Influence of Nitrous Oxide on Minimum Alveolar Concentration of Sevoflurane for Laryngeal Mask Insertion in Children

With approval from the Mito Saiseikai General Hospital ethics committee and written informed parental consent, we studied 120 unpremedicated children (age, 1–9 yr; American Society of Anesthesiologist's physical status I) scheduled for elective surgery under general anesthesia with the LMA ™. Exclusion criteria were a predicted or known difficult airway, acute upper respiratory tract infection, asthma or gastroesophageal reflux, or a parental/patient request for premedication. Children were randomly assigned, by opening a opaque sealed envelope, to receive 1 of 15 combinations of end-tidal concentrations of nitrous oxide and sevoflurane for inhalational induction via  a facemask: 0% nitrous oxide with 1.2, 1.4, 1.6, 1.8, or 2.0% sevoflurane; 33% nitrous oxide with 0.8, 1.0, 1.2, 1.4, or 1.6% sevoflurane; or 67% nitrous oxide with 0.4, 0.6, 0.8, 1.0, or 1.2% sevoflurane (eight patients per anesthetic combination).

An electrocardiograph, pulse oximeter, gas analyzer, and noninvasive blood pressure monitor were applied before induction. The sidestream-type infrared multigas analyzer, which measured nitrous oxide to an accuracy of ± 1.5% and sevoflurane to an accuracy of ± 0.1%, was calibrated before each use, and the inspired and end-tidal concentrations of sevoflurane, nitrous oxide, carbon dioxide, and oxygen were continually measured and recorded. Before LMA ™ insertion, the end-tidal concentrations were sampled from the nose via  a cannula using an infant circuit with an intracircuit volume of 800 ml and fresh gas flows of 6 l/min. After LMA ™ insertion, they were sampled from the distal end of the LMA ™ shaft using a cannula inserted through a self-sealing connector such that its tip was within 1 cm from the mask aperture bars. Anesthesia was induced with up to 6% sevoflurane and the designated concentration of nitrous oxide in oxygen. When the eyelash reflex disappeared, ventilation was assisted manually to maintain the end-tidal carbon dioxide partial pressure at 32–36 mmHg. The peripheral line was inserted after induction of anesthesia.

Before LMA ™ insertion was attempted, the end-tidal concentrations of sevoflurane and nitrous oxide were kept constant at the predetermined value for 15 min to allow equilibration between the alveolar and central nervous system concentrations. A single experienced LMA ™ user (S. K., > 1,000 LMA ™ uses) inserted and fixed the LMA ™ (size 1.5, 5–10 kg; size 2, > 10–20 kg; size 2.5, > 20–30 kg; size 3, > 30–50 kg) using the standard technique. 3The cuff of the LMA ™ was inflated with air to obtain 60 cm H²O of intracuff pressure.

The response of the patient was observed for 1 min after the LMA ™ insertion and classified as “no movement” or “movement.”No movement  was defined as the absence of purposeful movement of the extremities, coughing, bucking, and breath holding/laryngospasm. Movement  was defined as difficulties of mouth opening before the insertion. All responses were assessed by three independent observers (an anesthesiologist, a surgeon, and a nurse) who were unaware of the end-tidal anesthetic concentrations being used. When at least two of the observers documented any responses, the case was described as “movement.” The patients who moved were treated by deepening anesthesia with sevoflurane and/or intravenous propofol. The insertion time (from removal of the facemask to capnographic confirmation) was also recorded. Any adverse events were recorded.

MAC^LMIwas determined using a logistic regression model where P , the probability of no response, is:

where P is the probability of no movement, X¹is the end-tidal nitrous oxide concentration, X²is the end-tidal sevoflurane concentration, β⁰is the regression intercept constant, β¹is the coefficient for nitrous oxide, β²is the coefficient for sevoflurane, and β¹²is the coefficient for the product of the end-tidal nitrous oxide and sevoflurane concentrations (the interaction coefficient).

The main effects components, β¹and β², determined whether nitrous oxide and sevoflurane independently affected the response to LMA ™ insertion. The interaction coefficient, β¹², determined whether nitrous oxide and sevoflurane interacted to affect the response to LMA ™ insertion. The likelihood ratio test was used to determine which of the independent variables significantly affected the model. Age was not included in our logistic model because sevoflurane MAC remains constant in children aged 6 months to 12 yr. 4 

To determine MAC^LMI, the probability of no response in 50% of patients was evaluated at P = 0.5, and the above equation was solved for X². Likewise, to determine the concentration of sevoflurane required to prevent movement in 95% of children (E⁹⁵), the probability of no movement was evaluated at P = 0.95, and the equation was solved for X². The chi-square test and one-way analysis of variance after Bonferroni-Dunn test were used to compare the sex, age, weight, and height of the patients. P < 0.05 was considered statistically significant.

