Cough During Emergence from Isoflurane Anesthesia

Abstract

We evaluated the effects of smoking history and albuterol treatment on the amplitude and frequency of cough during emergence from anesthesia.Before induction of anesthesia, 68 patients were randomized to receive two puffs of a placebo or two puffs of albuterol via a metered dose inhaler. Anesthesia was then induced with thiopental, fentanyl, and succinylcholine. The patients' tracheas were intubated with an 8.0 mm-endotracheal tube, and isoflurane administration was initiated. At the end of surgery, isoflurane was discontinued, and the pressure in the endotracheal tube cuff was monitored via the pilot balloon while the end-tidal isoflurane concentration was recorded. Of the 68 patients, 52 coughed before responding to command, but the incidence did not differ between smokers and non-smokers (33 of 43 vs 19 of 25), nor did it differ between albuterol-treated and untreated patients. There was no difference in the frequency or amplitude of coughs between smokers and nonsmokers, nor did albuterol affect either variable. The mean end-tidal concentration at which cough first occurred was 0.30% +/- 0.02%, and only 5% of patients coughed at values >0.6%. We conclude that 1) cough is frequent during emergence; 2) smoking does not affect emergence cough; 3) albuterol treatment does not affect emergence cough; and 4) patients are unlikely to cough at end-tidal values of isoflurane >0.6%. Implications: Most patients cough as they awaken from general anesthesia given via an endotracheal tube. In our study population, cough was frequent but generally did not occur until the end-tidal concentration of isoflurane was <0.6%. Smokers were no more likely to cough than nonsmokers, and the beta-adrenergic agonist albuterol did not prevent cough.

(Anesth Analg 1998;87:1170-4)

A smooth emergence from general anesthesia with minimal coughing is often considered a hall-mark of an experienced anesthesiologist, and clinicians generally attempt to prevent patients from coughing at the end of a procedure. Coughing during emergence from a general anesthetic may be a problem after ocular surgery or neurosurgery because of increased intraocular and intracranial pressures ^[1]. Other possible adverse effects of cough include high intrathoracic pressures resulting in increased venous and intraabdominal pressures with consequent venous bleeding. Rarely, severe coughing may also disrupt an abdominal wound closure.

Despite the nearly ubiquitous occurrence of cough in tracheally intubated patients on emergence from anesthesia, there are few studies of the phenomenon ^[2]. Many clinicians believe that cough is more common during emergence in smokers and patients with chronic lung disease, but we could find no evidence that this had been studied. The current study was therefore undertaken to determine the frequency and severity of coughing on emergence from anesthesia and to determine whether smoking history or respiratory function affects the degree of coughing. Because coughing from some causes may be suppressed by a beta-adrenergic agonist ^[3-6], we also studied whether the administration of albuterol affected the cough response during emergence.

Methods

After approval of the study by our human subjects committees, we obtained written consent from 68 patients, ASA physical status I-III, at the Veterans Affairs Puget Sound Hospital. Patients were excluded if they were being treated with a beta (_2-adrenergic) agonist, corticosteroids, a cholinergic antagonist, theophylline, or an angiotensin-converting enzyme inhibitor or if they were undergoing thoracic or lower airway surgery.

Baseline peak expiratory flow was assessed with the patient awake in a sitting position ^[7], and baseline heart rate and blood pressure were measured before anesthetic induction. Patients were randomly allocated using a coin toss to receive either albuterol or placebo from canisters that were prepared in the hospital pharmacy so that the contents could not be identified by the investigators. Two puffs were given from the chosen metered dose inhaler and were given via a spacer (Aerochamber; Monaghan Medical, Plattsburgh, NY). Anesthesia was induced with 4 mg/kg thiopental and 2 [micro sign]g/kg fentanyl IV, and intubation of the trachea was facilitated with 1 mg/kg succinylcholine IV. A cuffed 8.0-mm inner diameter endotracheal tube was placed, and controlled ventilation was set at a tidal volume of 10 mL/kg at a rate of 8 breaths/min and an inspiratory to expiratory ratio of 1:3 with inspiratory flow at 30 L/m.

Immediately after intubation and before delivery of isoflurane, respiratory resistance (R_rs) was measured, and heart rate and blood pressure were recorded.

