Safe and smooth induction of anaesthesia with volatile anaesthetics is particularly important in the practice of clinical anaesthesia. Stimulation of the airway during induction with mild pungent volatile anaesthetics such as enflurane and isoflurane may elicit respiratory reflexes including breath-holding, coughing and laryngospasm . These responses may not only delay the induction of anaesthesia but may occasionally put the patients at risk for hypoxaemia. Volatile anaesthetics which are not pungent allow more rapid and smoother induction [2,3], compared with pungent volatile anaesthetics. It has been recently reported that stimulation of the nasal mucosa with high concentrations of enflurane, isoflurane, and halothane elicits respiratory reflex responses and changes the respiratory pattern in lightly anaesthetized humans. This suggests that the nose can be an important reflexogenic site during induction of anaesthesia with a pungent volatile agent . Assuming that the nose is a major source of the respiratory reflexes, observed during anaesthetic induction, it is possible that inhalation of volatile anaesthetics through the oral route may attenuate this reflex response.
The recent study of Doi and Ikeda  compared the airway irritation produced by the four anaesthetic agents, halothane, enflurane, isoflurane and sevoflurane, and showed that sevoflurane is the least irritant of these anaesthetics. In addition, Yurino and Kimura  reported that sevoflurane produces a more rapid and smoother induction of anaesthesia than does isoflurane. In these studies, however, the subjects were breathing through a facemask, and the effects of the breathing route on respiratory responses to inhalation of volatile anaesthetics were ignored.
The present study was designed to test the hypothesis that the effects of nasal inhalation of sevoflurane on the breathing pattern during induction of anaesthesia may differ from those of oral inhalation and that the oral breathing route may provide more rapid and smoother induction of anaesthesia than does the nasal breathing route.
Patients and methods
After obtaining approval from the Institutional Ethics Committee and informed consent, 20 ASA Class I or II patients were studied. All patients were scheduled for elective surgery under general anaesthesia. None had clinical evidence of any respiratory, cardiovascular, neuromuscular, or upper airway disease.
The patients received 50 mg of hydroxyzine and 0.5 mg of atropine sulphate intramuscularly 30-45 min before induction of anaesthesia. Patients were randomly allocated into two groups: in one anaesthesia was induced by inhaling sevoflurane through the nose (N-group) or in two through the mouth (O-group). The two groups were similar with no significant differences in age, sex, weight or height (Table 1).
In the operating room, each patient in both groups rested in the supine position. Heart rate was monitored continuously (ECG). Non-invasive measurement blood pressure recordings were made at 5 min intervals and arterial oxygen saturation (Spo2) was monitored continuously with a pulse oximeter (Oxypal, Nihon Koden, Japan). The subject was fitted with a tight fitting custom-made partitioned facemask (Senkousha, Japan). The mask, which was used in a previous study , has a hard rubber septum between nose and mouth to construct separate nasal and oral chambers, each with its own opening port (internal diameter 9 mm, length 40 mm). The rubber septum was positioned under the upper lip and on the upper gum of each patient so that the lips were apart and the mouth slightly open. Each mask chamber was tested separately for air leaks by pressurization to ±10 cmH2O for 10 s. Dead space in the nasal and oral chambers were 90 and 120 mL, respectively. Patients in N-group could breathe only through the nasal port; the oral port being occluded with a rubber stopper. Patients in O-group could breathe only through the oral port; the nasal port being occluded. The open port of the mask was connected to an experimental apparatus incorporated in a semiclosed anaesthetic system. Administration of oxygen was started at a flow rate of 6 L min−1. Patients were allowed to breathe oxygen through the mask for at least 5 min to acclimatize to the mask and to achieve denitrogenation. Ventilatory airflow was measured using a Fleisch pneumotachograph (No. 2), and tidal volume (VT) was obtained by electrical integration of the inspired flow. Endtidal carbon dioxide tension (PETCO2) was continuously monitored using an infrared carbon dioxide analyser (Capnomac Ultima, Datex, Finland). Before induction of anaesthesia, each patient was instructed to keep his or her eyes open for as long as possible. When the breathing pattern was stable, sevoflurane (5%) was introduced using a precalibrated vaporizer incorporated in the anaesthesia machine, while the concentration of sevoflurane was monitored with an anaesthetic gas monitor (Capnomac Ultima, Datex, Finland). All respiratory variables were recorded with a Nihon Kohden eight-channel recorder (RJG-4128).
