FormalPara Clinician’s capsule

What is known about the topic?

 Concern over transmission of COVID-19 during emergent orotracheal intubation has increased the complexity of this already high-risk procedure.

What did this study ask?

 Does implementation of a “COVID-protected RSI Protocol” in the ED and Critical Care settings impact intubation-related outcomes?

What did this study find?

 First-pass success rates increased from 73 to 82% following the introduction of a “COVID-protected RSI Protocol”.

Why does this study matter to clinicians?

 A “COVID-protected RSI Protocol” was implemented and associated with improved first-pass success rates and decreased adverse events.

Introduction

Approximately 1–2% of people who contract COVID-19 become critically ill [1], many requiring intubation and mechanical ventilation. Emergency intubation has been associated with an increased risk of adverse events [6,7,8,5], especially if there are multiple or prolonged attempts [10,11,8]. Numerous algorithmic approaches have been published [13,14,11] to mitigate these events. The risk of transmission of COVID-19 to the healthcare team during intubation resulted in a rapid shift from protocols focused exclusively on patient outcomes to protocols focused on both patient and healthcare worker (HCW) safety [12]. This resulted in the implementation of numerous novel protocols and guidelines for the safe intubation of patients with or at risk of having COVID-19 [16,17,18,19,20,21,22,23,24,25,22]. To date, these protocols have been site-specific, variable, and based on extrapolation of previous knowledge to new clinical contexts.

Prior to the COVID-19 pandemic, a prospective observational study characterizing current practices and patient outcomes related to emergent intubation by emergency medicine (EM) and Critical Care teams was already underway at our center. During this study, a natural experiment emerged when in response to the pandemic, the previously varied approaches to emergent intubation were quickly standardized to a “COVID-Protected Rapid Sequence Intubation (RSI) Protocol”. This protocolized approach was developed by local airway management experts and included multiple equipment, personnel, technique, and implementation considerations including an in situ simulation approach (see Appendix 1). The resultant study aimed to answer the question; in adult ED patients undergoing emergent intubation by EM and Critical Care teams, does the use of a “COVID-Protected RSI Protocol” impact first-pass success or other intubation-related outcomes?

Methods

Study design and time period

We conducted a single-center, prospective observational, quasi-experimental study. Data were prospectively collected on consecutive emergent intubations performed by the EM and Critical Care teams before and after implementing a standardized “COVID-Protected Rapid Sequence Intubation Protocol” from September 2019 to November 2020. The protocol was introduced in April 2020. Ethics approval for this study was obtained from the University of Manitoba Health Research Ethics Board. Patient consent was not required.

Study setting

This study was performed in Winnipeg, Manitoba at the Health Sciences Center, a provincial tertiary care center for trauma, transplants, burns, neurosciences, complex cancer care, and pediatrics, serving approximately 1.3 million people.

Study population

Patients were eligible to be included in the study if they presented to or were admitted to the Health Sciences Center and required emergent intubation. Consecutive adult patients (≥ 17 years old) intubated by EM or Critical Care teams were enrolled. Patients intubated in the operating room were excluded.

Intervention

The intervention of interest was the “COVID-Protected RSI Protocol”, hereafter referred to as “the protocol”. This protocolized approach to emergent airway management was institutionally adopted for all emergent intubations performed by EM or Critical Care teams at our institution in response to the COVID-19 pandemic. The protocol required full AGMP personal protective equipment and included an in situ simulation training rollout which was provided to all intubating services, an airway checklist, RSI by an experienced intubator, video laryngoscopy (VL), and avoidance of hand-ventilation (see Appendix 1).

Outcome measures

The primary outcome was first-pass success for intubations performed by EM or Critical Care teams before and after institution of the protocol. Similar to other studies, first-pass success was defined as the successful intubation of the trachea with a single insertion of a laryngoscope blade [23]. Pre-specified secondary outcomes included best Modified Cormack–Lehane view, post-intubation hypoxemia, post-intubation hypotension, esophageal intubation, cannot intubate/cannot oxygenate scenarios, CPR post intubation, Intensive Care Unit (ICU) mortality, ICU length of stay (LOS), and total mechanical ventilation days.