Measurements of R_rs were made using the isovolume method after correcting for the resistance of the endotracheal tube ^[8,9]. Before each study, the pneumotachograph (Capnomac Ultima; Datex, Tewksbury, MA) was calibrated for volume using a 1-L syringe. If the surgery lasted >2 h, the dose of inhalant was repeated via an in-line spacer. Other than the initial fentanyl dose, no narcotic was given until after the study. An isoflurane/nitrous oxide anesthetic was used for 58 of the 68 cases. For 10 of the cases involving intraabdominal surgery, patients received isoflurane alone.

At the conclusion of surgery, muscle relaxation (if used) was reversed as determined by a lack of fade to train-of-four stimulation. With the patient spontaneously breathing, nitrous oxide (if used) was turned off and oxygen flow was increased to 8 L/min. The isoflurane vaporizer was then turned off.

Throughout emergence, endotracheal tube cuff pressure was continuously recorded using a pressure transducer connected to the valve of the pilot balloon. Cuff pressure was initially adjusted to 30 mm Hg, and the cuff pressure was then continuously recorded on a strip chart recorder. The patient was not disturbed further; other than a repeated verbal request ("open your eyes"), all stimulation was avoided during emergence. The time of the first cough was noted. and tracheal extubation was performed when patients responded by opening their eyes or by reaching for the tracheal tube.

The end-tidal isoflurane concentration at the time of the first cough was recorded. The number of coughs and maximal amplitude of the coughs were noted and recorded. A minimal increase of 20 mm Hg in cuff pressure was required for a pressure deflection to be counted as a cough, and each pressure deflection was considered as a single cough if there were multiple peaks without a return to baseline (Figure 1) ^[10]. Forced expiration without a prior inspiration (bucking) was not counted as a cough. These were readily identified both from the lack of significant deflection on the pressure trace and the lack of inspiration by the patient before the effort.

Figure 1: Pressure tracing obtained from the pilot balloon demonstrates the characteristic pattern associated with coughs. As described in Methods, the this trace represents three individual coughs with return to baseline.

The relative influence of smoking and albuterol treatment was examined using analysis of variance with treatment and current smoking as factors. For variables involving comparisons of proportions between smokers and nonsmokers and between albuterol and placebo groups, chi squared tests were used unless there were less than five patients in a given group, in which case Fisher's exact test was used.

Results

Demographic data for the entire study group and all subgroups (smokers versus nonsmokers and albuterol versus placebo) are presented in Table 1. Peak expiratory flow as a percentage of the predicted value did not differ between the albuterol and placebo groups but did correlate with the number of pack-years of smoking (r = -0.36, P < 0.01). The only significant demographic difference between the groups was a lower mean weight among smokers compared with nonsmokers. The patients primarily underwent peripheral procedures; a few underwent abdominal procedures. There were no differences in ASA physical status or in the proportion of patients undergoing abdominal procedures between groups (Table 1 and Table 2).

Table 1: Comparative Data for Placebo- Versus Albuterol-Treated Patients

Table 2: Comparative Data for Smokers Versus Nonsmokers

A lower R_rs value in the albuterol-treated patients documented that the administered bronchodilator was effective (Table 1).

Of the 68 patients, 52 coughed before responding to command. There was no difference between the albuterol and placebo groups in the number or amplitude of the coughs, nor was there an effect of current smoking on the proportion of patients who coughed (Table 2). We also found that cough amplitude and numbers did not correlate with pack-years of smoking, peak expiratory flow, or R_rs after intubation.

The mean end-tidal isoflurane concentration measured at the time of the first cough was 0.30% +/- 0.02% (Figure 2), not including patients who did not cough at all before awakening. There was no difference between smokers and nonsmokers, nor was there an effect of albuterol treatment.

Figure 2: Histogram of the end-tidal isoflurane concentration at which cough first occurred for 49 of the 52 patients that coughed. Sixteen patients did not cough before they responded to command. End-tidal readings were not recorded for three patients because of technical problems.

Discussion

The findings of this study are that 1) after a primarily inhaled anesthetic with isoflurane, mot patients cough on awakening and do so at a relatively reproducible end-tidal anesthetic concentration; 2) smokers are not more likely to cough on emergence than are nonsmokers; and 3) the administration of a beta-agonist does not affect the likelihood of coughing during emergence.

Studies of the respiratory reflexes to tracheal mucosal stimulation during anesthesia in humans are limited; the most comprehensive work was performed by Nishino et al. ^[11]. Using injection of distilled water at differing depths of enflurane anesthesia, they detailed six different responses to tracheal stimulation: the apneic reflex, the expiration reflex, spasmodic panting breathing, the cough reflex, the slowing of breathing, and rapid shallow breathing. Of these reflexes, the cough reflex was the most readily suppressed by anesthesia. However, most patients still coughed after the injection of distilled water at 0.7 minimum alveolar anesthetic concentration (MAC).