In each group, respiratory responses to inhalation of sevoflurane were analysed with respect to changes in the inspiratory time (TI), expiratory time (TE), respiratory frequency (f), VT, and minute ventilation (V˙I). Control values for TI, TE, f, and VT were obtained by averaging the values for the 10 breaths immediately before the inhalation of sevoflurane. After the start of sevoflurane inhalation, values for the respiratory variables were measured breath-by-breath from the first to the fifth breath. Those from the sixth breath were obtained every 10 s by averaging the values for the breaths observed in each 10 s. V˙I was obtained by calculation from f and VT. Measurements of respiratory responses were continued until an obstructive respiratory pattern emerged. Time from the start of inhalation of sevoflurane to onset of sleep (TI-S) and time from the start of sevoflurane to emergence of obstructive breathing pattern (TI-OB) were measured and the groups were compared. Onset of sleep was defined as the time at which patients spontaneously closed their eyes with no response to verbal command. This command was repeated every 10 s until the subject no longer responded. Obstructive respiration was defined as paradoxical movement of rib cage and abdomen.
On the second or third post-operative day, all patients were asked whether they had noticed the smell of sevoflurane, felt difficulty in breathing, or would be willing to undergo the same anaesthetic induction again.
Statistical analysis was performed using Student's t-test, χ2-test, and analysis of variance (two-way) followed by Dunnett's test for comparisons between control values and those obtained after the start of sevoflurane inhalation. P<0.05 was considered statistically significant.
No airway reflex responses such as breath-holding, coughing and laryngospasm were observed during anaesthetic induction in any subject. No complication occurred in N-group but there was one episode of movement and excitement during obstructive breathing in O-group. Figure 1 shows an example of respiratory responses to inhalation of sevoflurane in one patient from each group. After the start of inhalation of sevoflurane, there was a precipitate decrease in VT in the patient who inhaled sevoflurane through the nose, although this response decreased rapidly. In the patient who inhaled sevoflurane through the mouth, however, the decrease in VT immediately after the start of inhalation of sevoflurane was less remarkable. Inhalation of sevoflurane made little difference to TI and TE.
Figure 2 shows changes in TI, TE, VT and V˙I during inhalation of sevoflurane in all patients. There were no significant differences in control values of TI, TE, VT, and V˙I between the two groups. Compared with the control value, a significant decrease in VT (P<0.01) was observed at the second breath (b2) and the third breath (b3) after sevoflurane was administered through the nasal route. This change, however, was decreased after the fourth breath (b4). In contrast, during administration of sevoflurane through the mouth, there was no significant decrease in VT after the start of sevoflurane inhalation. At the third breath after the start of sevoflurane inhalation, V˙I was significantly decreased in N-group (P<0.01), whereas in the O-group no significant decrease in V˙I was observed.
Figure 3 shows changes in end-tidal concentration of sevoflurane (FETSEVO) after the start of inhalation of sevoflurane. The end-tidal concentrations of sevoflurane during inhalation through the mouth were significantly higher than those through the nasal route at 10, 20, 30, and 40 s after the start of sevoflurane inhalation.
The average value of TI-S in N-group was significantly longer than that in O-group (N-group, 115±8.4 s; O-group, 86.2±4.4 s [mean±SE], P<0.01). The average value of TI-OB in N-group was significantly longer than that in O-group (N-group, 172±13.4 s; O-group, 118.1±9.7 s [mean±SE], P<0.01).