There were no demographic differences among groups (table 1). There were no differences among the three independent observers. Coefficient estimates for the logistic regression model are given in table 2. Based on the likelihood ratio test, the interaction coefficient for nitrous oxide and sevoflurane, β¹², did not differ significantly from zero (P = 0.7843) and was removed from the model. The logistic regression curves of the probability of no movement in response to intubation in the presence of sevoflurane and 0, 33, and 67% nitrous oxide are shown in figure 1. The number of move and no-move patients in each pair of nitrous oxide and sevoflurane concentrations are given in table 3. The MAC^LMIof sevoflurane without nitrous oxide was 1.57% (95% confidence limit: 1.42–1.72%). The addition of 33% and 67% nitrous oxide decreased the MAC^LMIfrom 1.57% (1.42–1.72) to 1.23% (1.07–1.39) and 0.80% (0.68–0.92), respectively (P < 0.001). The addition of 33% and 67% nitrous oxide decreased the E⁹⁵from 1.99 (1.81–2.57) to 1.70% (1.07–1.39) and 1.07% (0.94–1.50), respectively (P < 0.001). Insertion was easy in all patients. Insertion time did not differ among groups (8 ± 3 s). Immediately after the LMA ™ insertion, breath holding/laryngospasm that was unrelated to the nitrous oxide concentration or the sevoflurane concentration within the nitrous oxide groups occurred in 10 children. These patients were easily treated with intravenous propofol administration. There were no other adverse events.

We found that nitrous oxide at end-tidal concentrations of 33 and 67% were associated with a linear, dose-related reduction in sevoflurane MAC^LMIfrom 1.57% to 1.23% and 0.80%, corresponding to reductions of 22% and 49%, respectively, and that the interaction between nitrous oxide and sevoflurane was additive. Our results for sevoflurane MAC^LMIwere lower than those of Taguchi et al.  1(2.00%) but similar to those of Aantaa et al.  2(1.57%). This may be because of differences in insertion skill among the LMA ™ users participating in these trials.

Several aspects of study design can influence the validity of estimates of anesthetic potency. First, the stimulus applied by the LMA ™ should be similar and clinically reproducible. In our study, all insertions were easy and performed by a single experienced user. Higher anesthetic concentrations may be required for difficult insertions or for inexperienced users. Second, the technique used to sample respiratory gases should provide a reliable estimate of the end-tidal anesthetic concentration as the latter, at equilibrium, is taken to represent the concentration of anesthetic in the blood and brain. We took great care to minimize dead space for sampling the gases and, in all patients, a square capnograph was obtained. The equilibration time used in the current study have been validated in many previous studies. 1,2,4–9Third, appropriate mathematical methods should be applied to the dose–response data. We used logistic regression analysis, which has been shown in previous studies to yield MAC values 2,7,10,11that are similar to those determined by the method described by Dixon. 12In contrast to Dixon's approach, our study design permitted prospective randomization of all patients and yielded information about the interaction between independent variables.

The effects of nitrous oxide on volatile agent potency has been reported for skin incision (MAC^SI) 5,6and for tracheal intubation (MAC^TI). 7,8The effect appears to vary with the type of MAC and type of volatile agent. For halothane MAC^SI, 5isoflurane MAC^SI, 6and sevoflurane MAC^TI, 7the effect is linear and additive, whereas for sevoflurane MAC^SI4and desflurane MAC^SI, 13the effect is nonlinear and additive with 60% nitrous oxide reducing MAC by approximately 25% rather than 55%. Interestingly, our findings and those of Swan et al.  7show that nitrous oxide reduces the MAC of sevoflurane for instrumentation of the airway in a linear and additive fashion, but the findings of Lerman et al.  4show that sevoflurane MAC^SIis reduced in a nonlinear and additive fashion. Perhaps the influence of nitrous oxide on MAC also depends on the type of stimulus in addition to the type of MAC and volatile agent.

We conclude that nitrous oxide and sevoflurane suppress the responses to LMA ™ insertion in a linear and additive fashion in children.