We did not use an active stimulus to the tracheal mucosa, but rather observed whether coughing occurred. During awakening from endotracheal anesthesia, the trachea may be stimulated by the endotracheal tube, by noxious effects of the anesthetic gas itself, or by uncleared secretions. Our study probably reflects some combination of these effects. Our patients first coughed at a mean concentration of 0.30% end-tidal isoflurane, a much lower value than that at which active stimulation caused cough in Nishino et al.'s study ^[11]. We speculate that there is some adaptation to the stimulus provided by the endotracheal tube once it has been in situ, compared with the active stimulation provided by distilled water.

We frequently observed bucking-tightening of abdominal muscles with forced expiration-before actual coughing began. This correlates with Nishino et al.'s ^[11] observed expiration reflex, which occurred at deeper levels of anesthesia. This reflex is often confused with coughing, but it occurs without a prior deep inspiration and, hence, results in little airflow and little change in intrathoracic pressure.

We used the change in cuff pressure as an indication of cough. Maximal cough pressure is a frequently used measurement of cough efficiency ^[12-14]. Cough efficiency is affected to some extent by the presence of an endotracheal tube, but a human volunteer study found that tracheally intubated volunteers still develop the peak driving pressure needed for coughing despite the inability to close the glottis ^[10]. Our pressure curves demonstrated that anesthetized patients develop the multipeak pressure tracings during cough that also characterize cough in intubated and nonintubated awake volunteers.

We attempted to determine whether there was a relationship between cough and reflex bronchoconstriction after intubation and whether current smoking was a factor affecting cough during emergence. The often simultaneous occurrence of cough and bronchoconstriction makes a connection logical. Blocking vagal conduction can block cough induced mechanically or with citric acid ^[15] and can also reduce bronchoconstriction ^[16]. Contraction of airway smooth muscle is an essential step in the initiation of cough, and isoproterenol has been found to inhibit both citric acid-induced cough and increases in airway resistance in guinea pigs ^[17]. However, other studies suggest that cough and bronchoconstriction follow separate afferent neural pathways ^[18]. Changes in osmolarity cause both cough and bronchoconstriction, but lidocaine blocks only the cough, which suggests that the two reflexes may have a common etiology but differing mechanisms ^[19]. Cough is a central nervous system-mediated reflex, but bronchoconstriction may either be reflexly transmitted via the central nervous system or result from direct mediator release ^[19,20].

We found no correlation between the postintubation respiratory resistance and either the amplitude or frequency of cough. Patients treated with albuterol had significantly lower postintubation R_rs values but did not demonstrate any difference in cough frequency or amplitude. beta-agonists have antitussive effects mediated via effects on the permeability of endothelial and epithelial tissues, the inhibition of mediator release, and the stimulation of mucociliary clearance ^[20]. Cough due to fentanyl or to inhalation of hypotonic solutions can be ameliorated with a beta-agonist ^[4], although cough induced using citric acid as an irritant is not diminished by albuterol ^[5]. We did not find any protection against the cough reflex during emergence from general endotracheal anesthesia.

Finally, we analyzed our data to determine whether smokers coughed more than nonsmokers. Smokers cough more during induction ^[21], so it seemed reasonable that the same might be true during emergence. Alternatively, it could be argued that smokers might be more tolerant of the irritants during emergence given their chronic exposure to irritant smoke. We found that current smokers could not be differentiated from nonsmokers by the end-tidal anesthetic concentration at which cough occurred or by the number or amplitude of the coughs.

The end-tidal value of 0.30% should not be interpreted as a value analogous to a MAC value because the patients were not in a steady state. Determination of MAC values requires that a steady state be reached before applying a stimulus. In the present study, the stimulus already existed, so a true MAC could not be determined. Concentrations in the neural tissues or in airway smooth muscle might differ significantly from the end-tidal concentration; therefore, the values observed should not be compared with MAC data. For the clinician, knowing the non-steady-state value is more reflective of operating room conditions. From a clinical perspective, the mean value for cough may be of less interest than concentration above which one can be reasonably confident that coughing will not begin. Of the patients, <5% coughed at end-tidal concentrations <0.6%, and none coughed at end-tidal concentrations <0.95%. These data may serve as a guideline for anesthetic concentrations at which cough is unlikely to occur.