Study design, setting, and participants
This study was a secondary analysis of the Japanese Emergency Airway Network (JEAN) Registry, a prospective observational multicenter data registry designed to characterize the current ED airway management in Japan. The study setting, methods of measurement, and measured variables were as previously described [6, 9,10,11]. Briefly, JEAN was initiated in April 2010 as consortium of 13 academic and community medical centers across Japan. All 13 EDs had emergency attending physicians, and 12 had affiliations with an emergency medicine residency program. The participating institutions were certified as Level I (n = 11) or Level II equivalent (n = 2) trauma centers and had a mean ED census of approximately 29,000 patient visits per year (range 10,000–67,000). In this multicenter observational study with an aim to describe the current airway management in Japanese EDs, we did not specify the drugs or dosages to be used. A prefixed or predetermined dose was not used across the sites. The Institutional Review Board of each participating hospital (Fukui University Hospital, Fukui Prefectural Hospital, Japanese Red Cross Medical Center of Wakayama, Kameda Medical Center, National Center for Global Health and Medicine, Nagoya Ekisaikai Hospital, Obama Municipal Hospital, Okinawa Chubu Prefectural Hospital, Osaka Saiseikai Senri Hospital, Shonan Kamakura General Hospital, Kurashiki Central Hospital, St. Marianna University School of Medicine Hospital and Niigata City General Hospital) approved the protocol with a waiver of informed consent obtained prior to data collection because it was not feasible to obtain consent from patients who require a life-saving emergency procedure-emergency airway management, and this observational study was considered a “minimal risk” study.
Data were gathered prospectively over a 29-month period (April 2010 to August 2012). Adult and pediatric patients who presented to one of these EDs and underwent emergency intubation were eligible for inclusion.
Outcome measures
The primary outcome measure was intubation success on first attempt. The secondary outcome measures were intubation success in ≤2 attempts and intubation-related adverse event rate. Intubation success was defined as proper placement of the endotracheal tube through the vocal cords with confirmation by quantitative or colorimetric end-tidal carbon dioxide monitoring [4, 6, 9, 10]. We defined the intubation-related adverse events as any adverse events that were potentially related to the procedure itself or occurred during the ED course after intubation. We did not define a specific time-period of intubation-related adverse events, as half-life of sedatives and neuromuscular blocking agents and the physiologic reserve of patients varied widely. Adverse event included esophageal intubation, main bronchial intubation, lip or dental trauma, vomiting, airway trauma, dysrhythmia, hypotension, hypoxemia and death [9]. Esophageal intubation was defined as misplacement of the tracheal tube in the upper esophagus or hypopharynx with a lapse of time and desaturation (pulse oximetry saturation <90%) before the removal of the misplaced tube. Vomiting was defined as gastric contents that required suction removal during laryngoscopy in a previously clear airway. A previous clear airway was defined as airway without visualized gastric contents during laryngoscopy. Hypoxemia was defined as pulse oximetry saturation <90% during an intubation attempt that was potentially-related to the intubation procedure, not secondary to esophageal intubation. Therefore, unchanged hypoxemia was not considered as an adverse event. Hypotension was defined as systolic blood pressure <90 mm Hg. Cardiac arrest included asystole or pulseless electric activity and cardiopulmonary resuscitation during or after intubation. By contrast, we excluded patients who already had cardiac arrest before the intubation procedures.
Data analysis
For the purpose of this analysis, we identified patients who underwent their first intubation attempt using NMB alone and those with RSI. We excluded patients involving cardiopulmonary arrest prior to emergency airway management, subsequent intubation with alternate methods when the first attempt failed. We also excluded intubation using non-direct-laryngoscopy (e.g. video laryngoscopy or fibroscopy) or the use of a gum-elastic-bougie because all cases which used these devices were intubated with only RSI.
We compared the outcomes between intubation attempts using NMB alone and RSI, by fitting two logistic regression models (unadjusted and adjusted for selected variables) using each of the three outcomes as dependent variables. A set of potential confounders was selected a priori based on biological plausibility and a priori knowledge. These included age, sex, body weight, primary indication for intubation (shock, altered mental status, or others) [2, 12, 13], and training level of intubator [4, 9, 10, 14, 15]. In the sensitivity analyses, we stratified the model by indication and intubator characteristics. All analyses were conducted with JMP statistical software version 9 (SAS Institute Inc., Cary, NC, USA).