Coronavirus disease 2019 (COVID-19) is characterized mainly by moderate/severe pneumonia due to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)1 in 42% of patients, and among them 61–81% of patients require intensive care unit (ICU) admission.2 COVID-19 pneumonia, and related hypoxemia and desaturation patterns, are atypical and not comparable to that of the “standard” acute respiratory distress syndrome (ARDS) patients, as these patients show a low arterial partial oxygen pressure / fraction of inspired oxygen (PaO2/FiO2) ratio and initially higher compliance and pulmonary shunt fraction.3

The Emergency Endotracheal intubation (ETT) of SARS Cov2 patient is a high-risk procedure and represent an additional challenge to all intensivists due to barrier enclosures.4 Silent hypoxia often leads to late intubation which is linked to a high risk of complications. Non-survivors showed worse blood gas analyzes than survivors, both before and after intubation.5 Few studies reported the implications and adverse events of performing endotracheal intubation for critically ill COVID-19 patients admitted to ICUs.6 The choice of supplementary oxygen delivery interface is challenging and the decision to provide invasive ventilatory support is crucial.

The application of non-invasive support, with e.g. as high-flow nasal oxygen (HFNO), continuous positive airway pressure (CPAP) and non-invasive ventilation (NIV), might correct hypoxemia and could avoid endotracheal intubation.7 The Italian outbreak experience resulted in a series of clinical protocols, suggesting the use of a short NIV trials with the need of intubation with a PaO2/FiO2 below 150.8 In this setting, it is advocated to prefer anticipated endotracheal intubation in order to minimize related complications.8 In the setting of critically ill-patients’ intubation, hypoxia, hemodynamic failure and cardiac arrest are common and well described events and given the COVID-19 physio-pathological evidence, these might be much more common in this context due to limited safe apnea time and poor oxygenation reserve.9 The aim of the present study was to determine the incidence of major adverse events related to tracheal intubation in COVID-19 patients defined as the onset of severe hypoxemia, hemodynamic instability and cardiac arrest within 30 minutes from the procedure

This was a prospective, observational, dual-center study. Data were collected prospectively on patients admitted into two ICUs.

The study was designed in accordance with the Declaration of Helsinki. The ethics committees waived the need for written consent, allowing constant verbal consent (N° 125/20 of December 15, 2020). This was attained from all participants after explaining purposes, significance and advantages of the research by the trained data collectors and through a covering letter provided to each respondent. Patients were also informed that they had the right to withdraw from the study at any time. For each enrolled patient, their personal information was recorded and kept in a secure folder. To protect their privacy, patients’ information were only accessible to the researchers. Reporting was adapted from the STROBE checklist for observational research.10 The study was registered (ClinicalTrial.Gov Trials Register NCT04909476) on June 01, 2021.

All patients admitted for respiratory failure that needed advanced airways management were eligible for the study. Inclusion criteria were: 18 years of age and older; a positive SARS-CoV-2 polymerase chain reaction (PCR) test; intubation for respiratory failure or muscle exhaustion. Exclusion criteria were:

• -

  a negative SARS-CoV-2 PCR test;

• -

  intubated outside the hospital, thus secondary transfer;

• -

  intubated because of a cardiac arrest;

• -

  undergoing general anaesthesia for surgery;

• -

  incomplete data for enrollment.

All intubations were performed by attending physiscian specialized in anesthesia and intensive care or by residents under supervision.All anesthesia and intensive care residents had at least 4 yr of intraoperative anesthesiology experience.

Demographic variables and data about the procedure were collected prospectively for all patients at intubation using a Case Report Form (CRF). Hemodynamic parameters and temperature were measured continuously through invasive monitoring. Pre-procedural collected data included date and time of the intubation, the level of monitoring, patient’s parameters, ongoing vasoactive drugs and oxygen delivery method. Before the procedure patients were evaluated for anticipated difficult airways, according to SIAARTI task force Recommendation, and based on carefully assessment of the need for tracheal intubation, blood gas analysis and radiological findings.11 Healthcare professional performance data included the number of attempts at intubation, the expertise of the different attempts, the number of laryngoscopies per week usually performed, the hours on shift already worked before intubation. We recorded pre-oxygenation data focusing on type and tyming of the device used and the peripheral oxygen saturation (SpO2) at the end. Drugs and dosage, the device used, grading of Cormack-Lehane views, the lowest SpO2 during laryngoscopy were also recorded. End-tidal carbon dioxide (EtCO2) level was also recorded when available. We recorded the occurrence of difficult airways (more than 3 attempts of intubation), the need of front of neck airways, the onset of any arrhythmias, the detection of esophageal intubation. Within 30 minutes from the intubation, we recorded the occurrence of severe hypoxia, defined as a SpO2 <80%, hemodynamic instability defined as SBP 65 mmHg recoded at least once or SBP < 90 mmHg for 30 minutes, a new requirement or increase of vasopressors, fluid bolus >15 mL/kg to maintain the target blood pressure, and the occurrence of cardiac arrest. Within 6 hours from the intubation the onset of new pneumomediastinum and/or pneumothorax was collected.