Data collection

We developed and pre-tested a standardized data collection form that was completed by the Respiratory Therapist (RT) attending to each intubation (see Appendix 2). The RT would liaise with the physician performing the intubation for clarification of details as required. These forms were stored in the RT department offices. Additional variables were extracted through a paired health records review.

Data analysis

All statistics were conducted using SAS version 9.4 (SAS Institute, Cary NC). Baseline characteristics, location, equipment, medication, and operator characteristics are reported as mean, median, and interquartile range (IQR) or as frequencies and proportions according to type of variable. Our primary outcome, the proportions of patients with first-pass success before and after the introduction of the protocol, was compared using a Pearson chi-squared test. Secondary outcomes were compared between protocol periods using Pearson χ2 tests for dichotomous or categorical variables and non-parametric Mann–Whitney tests for continuous variables. For comparisons of baseline characteristics and secondary outcomes, we present p values corrected for the false discovery rate (FDR) [24] to account for multiple comparisons.

Subgroup analyses explored whether the effect of the protocol intervention on the outcome varied across the levels of a priori defined subgroups. Subgroup analyses were performed using multivariable logistic regression with heterogeneity of treatment effect tested by including an interaction between the protocol intervention and subgrouping variable. We evaluated primary and secondary outcomes between ED, ICU and ward settings; by provider experience; anticipated anatomical difficulty; anticipated physiologic difficulty; use of paralytic and by video versus direct laryngoscopy. Provider experience was dichotomized with an ‘experienced intubator’ being defined as either an attending staff physician in critical care or emergency medicine or any third-, fourth-, or fifth-year emergency medicine, anesthesiology or critical care medicine resident.

Sample size

The primary outcome (first-pass success) was tested as a comparison of proportions. Several studies have reported first-pass success rates of 80–87% for intubations performed in the ICU and Emergency Room setting [23, 25]. Based on this literature and the high prevalence of junior trainees performing initial intubation attempts at our institution, we made a conservative estimation of 75% first-pass success in the ‘before’ group. This rate of first-pass success was used as the comparative baseline. It was estimated that N = 200 intubations would be required in the post-intervention period to detect a 10 percentage-point improvement in first-pass success with 80% power, assuming a 5% type-1 error rate.

Results

During the 14-month study period, a total of 630 patients were enrolled; 416 patients in the pre-protocol period and 214 patients in the post-protocol period (Fig. 1).

Fig. 1
figure 1

Enrollment, intervention, and analysis

Full size image

Baseline characteristics

Patients intubated pre-pandemic and patients intubated during the pandemic using the protocol differed significantly with respect to: location (ED 41% vs 51%, ICU 46% vs 44%, Ward 11% vs 4%); use of direct laryngoscopy (DL) versus video laryngoscopy (VL) (VL 54% vs 85%); intubator experience (experienced 46% vs 68%); use of paralytic (67% vs 86%), type of sedative used (Midazolam 28% vs 16%, Fentanyl 26% vs 12%, Ketamine 43% vs 69%, etomidate 14% vs 7%, Propofol 37% vs 18%); paralytic use (use pre-protocol 67% vs 86%); pre-oxygenation method (BVM 87% vs 80%, Nasal prong 36% vs 25%); and use of nasal prongs for oxygenation during intubation (36% vs 19%) (Table 1).

Table 1 Patient, Location, Equipment, Medication, and Operator Characteristics
Full size table

Primary outcome

The first-pass success rate in the pre-COVID period was found to be 73.1% (n = 304). Following the introduction of the protocol, the first-pass success rate increased to 82.2% (n = 176, p = 0.01) (Table 2).

Table 2 Primary and secondary outcomes
Full size table

Secondary outcome

There were statistically significant improvements in the secondary outcomes of Modified Cormack–Lehane view (47% Grade I vs 67%, p = 0.019), and rates of esophageal intubation (5% vs 0.5%, p = 0.017). There were no statistically significant differences in the secondary outcomes of hypoxia (10.1% vs 12.9%, p = 0.42), hypotension (12.3% vs 8.8%, p = 0.34), cannot intubate/cannot oxygenate scenarios (no events), CPR post intubation (0.5% vs. 2.3%, p = 0.11), ICU mortality (17.3% vs. 18.1%, p = 0.85), mechanical ventilation days (3 vs 3 days, p = 0.38) or ICU length of stay (5.46 vs. 7.23 days, p = 0.05) (Table 2).

