As a result of the COVID-19 pandemic, unprecedented numbers of patients required advanced support for respiratory failure worldwide.1,2,3 Endotracheal intubation in COVID-19 patients is a highly specialized procedure4 that carries risk both to patients and attending healthcare workers (HCWs). Healthcare workers involved in airway management for patients infected with coronaviruses may be at high risk of infection due to aerosol and droplet exposure during the procedure. During the severe acute respiratory syndrome coronavirus-1 (SARS-CoV-1) outbreak, HCWs performing endotracheal intubation had a relative risk [RR] of 13 for developing SARS compared with others caring for these patients.5,6 Initial data from the COVID-19 pandemic (caused by SARS-CoV-2) suggests that 10% of HCWs developed an infection, although causality with professional exposure as opposed to social exposure could not be established with certainty.7 Expert recommendations were released to guide safe endotracheal intubation in these settings, and included directions on team composition and dynamics, use of videolaryngoscopy, and rapid sequence induction (RSI) with administration of hypnotics and neuromuscular blocking agents (NMBAs).8,9 Given the underlying physiologic compromise in critically ill patients, the process is associated with a significant risk of life-threatening severe adverse events (SAEs), such as hypoxemia, hypotension, and cardiac arrest.3,10,11,12 To optimize resources and reduce the risk to patients and HCWs, expert airway teams were developed in several centres in Canada and in other countries to perform endotracheal intubations in COVID-19 patients. The structure of these airway teams differed across countries and healthcare systems. A group from the UK recently published the experience of a London hospital,13 showing that a highly-skilled designated intubation team utilizing a protocolized, early tracheal intubation model may improve patient and staff safety.

To the best of our knowledge, the Canadian experience with similar specialized teams has not been reported to date. In this retrospective study, we describe patient outcomes at two academic hospitals in Toronto where airway management teams were developed and responsible for urgent airway management in all patients with suspected or known COVID-19. Our hypothesis was that the establishment of designated airway teams would be associated with a high rate of successful first-attempt intubation and maintain patient and provider safety.

Methods

At the onset of the pandemic, a dedicated rapid response team for airway management of all COVID-19 suspected or confirmed patients (internally named the “Emergency rapid intubation team” [ERIT]) was developed at both Toronto General and Toronto Western Hospital, and its activities were designed to follow a shared protocolized approach. Prior to the institution of this team, airway management in the emergency rooms (ERs) and intensive care units (ICUs) was provided by ER and ICU physicians, with anesthesiologists available as back-up in case of an unanticipated or anticipated difficult airway. On medical and surgical wards, airway management was typically provided either by ICU physicians or, in case of code blue scenarios, by an anesthesiologist or supervised anesthesia trainee (resident or fellow). Videolaryngoscopy was not the routine first-line device prior to ERIT, and RSI was not the standard in all cases, with the choice of device and medications at the discretion of individual providers.

The ERIT was active 24 hr a day and seven days a week. The ERIT team was composed of one attending anesthesiologist, two operating room registered nurses (OR RNs), one anesthesia assistant (AA), and one patient attendant, with each team working a 12-hr shift. Prepacked boxes containing the required equipment were developed and stored in a designated location. Dedicated pagers and overhead public address system activation codes were utilized to request ERIT assistance, and the team responded to requests from anywhere in the hospital, including ICUs and ERs.

Didactic teaching and simulation sessions were attended by team members before the “go-live” date. The team would meet at the beginning of each shift to review algorithms and simulate clinical scenarios, including donning and doffing of personal protective equipment (PPE), which was in accordance with the recommendations from the World Health Organization (WHO).14 The roles of each team member were preidentified and designated as follows: airway manager (anesthesiologist), assistance with airway equipment and airway management (AA), assistance with medications, monitoring, hemodynamics (OR RN1), “runner” and safety officer for PPE donning and doffing (OR RN2), and assistance with provision of additional equipment and transportation (attendant). Procedures were documented on standardized clinical forms.

