Laryngeal Mask Airway and Children’s Risk of Perioperative Respiratory Complications: Randomized Controlled Studies Are Required to Discriminate Cause and Effect

Abstract

We thank Dr. Eikermann et al. for their interest in our article1and their suggestions for conducting randomized controlled studies to determine the optimal time after an upper respiratory infection (URI) for providing anesthesia and to characterize the optimal technique for airway management.First, we would like to emphasize that the work that claims the use of a laryngeal mask airway (LMA) as an alternative to tracheal intubation in children with recent URI was completed in a randomized controlled study that included 41 children in each group.2Although the incidence of laryngospasm was 10% with the LMA (twice higher than with an endotracheal tube), this did not reach statistical significance because of the low number of children involved. To our knowledge, our study is the first to report a large number of children with a recent URI and to provide some evidence that despite the use of an LMA, the incidence of respiratory complications remains high. However, we agree with the authors that multiple attempts to insert the LMA are often associated in clinical practice with difficult anatomical conditions or light or inadequate anesthesia and are independent of the presence of an URI. However, if laryngospasm were an epiphenomenon, one would observe the higher incidence of complications at insertion, which was not the case in our study because the higher odds ratio for laryngospasm was observed intraoperatively and even in the postoperative care unit. Furthermore, multiple attempts to insert the LMA were only found in the univariate analysis for all respiratory complications, whereas URI was almost the only factor that remained in the multivariate analysis, which further confirms that a recent URI is a risk factor for the occurrence of perioperative respiratory complications with the use of an LMA.1The second point raised by the authors on the time delay after a recent URI before proceeding with anesthesia is of interest. Although we agree that we cannot provide strong evidence that anesthesiologists should consider at least a 2-week interval, we still believe that the absence of evidence does not translate into the evidence of absence. Please note that the underlying pathophysiologic mechanisms involved in the occurrence of respiratory complications after insertion of an LMA are completely different than those observed in the case of an endotracheal tube. The nonadrenergic, noncholinergic autonomic nervous system is primarily involved with the LMA stimulating the sensory nerves (C fibers), whereas the cholinergic system is the main pathway that is activated by the insertion of an endotracheal tube. This explains in part why the incidence of bronchospasm is negligible in the presence of an LMA. Furthermore, we disagree with the authors that there was no “control group”per se because fever can be present independently of a URI and cough is the second most common symptom in childhood (10–20% of preschool children), not necessarily associated with a recent URI.3We based our definition of URI on the parents’ statements because parental confirmation of the presence of a cold has already been identified as a predicting factor for the occurrence of adverse events during anesthesia.4Therefore, we believe that our categorization into two groups precludes the possibility of mixing URIs of different severities.We finally agree with the authors that URI dilemma remains an issue. Please note that there is no evidence in the literature that waiting several weeks after an URI will decrease the incidence of respiratory complications. Although we agree that a randomized controlled study may definitely add some evidence to our statement, we would like to point out that prospective observational studies produce invaluable information about perioperative morbidity in pediatric anesthesia. We are aware that such studies include many confounding factors that were, however, closely examined in our study by integrating different models in the multivariate analysis. Therefore, we believe that our results are of great importance in designing for the future randomized control studies to confirm our findings without integrating these confounding factors.*Geneva Children’s Hospital, Geneva, Switzerland. walid.habre@hcuge.ch