Table 2 shows the post-operative interview responses. There was no significant differences between the two groups. One patient in the N-group hopes for a different type of induction of anaesthesia next time.
In the current study we examined the respiratory pattern during anaesthetic induction with sevoflurane by inhalation through the nose or through the mouth. Our results show that the inhalation of a high concentration of sevoflurane, by either route, did not cause adverse respiratory responses such as breath-holding, coughing or laryngospasm. The absence of respiratory responses to sevoflurane inhalation is compatible with the fact that sevoflurane produces little airway irritation . Also, at post-operative interview, all the patients, except one who breathed through the nose, answered that they would hope to have the same type of anaesthetic induction again. Thus, the induction of anaesthesia with a high concentration of sevoflurane, whether it is administered through the nose or the mouth, appears to be an acceptable method for clinical use. Despite the absence of respiratory reflexes during sevoflurane inhalation, our data showed that there were considerable differences in respiratory pattern during nasal and oral inhalation of sevoflurane. Thus, although neither respiratory frequency nor respiratory cycle duration changed, minute ventilation (V˙I) and tidal volume (VT) immediately decreased during inhalation of sevoflurane through the nose but not when through the mouth. The precise mechanism for this difference was not examined in this study, but the decrease in VT during nasal inhalation of sevoflurane was likely to be due to stimulation of the nasal mucosa by sevoflurane. In fact, there is evidence to suggest that stimulation of nasal mucosa with commonly-used volatile anaesthetics such as isoflurane, enflurane and halothane can elicit nasal reflexes and cause changes in the breathing pattern . Although when compared with these other anaesthetic agents, sevoflurane is considered to be less irritating to the airway mucosa , it is still possible that sevoflurane can stimulate the nasal mucosa and elicit the nasal reflexes. Our finding that oral inhalation of sevoflurane caused little or no change in the breathing pattern may support this notion.
It was also shown in this study that both TI-S and TI-OB during inhalation through the mouth were significantly shorter than during inhalation through the nose, suggesting that inhalation of sevoflurane through the mouth produces induction of anaesthesia more rapidly than nasal inhalation. The finding that FET SEVO was significantly higher during oral inhalation than during nasal inhalation for 40 s after the start of sevoflurane inhalation is compatible with these observations.
This difference in the rate of rise of anaesthetic concentration between nasal and oal inhalation of sevoflurane can be partly explained by a difference in the dead space of the nose and mouth. It is apparent that, for a given minute ventilation, the greater the dead space, the less the alveolar ventilation. In this context, it is worth noting that the more the alveolar ventilation decreases, and more slowly the alveolar sevoflurane concentration rises. Tanaka et al. have shown that the dead space during nasal breathing is 40 mL greater than that during oral breathing. Although the dead space in the nasal chamber of our partitioned mask is 30 mL less than that in oral chamber, there is a possibility that the dead space during nasal breathing was substantially greater than that during oral breathing.
Another possible explanation for the difference in anaesthetic concentrations during nasal or oral inhalation of sevoflurane is that the immediate decrease in ventilation at the second and the third breath, after nasal inhalation of sevoflurane, delays the rise in anaesthetic concentrations in the anaesthetic circuit and thereby delays induction.
As ventilation is easily influenced by voluntary control under experimental conditions , particularly when the subjects' attention is focused on their breathing, the difference in breathing pattern between nasal and oral inhalation of sevoflurane might have been caused by conscious intervention. However, the decrease in VT during nasal inhalation of sevoflurane was consistently observed within 10 s after the start of sevoflurane inhalation, suggesting that the change was more likely to have been produced by a reflex effect. Furthermore, in order to minimize the possible effects of the subjects' attention to breathing, we instructed the subjects to keep their eyes open to distract their attention from breathing.
In conclusion, our results show that the effects of nasal inhalation of sevoflurane on breathing pattern are slightly different from those of oral inhalation of sevoflurane, and that of the two, oral inhalation of sevoflurane may induce anaesthesia more rapidly.