The protocol included hemodynamic optimization, preoxygenation, use of RSI (for NMBAs, either succinylcholine at 1.5 mg·kg−1 or rocuronium at 1.2 mg·kg−1 were recommended; the choice of hypnotics was left to the discretion of the attending anesthesiologist and included one or more of the following agents: ketamine, propofol, midazolam, and fentanyl) and videolaryngoscopy as a first-line device (Storz C-MAC® airway management cart with reusable blades, KARL STORZ SE & Co. KG, Tuttlingen, Germany). The blade size and type (Macintosh-blade or D-blade) used for the initial attempt was chosen by the individual anesthesiologist. We did not routinely use apneic oxygenation during intubation with low-flow or high-flow nasal oxygen. Following recommendations from experts published at the onset of the pandemic, our strategy included intubation early in the course of illness for patients with respiratory failure, as judged by the attending physician. Typically, intubation was performed in patients with any of the following: persistent or worsening tachypnea (RR > 25) or a trajectory of increasing respiratory distress despite medical management, marginal and/or rapidly deteriorating oxygen saturations despite supplemental oxygenation, and bilateral infiltrates on chest imaging.15 The clinical evaluation of potential risk factors for difficult airway and hemodynamic instability was performed at the discretion of the attending anesthesiologist.

We retrospectively collected data on all airway management procedures performed by the ERIT. The clinical registry was used to identify patients who received care by this team, and a retrospective chart review was conducted collecting data on demographics, specifics of airway management (e.g., technique, location, number of attempts) and related SAEs (hypoxemia, hypotension, and cardiac arrest within 15 min of intubation) in confirmed or suspected COVID-19 patients at participating centres. Anonymized data were entered in a centralized, encrypted and password-protected database. The study was approved by the University Health Network Research Ethics Board (REB 20-5620, 2 August 2020). This study conforms with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines16 and was preregistered prior to commencement in January 2021 of data extraction and analysis under ClinicalTrials.gov Identifier NCT04689724 (registration date, 30 December 2020).

Data were examined for missing and incomplete data. Results are presented as means (standard deviation) for continuous variables or absolute values with percentages (%) for count data. Contingency table data were compared using the Chi square test for cell count values equal to or greater than 5, and Fisher’s exact test for cell counts less than 5. An exploratory analysis examining the association of preoperative patient characteristics and induction medication dosing with outcomes was conducted. The co-primary outcomes of interest were hypotension (defined as a systolic blood pressure (SBP) < 90 mm Hg within 15 min of induction) and hypoxemia (defined as a postinduction oxygen saturation < 90% within 15 min of induction). Relative risk ratios were calculated for individual predictors using log-binomial regression, given the incidence of our outcomes of interest was > 10%.17 Induction medication dosing, preinduction hypoxemia, and preinduction hypotension were examined as predictors of interest in separate models, each adjusted for the baseline patient characteristics of age, sex, and overall condition (awake, drowsy, or cardiopulmonary arrest), which were a priori specified given our limited number of outcome events.18

Results

From 3 April to 17 June 2020, 123 COVID-19 patients (suspected or confirmed) with respiratory failure were assessed by the specialized airway teams at the two hospitals, with 117 (95%) ultimately requiring airway management. Table 1 summarizes the details. Incomplete data were noted in five (4%) procedures.

Table 1 Characteristics of patients assessed by the emergency response intubation team
Full size table

A total of 43 anesthesiologists were involved as the primary airway managers in at least one intubation during the study period. The majority of airway management procedures (n = 70/117; 60%) were performed in the ICU. A total of 29 (25%) were performed in the ER, and 20 (17%) in medical/surgical wards (either high acuity step down unit beds [level 2], or regular ward beds) (Table 2). One instance of airway management (1%) was conducted in the OR.

Table 2 Intubation location characteristics and induction strategy
Full size table

Procedure

Rapid sequence induction (predetermined induction doses, high-dose neuromuscular blockade, no bag-mask ventilation, with or without cricoid pressure) or modified RSI (predetermined induction medication doses with some element of titration depending on patient response, no bag-mask ventilation) was employed in all cases. A videolaryngoscope was the final successful device used for airway management in most cases requiring endotracheal intubation (109/117; 93%), with a high first-attempt success rate (108/117; 92%) (Table 2). In seven cases (6%), there was a pre-existing endotracheal airway and an exchange was required. Out of the nine intubations with more than one documented attempt, 7/9 (78%) achieved success with videolaryngoscopy (Table 3). More than two attempts at intubation were required in 2/117 (2%) cases; one intubation required three attempts and another required four attempts (Table 3). A flexible reusable bronchoscope was required in 2/117 (2%) cases, and a tracheostomy (tube exchange) was involved in one case (1%) (Table 4).

