurane chosen for administration to the upper airway was a approximately 2 minimum alveolar concentration-equivalent dose, (although this has no clinical relevance in terms of systemic/alveolar potency because the anesthetic was not taken up systemically). The data demonstrate that sites within the larynx, oropharynx, or nasopharynx contribute to the neurocirculatory activation observed with desflurane but that the relative contribution to the net neurocirculatory activation from upper airway sites is substantially less than the contribution from lower airway regions. We had presumed that, because of the noxious nature of desflurane, the contributions from upper airway irritation would play a more significant role.

In study 2, desflurane was selectively given to the right or left lung without altering systemic desflurane concentrations. Several interesting results were obtained. First, an increase in desflurane to either lung led to increases in HR and MAP, but did not increase SNA to skeletal muscle blood vessels in the lower extremity. Second, although the surface area of the right lung is greater than that of the left lung, there was only a suggestion that right lung responses might exceed those initiated from the left lung. Third, the sum of the hemodynamic responses to right and left lungs was nearly equivalent to the HR and MAP response when both lungs and systemic sites were exposed to desflurane. This suggests that a simple summation of responses from the right and left lung occurs and this makes it difficult to propose any unusual interaction, potentiation, or inhibition when both lungs and systemic sites are simultaneously exposed to desflurane. The data also suggest that increases in SNA to the peripheral vasculature in response to desflurane contribute only minimally to the increase in the hemodynamic response. This might be caused by a powerful influence of the volatile anesthetics to relax vascular smooth muscle, thereby opposing neural vasoconstrictor influences and inducing reflex increases in sympathetic vasoconstrictor activity in the periphery. This has been apparent when higher steady-state concentrations of desflurane have been compared to equipotent concentrations of isoflurane and sevoflurane; similar BP decreases occurred despite significantly higher levels of SNA and norepinephrine. [5,6].

Peripheral recordings of SNA were only increased when both lungs and systemic sites were exposed to increases in the desflurane anesthetic. This observation can be interpreted several ways:(1) HR increases in response to desflurane might be mediated solely by withdrawal of vagal nerve activity to the sinoatrial node. Therefore, HR increases would account for the majority of the MAP increases with little or no need for sympathetically induced MAP increases; or (2) SNA activation to the heart (and perhaps other regional circulations) might differ from muscle SNA responses. The possibility of selective sympathetic responses to desflurane is consistent with our understanding of the highly differentiated sympathetic control mechanisms in humans. [12,13]The selectivity of SNA (and HR) control mechanisms in response to desflurane has been demonstrated in several studies that employed opioids or clonidine in an attempt to reduce neurocirculatory responses to desflurane. [14–16]These pretreatment strategies led to partial attenuation of the HR component but preserved or augmented SNA and norepinephrine responses to desflurane.

The SNA responses in this study suggest that there may be sites outside the airways that respond to increasing blood concentrations of desflurane or that peripheral SNA is only activated when both lungs are exposed to desflurane, leading to a large afferent neural input into the CNS. Because the solubility of desflurane permits rapid uptake into the CNS, we have speculated that there might be a transient disinhibition of centers modulating sympathetic efferent outflow. [3]In support of this, a recent study also has suggested that, in addition to airway sites, a second site mediating desflurane responses might exist in a highly perfused tissue. [17].

If airway sites are pivotal in initiating a response to desflurane then airway or systemic lidocaine might prove effective in attenuating the response. Lidocaine has been reported to lessen the hemodynamic response to laryngoscopy and tracheal intubation when applied topically or given systemically to patients. [18,19]It also reduces the bronchospasm associated with histamine. [20]Study 3 involved the selective administration of lidocaine to block neural pathways from the cranial nerves supplying the tongue and oropharyngeal mucosa. In addition, transtracheal injection and superior laryngeal nerve block with lidocaine and inhalation of aerosolized lidocaine were employed in an attempt to attenuate/block vagal afferent nerves and upper and lower airway receptors that mediate responses to airway nociceptive input. This pretreatment had no effect on the neurocirculatory activation associated with rapid increases in the inspired desflurane concentration.

To evaluate the potential effectiveness of systemic lidocaine without topical application and nerve blocks, we employed systemic infusions of lidocaine chosen to replicate the concentrations recorded during selective nerve blocks and topical lidocaine. This approach also failed to attenuate the neurocirculatory activation associated with desflurane. Large systemic concentrations of lidocaine can produce CNS excitation and seizures. The relatively high lidocaine concentrations achieved in this study probably account for the small increases in HR and BP noted before anesthetic induction and the excitation/agitation of pretreated subjects before induction of anesthesia. The lack of effectiveness of local or systemic lidocaine can be interpreted in several ways:(1) lidocaine did not block all of the afferent stimulation sites and pathways in the airways or CNS that are associated with desflurane's response. This could occur if there was insufficient distribution of the local anesthetic or insufficient density of the blockade;(2) afferent sites that are associated with the response to desflurane are present in regions outside the site of action of lidocaine; or (3) lidocaine was effective at desflurane stimulation sites, but the CNS excitatory properties of lidocaine (or desflurane) opposed or superseded the effector response.