Subgroup analysis

There was a significant interaction between the study time period (pre-protocol versus post-protocol period) and anatomical difficulty with respect to the risk of hypoxemia. For those with a non-anatomically difficult airway, the risk of hypoxemia increased from 8.9% in the pre-protocol period to 16.4% in the post-period (OR = 2.0, 95% CI = 1.15, 3.5, p = 0.01) while for those with anatomically difficult airways (2 or more difficult features), the risk decreased from 20.8% to 8.9% (OR = 0.37, 95% CI = 0.12, 1.15, p = 0.01). No other interactions were significant. (Table 3).

Table 3 Subgroup analysis; the interaction between the protocol intervention and subgrouping variable
Full size table

Discussion

Interpretation of findings

This single-center prospective observational study comparing intubation-related outcomes before and after the introduction of a “COVID-Protected RSI Protocol” found a statistically significant increase in first-pass success in intubations performed using the protocol without any significant increase in adverse outcomes.

Significant improvements in first-pass success observed in the post-protocol cohort were associated with several contextual changes including an increased proportion of intubations occurring in the ED, increased use of RSI, VL, and increased proportions of intubations performed by experienced operators. Other key elements of the protocol included the use of a standardized airway checklist and an in situ simulation implementation approach. While our study design precluded causal inference, these findings are in keeping with current airway management paradigms. The implementation of the protocol was temporally associated with decreases in esophageal intubations. No significant differences were found in pre-specified adverse events overall. Whether the protocol prevented or reduced transmission of COVID-19 to staff during intubation was not evaluated in our study. However, upon completion of the study, local administration in the ED and Critical Care indicated to no occupational cases of COVID-19 and were traced back to included intubations. In subgroup analyses, we observed increases in hypoxemia during intubation of anatomically easy airways for the protocol-exposed group and decreases in hypoxemia during intubation of anatomically difficult airways. We theorize that an increase in the use of paralytics and VL in the protocol may have led to better intubating conditions and quicker success in anatomically difficult airways, whereas the disincentive to manual bagging and pre-oxygenation may have led to slightly more hypoxia in anatomically easy airways, although this hypothesis was not expressly tested. These findings may also have been spurious due to multiple comparisons.

Comparison to previous studies

Subgroup analyses revealed no difference in the relationship between experience of the intubator and first-pass success, hypoxemia or hypotension in the pre- and post-COVID periods. The improvement in first-pass success noted in the post-COVID period was, therefore, not a function of the experience of the intubator. In fact, in a forthcoming publication of the original study evaluating predictors of first-pass success and adverse outcomes from the pre-COVID period, we found that intubator experience was significantly positively related to the odds of first-pass success [26]. Extant literature certainly supports greater operator experience being associated with first-pass success [7, 27]. Indeed, a recent evaluation of an attending anesthesiology-led airway management team during the COVID-19 pandemic in a Canadian center found a 91% first-pass success rate [28]. Multiple intubation attempts put patients with COVID-19 at increased risk of deterioration and the healthcare team at potential risk of COVID-19 transmission. The protocol recommended patients with COVID-19 be intubated by the most-experienced available intubator as a way of mitigating these risks. Our teaching center’s lower than expected first-pass success rates in the pre-protocol period may have been influenced by a greater comparative number of inexperienced trainees performing intubations, particularly on critical care teams.