Table 3 Outcomes of patients within 15 min of airway management
Full size table
Table 4 Details of intubations requiring more than one attempt
Full size table

Serious adverse events

Hypoxemia

Within 15 min of airway management, hypoxemia (peripheral oxygen saturation [SpO2] < 90%) occurred in 28 patients (24%), with severe hypoxemia (SpO2 < 70%) present in ten patients (9%). Patients with preinduction hypoxemia had a higher RR of postintubation hypoxemia in unadjusted (RR, 2.39; 95% confidence interval [CI], 1.26 to 4.54; P ≤ 0.01) and adjusted models (aRR, 2.17; 95% CI, 1.08 to 4.35; P = 0.03). There were only 9 (8%) situations where multiple attempts were required for airway management. With multiple intubation attempts, the incidence of positive pressure ventilation prior to intubation was more common (2/9; 22%) compared with single intubation attempts (11/108; 10%), although the difference was not statistically significant (Fisher’s exact test, P = 0.28). The single incidence of supraglottic airway use occurred in an individual who had hypoxemia prior to induction of anesthesia and required multiple intubation attempts; hypoxemia was resolved after successful intubation.

Hypotension

Hypotension within 15 min of airway management (SBP < 90 mm Hg) occurred in 37/117 patients (32%), and severe hypotension (SBP < 65 mm Hg) in 11/117 patients (9%) (Table 3). Patients with pre-existing hypotension prior to intubation were more likely to have hypotension after intubation in unadjusted models (RR, 1.79; 95% CI, 1.05 to 3.06; P = 0.03], but not after adjustment for age, sex, and preinduction clinical condition (aRR, 1.55; 95% CI, 0.93 to 2.60; P = 0.09). A total of 51/117 (44%) patients received phenylephrine boluses at the time of induction, with a mean (SD) dose of 362 (39) µg and a dose range of 80–2,200 µg. A total of 12/117 patients (10%) received ephedrine with a mean (SD) dose of 17 (11) mg, and a range of 10–50 mg, and 8/117 patients (7%) required epinephrine boluses, with a mean (SD) dose of 269 (324) µg, and a dose range of 20–1,000 µg. Centrally infused vasopressors at the time of induction were used in 15/117 (13%) patients, with a dose range for norepinephrine of 0.03–0.3 µg·kg−1·min−1, a dose range for vasopressin of 2–6 units·hr−1, and a dose range for epinephrine of 0.2–0.3 µg·kg−1·min−1 (Table 5). The type of hypnotic used for induction was not associated with postinduction hypotension in either unadjusted or adjusted models. Neither was the total medication dose used for the agents fentanyl, midazolam, or propofol; however, there was a statistically significant association between the total ketamine dose used and hypotension in both unadjusted and adjusted models (Table 6).

Table 5 Induction medication details
Full size table
Table 6 Association of induction medication doses with postinduction hypotension
Full size table

Cardiac arrests

No new cardiac arrests were reported after induction and intubation (Table 4). At the time of ERIT team arrival, three (3%) patients were actively receiving chest compressions, and one patient (1%) was documented as being periarrest, although it was unclear if chest compressions had been initiated prior to ERIT activation, and were not required after team arrival.

Protective personal equipment (PPE)

Gloves, N95 masks, face shields, or goggles and caps were used in 100% of cases by the professional conducting the intubation. A gown was used in 98% of cases, and 100% of procedures were supervised by a designated “spotter” (Table 7).

Table 7 Personal protective equipment and practices among team members
Full size table

Discussion

We report the results of a retrospective study on the airway management of patients with suspected or confirmed COVID-19 requiring airway management at two academic hospitals during the first wave of the pandemic in Canada.

