This was a single-center, prospective, observational study with before-and-after comparisons. This study was approved by the institutional review board of Osaka University (approval no. 19142). Because all data for this study were obtained from patient medical records, the requirement for consent from individual patients was waived.
This study was performed at the Trauma and Acute Critical Care Center of Osaka University Hospital in Osaka, Japan. Our center is a tertiary care facility that receives only critically ill patients; about 1100 patients are transported to the center annually. The ED is equipped with two resuscitation rooms and one angiography room. Patients requiring hospitalization are admitted to the ICU attached to the center and are managed by dedicated intensive care physicians in the same department. For tracheal intubation, the team consists of one attending physician and at least two nurses. When an emergency medicine resident performs tracheal intubation, the team consists of the resident, an attending physician, and at least two nurses. The staff members are familiar with the use of HFNC. The study period was 1 year and 5 months, from October 2018 to March 2020. From October 2018 to September 2019, tracheal intubation was performed under conventional oxygen administration (conventional group), and from October 2019 to March 2020, tracheal intubation was performed using HFNC (HFNC group). A one-week period from September 24 to 30, 2019, was set as a training period to familiarize the medical staff with the HFNC procedure.
We included non-trauma patients of ≥ 18 years of age who required tracheal intubation regardless of having hypoxia and who were treated in the ED (i.e., those who required tracheal intubation for airway protection were included in our study). Patients with non-traumatic cardiopulmonary arrest at the time of admission or who had a do-not-resuscitate order were excluded from the study. Furthermore, patients whose records were missing SpO2 data and patients in the HFNC group in whom HFNC was not used because equipment was unavailable were excluded from the analysis.
In the conventional group, passive oxygenation and positive pressure ventilation using a bag-valve mask with a reservoir bag were provided from the time when tracheal intubation was decided. Oxygen was administered at a rate of 10 L/min. Sedatives, analgesics, and neuromuscular blocking agents were administered at the discretion of the emergency physicians. In the HFNC group, a nasal interface (Optiflow™, Fisher & Paykel Healthcare, Auckland, New Zealand) connected to HFNC (AirvoTM2, Fisher & Paykel Healthcare, Auckland, New Zealand) was attached to the patient from the time when tracheal intubation was decided. The inspiratory fraction of oxygen (FiO2) was set to 1.0 with a flow of 60 L/min. The gas was heated to 37 °C and humidified. The HFNC was used continuously until the tracheal tube was connected to the ventilator. Intubation equipment, sedatives, analgesics, neuromuscular blocking agents, and intubation maneuvers were performed according to the usual methods. All tracheal intubations in the ED were performed by board-certified emergency medicine physicians (post-graduate year > 5) and emergency medicine residents (post-graduate year 3 to 5) who had completed training in airway management. Intubation was performed using a Macintosh laryngoscope or McGRATH™ MAC video laryngoscope (Medtronic, Minneapolis, MN). We used medications for intubation recommended by the Japanese Society of Anesthesiologist . The following medications were available for tracheal intubation in the ED: propofol, midazolam and ketamine (as sedatives), fentanyl (as an analgesic), and rocuronium (as a neuromuscular blocking agent).
Demographic variables included age, sex, and body mass index. The clinical data collected in the ED included SpO2, heart rate, systolic blood pressure, respiratory rate, and Glasgow Coma Scale. Arterial blood gasses were collected immediately after arrival at the ED. The severity of the patient’s illness on admission was assessed by the Sequential Organ Failure Assessment (SOFA) score and the Acute Physiologic Assessment and Chronic Health Evaluation (APACHE) II score. The initial diagnosis on admission was noted according to the medical records. Septic shock was defined according to the Sepsis-3 criteria, and hemorrhagic shock was defined as shock with a systolic blood pressure of < 90 mmHg due to bleeding, such as gastrointestinal bleeding . Cardiogenic shock was extracted from the emergency physician’s medical record. All data were collected by an independent researcher using a standardized data collection form based on information provided in the electronic medical record, in which SpO2 was automatically recorded every 3 s during tracheal intubation. The primary endpoint was the lowest SpO2 during the intubation procedure beginning from the administration of the induction drug to tracheal tube placement. In the ED, SpO2 was measured by a BluPRO SpO2 reusable sensor (Nihon Kohden, Tokyo, Japan) placed on a fingertip. The lowest SpO2 was defined as the lowest level of SpO2 measured during tracheal intubation. The secondary outcome was the percentage of cases with SpO2 < 90% during the intubation procedure. Length of mechanical ventilation, length of ICU stay, ICU mortality, and 28-day mortality were also evaluated.
The sample size was estimated based on previous studies, assuming that the standard deviation of the lowest SpO2 was 12% and that the use of HFNC would improve the lowest SpO2 by 5% over conventional oxygen administration . We calculated an alpha error of 0.05 and a power of 0.8, assuming a dropout rate of 10% (e.g., missing data), and expected to enroll 100 patients in each group. However, enrollment was terminated mid-study due to restrictions on the use of HFNC in the ED after March 2020 due to the novel coronavirus infection epidemic. Termination was for administrative reasons. Hospital administrators banned the use of HFNC in our ED due to concerns about aerosol generation by the use of HFNC on patients with the novel coronavirus disease.
Continuous variables are shown as the median and interquartile range (IQR), and categorical variables are shown as frequencies and percentages. The Wilcoxon rank-sum test was used to test continuous variables, and Fisher’s exact test was used to test the nominal variables. A linear regression analysis, with adjustment for age, use of neuromuscular blocking agents, intubation by emergency medicine residents, and APACHE II score, was performed to evaluate the relationship between the use of HFNC and the lowest SpO2 during intubation in the conventional and HFNC groups. All analyses were performed using JMP 15 (SAS Institute Inc., Cary, NC, USA).