The primary endpoint was to determine the occurrence of at least 1 of the following events within 30 minutes from the start of the intubation procedure and to describe the types of major adverse peri-intubation events12: severe hypoxemia defined as an oxygen saturation as measured by pulse-oximetry (SpO2) <80%; hemodynamic instability defined as a SBP 65 mmHg recoded at least once or SBP < 90 mmHg for 30 minutes, a new requirement or increase of vasopressors, fluid bolus >15 mL/kg to maintain the target blood pressure; cardiac arrest. The secondary endpoint included the occurrence of cardiac arrhythmias, difficult intubation defined according to SIAARTI task force Recommendation, a cannot-intubate or cannot-oxygenate scenario (defined as impossibility to achieve a successful tracheal intubation and adequate patient’s oxygenation), emergency front-of neck airway, pulmonary aspiration of gastric contents, esophageal intubation, airway injury and the occurrence of pneumothorax or pneumomediastinum within 6 hours from the intubation.

Statistical analysis was performed using Stata software v. 13.1 (StataCorp LP, College Station, TX, USA). The cumulative incidence of major adverse events during intubation is estimated to be 0.28.13 The primary endpoint of our study was to estimate the cumulative incidence of a major adverse event in patients with COVID-19, a figure currently not estimated in the literature. We assumed a larger cumulative incidence of the major adverse event in the COVID-19 patients who need intubation. We considered that the major adverse event can occur in 0.45 of COVID-19 patients who need intubation; the sample size of 80 achieves 80% power to reject the null hypothesis of zero effect size when the population effect size is 0.30 and the significance level (alpha) is 0.050 using a two-sided one-sample z test. Supplement online 1 shows the dropout-Inflated Sample Size and the estimated power for a one-sample proportion test. The primary and secondary endpoints were expressed as absolute frequency and percentage (number of each event on total intubations). Normal distribution was assessed by Shapiro-Wilk test. Continuous variables were expressed as mean ± standard deviation or median [interquartile range], as adequate. Categorical variables were shown as absolute and percentage frequencies. Two groups were identified: patients who had at least one adverse event at intubation and patients without adverse events. The descriptive comparison between the two groups was performed by Student's T test or Wilcoxon's Ranksum test for continuous variables and by the Chi-squared/Fisher’s exact test for categorical ones. A logistic regression model with stepwise forward modality was carried out to define the risk factors for the major adverse event. A multivariable logistic regression model was performed including as covariates variables with P values < 0.05 in the bivariable analyses or with a clinically relevant meaning. A post-estimation analysis was performed to evaluate the goodness of the model with the Hosmer-Lemeshow tests, Akaike's information criterion, Bayesian information criterion, accuracy, sensitivity and specificity of the model.

Between November 2020 and May 2021, we enrolled 142 COVID-19 patients needing endotracheal intubation. Fig. 1 shows the flow study population. The type of our data and the electronic health record let us to collect all data without missing. However, we exceed the estimated sample size, improving the power of the study. Demographic data are represented in Table 1.

Figure 1.

Flow study population.

(0.14MB).

Airways Assessment and Indication of Intubation are shown in Table 2. Particularly, thyroimental distance was calculated as a straight line from mentum to thyroid notch with mouth closed and neck fully extended.

Table 3 presents intubation characteristics and performance. Midazolam was the most frequently used induction agent (110 patients (77.46%), 48 (43.63%) times in co-administration with propofol and 50 (45.45%) times in co-administration with ketamine. Rocuronium was the unique neuromuscular blocking administered with a median dose 1.11 mg/kg (0.94-1.25). The average doses for midazolam, propofol e ketamine were 0.06 (0.04-0.08) mg/kg, 1 (0.6 – 1.41) mg/kg and 1.11 (0.94 -1.33) mg/kg, respectively. Fentanyl was the only opioid used. All drugs were administered on actual body weight. Characteristics concerning methods used to confirm intubation and attemps are reported in Table 3. Particularly, waveform capnografy was used in the 21.1 %, while monitor capnometry was ised in 27.5 of patients. Colorimetric carbon dioxide detection was never used. Macintosh Laryngoscope was used as the primary device for tracheal intubation for 21(14.89%) patients and a video laryngoscope (GlideScope®). in 121(85.21%) patients. No patients experienced impossibility to achieve a successful tracheal intubation and adequate patient’s oxygenation. First-pass intubation success was achieved for 129 (90.84%) patients.

Adverse events are shown in Table 4. Particulalry, among the 142 enrolled patients, 105 (73.94%) experienced at least 1 major adverse peri-intubation event,and 35 (24.65%) more than one adverse events simultaneously.

Cardiovascular instability accounted for most of the events, followed by severe hypoxemia. Four patients (2.82 %) had a cardiac arrest following tracheal intubation. Particularly, among the secondary outcomes, 13 (9.15%) patients had new onset of supraventricular arrhythmia and at six hours from intubation, 16 (11.26%) patients developed pneumothorax/pneumomediastinum.

All pneumothorax/pneumomediastinum were successfully managed with chest tube drainage.

The rate of major adverse events was significantly lower with first-pass intubation success than it was for patients requiring 2 attempts (67.61% vs 84.62%; p < 0.001).

Multivariable analysis is shown in Table 5. The model correctly classified the occurrence of a major adverse events with a accuracy of 82%. Overall, 97 (31.7%) of patients discharged from ICU survived, while 67 (47.18%) died in hospital. No difference was found for ICU length of stay between patients with a major adverse peri-intubation event and patients without events: 21 (14.5-32) days vs 20 (14-32) days; (p = 0.44).