These studies have improved our understanding of sites involved in transducing the responses to desflurane. However, one interpretation of our data is that there are multiple locations, for example, there may be systemic, and upper and lower airway locations. The efferent response profile might be altered substantially by central integration of the entire afferent profile. Selective stimulation of a single site might not give a true indication of the importance of that site in the overall response. Consistent with the speculation of multiple sites, a recently published study suggests that there may be two sites mediating responses to desflurane in humans; an airway site and a second site in a highly perfused tissue. [17].

In study 1, during which isoflurane and desflurane were given to the lungs, the rate of rise of the alveolar concentration as a percentage of minimum alveolar concentration was not evaluated, but, given the physical properties of these two agents and as evidenced by a similar approach in an earlier study, the increase in the systemic concentration of isoflurane lags behind that of desflurane. [3]However, even when fresh gas flows and delivered concentrations of isoflurane have been increased in an attempt to approach the speed of uptake of desflurane, the rate of rise of isoflurane concentration still lags behind that of desflurane. [11]Thus, we cannot rule out the possibility that a greater rate of rise of the alveolar concentration of isoflurane might have initiated some degree of neurocirculatory response.

In study 3, we attempted to anesthetize the afferent pathways innervating the airways so as to block a potential site of stimulation by desflurane. Through a series of nerve blocks, topicalization, and nebulized inhalation of local anesthetic, we aggressively exposed as much of the airway to lidocaine as we deemed possible in an awake human subject. Without a positive control we cannot be certain as to the adequacy of airway anesthesia, and this limits the interpretation of our negative results. The smaller bronchioles and alveoli were the most likely sites to have been missed with the nebulized lidocaine. We theorized that any substance such as a saline spray down the trachea as a positive control would not be reasonable in awake subjects before lidocaine treatment. We worried that an injected substance might not reach the distal airways as would a gaseous agent and that the strength of the stimulus to the airway with a positive control substance might not match the strength of the desflurane stimulus thereby misleading the interpretation of such a response. In addition, we know that the response to desflurane accommodates with repeated exposures. Stimulation with a positive control may have likewise altered the subsequent response to stimulation by a volatile anesthetic.

We erroneously assumed that lidocaine would be at least partially effective in attenuating the response to desflurane thereby simplifying our interpretation. There were virtually no differences between the responses of the blocked group and those of the placebo group, leading us to one (of three) interpretations, i.e., that airway afferents, at least from the upper segments of the lung and oropharyngeal regions, may not play a significant role in the neurocirculatory response to desflurane. This is supported by the results of study 1, which demonstrated a lesser maximal response from the upper airway. Two other recent studies lend support to our data suggesting that both nebulized and intravenous lidocaine are ineffective at attenuating the sympathetic responses to increasing concentrations of desflurane in humans. [21,22].

Neural recordings reflect sympathetic activity that regulates the tone of blood vessels within skeletal muscle of the lower leg. Because skeletal muscle blood volume accounts for 30–40% of the recruitable blood volume, the activity of these nerves has significant effects on BP. These nerves are under baroreceptor control such that even changes in BP of 1 or 2 mmHg can substantially alter their activity. While we have employed these recordings to gain access to sympathetic nerves with important physiological roles, we recognize that a single recording site may not correspond to sympathetic activity in other beds such as the splanchnic or cardiac systems. Thus, we have avoided drawing generalizations about the entire sympathetic nervous system from this single measurement technique.

Individual subjects vary quite strikingly with regard to their basal SNA. Resting SNA of young healthy persons participating in this study vary by as much as one order of magnitude from person to person. However, measurements of SNA in the same person on different days tends to be quite reproducible. Therefore, interindividual variability of SNA is much greater than intraindividual variability. This leads us to believe that some persons regulate blood pressure and systemic vascular resistance predominantly via SNA alterations, while others employ different regulatory mechanisms (e.g., changes in HR or contractility). Because of this interindividual variability, we express responses of SNA as changes rather than in absolute terms whenever possible. In addition, we attempt to have the same subject participate in all phases of the study, to control for these differences in basal SNA between subjects.

In summary, there appear to be sites in both the upper and lower airways that respond to increasing concentrations of desflurane. Lower airway stimulation initiates a large hemodynamic (HR, MAP, and SNA) response, whereas upper airway stimulation initiates an increase of lesser magnitude in these same variables. Heart rate and MAP responses to desflurane can be initiated by selectively increasing concentrations to either right or left lung without altering systemic levels of desflurane. The application of rapid increases in inspired desflurane concentration to both lungs as seen clinically, supersedes the responses from either of these localized stimulation sites. Finally, neither blockade of cranial nerves IX and X combined with topical lidocaine nor systemic lidocaine were effective in attenuating the neurocirculatory activation associated with desflurane. The possibility of stimulatory sites outside the airway (but in near proximity to the airways) and not influenced by systemic lidocaine must be considered.

The authors thank Linda D. Messana, B.S., and Toni D. Uhrich, M.S., for their careful evaluation of the data, Jill Barney, M.S., for her careful critique and editing of this work, and William T. Schmeling, M.D., Ph.D. for creative thoughts on methodological approaches to study the site of action of desflurane.

*Ebert TJ: Autonomic balance and cardiac function. Current Opinion in Anaesthesiology 1992; 5:3–10.

**Lerman J: Sevoflurane and desflurane in paediatric patients. Current Opinion in Anesthesiology 1993; 6:527–531.