In this cohort managed by a designated expert team, we observed widespread use of videolaryngoscopy as a first-line device, high first-attempt success rates, and very high adherence to recommended PPE practices. These data appear overall in line with reports from other COVID-19 airway management teams.13 While accurate comparisons are not possible and some differences in the definitions exist, a previous, large epidemiological study in critically ill patients that included management by a variety of healthcare teams12 reported peri-intubation rates of cardiovascular instability of 42%, hypoxia (SpO2 < 80%) of 9%, and difficult intubation (intubation requiring more than two attempts) of 5%. Such findings reinforce the impact of pre-existing physiologic compromise on airway management in the critically ill.19 Important planning considerations include patient tolerance of the apneic interval that precedes intubation and ventilation, and the potential of medications used for anesthesia induction to contribute to hemodynamic instability.19 The choice of induction agent was not generally observed to have a significant association with hypotension in this study; however, a weak association was observed between increasing ketamine dose and postinduction hypotension. This may be related to the known myocardial depressant effects of ketamine, or may be due to residual confounding related to ketamine being chosen as an induction agent in sicker patients.20,21 Patient harm can be minimized by anticipating and preparing to manage these adverse effects, as well as having skilled and experienced team members working as a cohesive unit.22 This is borne out by the absence of periprocedural cardiac arrest in our population, which was associated with prompt administration of supplemental vasopressors in appropriate doses.

When we instituted the ERIT system at our centres, we decided to involve only experienced airway management providers to minimize risks to team members and patients. Hence, only consultant anesthesiologists were the designated airway manager on ERIT. For similar reasons, it was decided that the first-line device would be a videolaryngoscope, and that RSI would be employed in all cases. These factors may have contributed to the high first-attempt and overall success rates observed. Indeed, in the INTUBE study, having anesthesia as primary specialty (odds ratio, 0.53; 95% CI, 0.41 to 0.69), and the use of a videolaryngoscope (odds ratio, 0.60; 95% CI, 0.42 to 0.85) were significantly associated with a reduced likelihood of first-pass intubation failure.12 We observed a near-complete adherence to PPE protocols, with higher rates compared with other experiences described in the literature. In an international observational study of more than 4,000 emergent tracheal intubations in COVID-19 patients, Wong et al. reported that PPE in compliance with WHO standards was utilized in 87.5% of cases,23 compared with over 98% in our centres. Canadian data from the IntubateCOVID database showed that WHO standards were not met in 3.7% of procedures.24 The repeated simulation training and the presence of a “spotter” who ensured proper PPE practice among team members may have played an important role in these figures.25,26,27 Nevertheless, even with these considerations in mind, there were (albeit very rare) breaches in best practices (such as not wearing a gown), and providers should always be mindful of the importance of adequate PPE.

The study has several inherent limitations that merit discussion. First, those related to the retrospective design and limited sample size of the study. Second, we present the experience of two academic centres, part of the same institution, which adopted a similar approach. The findings may not necessarily be generalizable to other settings. Third, the periprocedural data were extracted from a standardized clinical form that was filled in by one of the ERIT team members immediately following the event, rather than by an independent observer; hence, there is a potential for reporting and observer bias. Additionally, we did not capture data regarding the clinical evaluation by the attending anesthesiologist of potential risk factors for difficult intubation.

We also have no direct comparison data from intubations in the critically ill population at our institution conducted by non-specialized airway teams, limiting our capability to ascertain the impact of the airway team per se on patient outcomes or HCW safety. No data are available on the incidence of COVID-19 infection in ERIT HCWs. We had initially planned, with REB approval, to interview team members to assess whether they had developed signs, symptoms, or confirmed infection related to ERIT activities; however, the response rate to a call for voluntary participation in this interview was too low to provide meaningful information.

Conclusion

We present data from a Canadian experience with a consultant anesthesiologist-led specialized team for airway management in patients with COVID-19 experiencing respiratory failure. A highly protocolized team-based approach was used, and only experienced providers performed airway management. In this setting, we observed high adherence to PPE recommendations, high first-attempt success rates at intubation, and periprocedural rates of severe hypotension and hypoxemia of 9% each. We believe that this model of emergency airway management in critically ill patients could be readily implemented in other Canadian settings, and may be beneficial for provider and patient safety during future pandemics or in subsequent waves of the present COVID-19 pandemic.