Our finding of improved Cormack–Lehane view post-protocol is consistent with several studies conducted in the ED and ICU that concluded VL improves glottic exposure and increases first-pass success rates [33,34,35,36,33]. Nevertheless, debate exists regarding the superiority of VL to DL [34]. A recent international prospective observational cohort study examining intubation evaluated 4476 episodes of intubation from 32 countries in patients with suspected or confirmed COVID did not show any improvement in first-pass success with VL [35]. Proposed advantages of VL compared to DL include decreased proximity of operators to the patient’s airway (potentially reducing transmission of droplet-transmitted or airborne respiratory infections), increased speed of intubation, quality of view obtained, and reduced incidence of peri-intubation adverse events [34]. The use of paralytic medications during RSI is another technical element of airway management that has been reported to improve laryngeal view and first-pass success. Airway expertise as well as comfort with multiple airway techniques including surgical airway approaches is required when using paralytics due to the risk of death to patients if the provider is unable to secure an airway [40,41,38]. As such the promotion of VL use along with an RSI approach by most-experienced operators in our intervention is suspected to have contributed to our observed increase in first-pass success.

Implementation of the protocol was combined with an in situ simulation implementation strategy. The rapid adoption of changes to usual practice, without an appropriate implementation strategy, can result in increased cognitive burden for practitioners. In the early stages of the pandemic, some experts in quality improvement suggested the medical treatment of COVID-19 may not be our greatest challenge, but rather the rapid implementation of new complex protocols and procedures [39]. Experience gained through computerized simulator training can improve airway management on scenario-based respiratory arrest assessments [40, 41]. A recent systematic review further supports that in situ simulation programs can improve patient outcomes [42]. In situ simulation allows teams to test processes, create a shared mental model and mitigate human factors that contribute to error. During the COVID-19 pandemic, in situ simulation has been shown to be effective in identifying latent safety threats including inadequate preparedness for infection control, knowledge gaps regarding evidence-based practices, as well as lack of leadership and communication [43]. One recent Canadian prospective observational study reported increased rates of first-pass success when intubators had more experience intubating COVID-19 patients and theorized that adaptation through experience was the factor that likely led to this improved outcome [35]. Accordingly, the early and systematic use of in situ simulation during the implementation of the protocol may have contributed to the increase in first-pass success observed.

Strengths and limitations

The main limitation of this study is the inability to establish a causal relationship between our protocol and the identified improved outcomes due to the observational nature of the study, a changing population, evolving understanding of the management of patients with COVID-19 and a multifaceted intervention. Moreover, first-pass success is not a patient-oriented variable; intubation-related mortality or major morbidity would be more patient-oriented but were uncommon. This study is also representative of a single-center experience, limiting its generalizability. The use of this protocol by other centers, as means of decreasing risk of COVID-19 transmission and improving first-pass success, may also be limited by availability of in situ simulation and video laryngoscopy. Furthermore, intubation by the most-experienced intubator is not always be feasible as it may negatively impact trainees. Additionally, we did not specifically adjudicate HCW safety.

The prospective nature of this study allowed for detailed collection of peri-intubation clinical variables that can be missing from the medical record and better adjudication of outcomes. Standardized data collection for all consecutive cases by a neutral party (RT) may have also reduced reporting bias. To our knowledge, there are no prospective studies comparing intubation outcomes before and after the introduction of an intubation protocol instituted during the COVID-19 pandemic. Other studies have prospectively described intubation practices and protocols during the pandemic but lack the advantage of comparative data prospectively collected before implementation.

Clinical and research implications

The improvement in first-pass success after implementing a “COVID-19-Protected RSI Protocol” is a compelling finding and supports an opportunity for further research to explore whether specific components of the protocol and its implementation are responsible for observed improvements. Moreover, anticipating future respiratory pandemics, with similar modes of transmission, our protocol can serve as a foundation to inform the development and implementation of safe and effective emergent endotracheal intubation protocols for both staff and patients. Additional research will be needed to establish causality in the association between the protocol and improved first-pass success and to evaluate the impact of this protocol on HCW safety.

Conclusion

We describe the implementation of a “COVID-Protected RSI Protocol” and identify correlates of efficacy and patient safety. Following the implementation of this standardized protocol, we observed an increase in first-pass success during emergent intubations. Standardization and simulation-supported implementation of the protocol may have contributed to observed improvements in clinical results. This study not only provides useful information and lessons for healthcare teams intubating patients with transmissible respiratory infections, but also identifies modifiable factors associated with improved first-pass success.