In the year under study, 48 intubations were performed, all of which were included in the analysis. The median patient age was 11 months (interquartile range, 3–62.7) and the median weight 12.5 kg (interquartile range, 5.5–20). Male patients amounted to 70.8% of the sample. In addition, 41.7% (20 patients) had heart disease, which was cyanotic heart disease in 6 cases. The most frequent reasons for admission were respiratory disease and neurologic disease (Fig. 1). Fifty-four percent of intubations were performed in patients not previously intubated, 25% were reintubations and 15% were endotracheal tube (ETT) exchanges. The most frequent indication for TI was respiratory failure (Fig. 2). The ETT was exchanged due to a problem with the size of the tube in one patient, to replace an uncuffed ETT for a cuffed one in another patient, and due to ETT obstruction in 3 patients.

Figure 1.

Reason for admission.

(0.09MB).

Figure 2.

Indication of intubation.

(0.12MB).

Premedication with atropine was administered in 21 intubations (43.8%). The induction agent used most frequently was etomidate (23 intubations; 48%), followed by midazolam (12 intubations; 25%), ketamine (6 intubations; 12.5%), propofol (2 intubations; 4.2%) and thiopental (1 intubation; 2.1%). The most frequently used analgesic was fentanyl (42 intubations; 87.5%) and the most frequently used muscle relaxant was rocuronium (43 intubations; 90%).

Most intubations were performed by intensive care specialists on staff (35 intubations; 73%), followed by intensive care medical residents (17%), paediatrics medical residents (6%) and anaesthesiologists (2%). The physician that carried out the second attempt was the same one that had carried out the first in 14 cases (74%), a different one in 16%, and an intensive care medical resident in all other cases (10%). The second attempt was successful in 68% of cases, and the reasons for failure were the same as in instances of first-pass failure.

We identified the following risk factors for difficult airway: aged less than 2 years (27 cases; 56%),4 previous history of difficult airway (9 cases; 19%), airway obstruction (7 cases; 15%), facial hypoplasia/retrognathia (6 cases; 12%), Cormack-Lehane grade ii-iv on examination (4 cases; 8%), limitation of neck motion (4 cases; 8%), malformative syndrome (3 cases; 6%) and restricted mouth opening (2 cases; 4%).

Bag-mask ventilation was administered before intubation in 30 of TIs (62.5%), with ventilation difficulties emerging in 5. Apnoeic oxygenation was used during TI in 11 cases. Every patient was intubated through the mouth with a cuffed ETT. Direct laryngoscopy was the method used most frequently in first intubation attempts (27 cases; 56%) and successive attempts, followed by video laryngoscopy with disposable and nondisposable devices. The laryngoscopic view during TI was classified as Cormack-Lehane grade I in 35 cases (73.8%), grade ii in 19% and grade iii in 7.2%. The BURP manoeuvre (back, upward, right and lateral pressure) was used in 42% of first attempts and 47% of second attempts.

Intubation was achieved in 100% of cases, in the first attempt in 60.4% (29 TIs), in the second attempt in 27% (13 TIs) and in the third attempt in 10% (5 TIs). One patient required 5 attempts. We did not find differences in the first-pass success rate between staff physicians and medical residents (Table 1). The most frequent reason for failed TI was difficulty inserting the ETT in the trachea (53%), followed by not visualizing the glottis (37%).

The multivariate analysis did not evince an association between first-pass failure and any of the patient characteristics or indications for intubation, the experience of the physician that performed intubation, health care-related variables or the type of laryngoscopy used, except for difficult bag-mask ventilation (Table 1). In the multivariate analysis adjusted for age and difficult airway, difficult bag-mask ventilation continued to be an independent risk factor for failed intubation (odds ratio, 2.33; 95% confidence interval, 1.097–4.954; P = .028).

A total of 6 adverse events were detected in association with intubation (12.5% of TIs). This included the identification of 4 serious TIAEs (8.3%): hypotension requiring intervention (2), cardiac arrest with return of spontaneous circulation (1) and delayed recognition of oesophageal intubation (1) (Table 2). None of the patients died during or as a consequence of the procedure.

Desaturations during intubation attempts were mostly mild. There was a desaturation of less than 10% relative to the baseline oxygen saturation (SpO2) in 62.2% of first intubation attempts and of 10% or greater in 37.8%, greater than 20% in 21.6% of cases (Fig. 3). We did not find an association between desaturation and the number of intubation attempts (P = .569), the presence of heart disease (P = .769) or the type of physician performing the procedure (P = .572).

Figure 3.

Decrease in transcutaneous saturation during intubation. Percentage of intubations corresponding to each interval of maximum saturation decrease relative to baseline during intubation.

(0.09MB).

In 10% of events (n = 5), there were problems related to supplies during TI, such as specific supplies not being stocked in the airway cart. In one case, there were issues with vital signs monitoring, and in another, lack of team organization and role assignment. No events were documented in association with the readiness or use of intubation drugs.

The univariate analysis revealed that both difficult bag-mask ventilation and multiple intubation attempts were associated with the development of TIAEs (Table 3) and serious TIAEs, although these factors were no longer significant in the multivariate analysis. On the other hand, role assignment was associated with a lower incidence of TIAEs (P = .039); and the use of the airway cart (P = .031), role assignment (P = .016) and the review of the patient’s history and risk factors (P = .037) were associated with a lower incidence of serious TIAEs, an association that persisted in the multivariate analysis adjusted for age and the presence of difficult airway (Table 4).

In most cases, intubation was performed according to the RSI protocol. In more than 90% of instances, the practitioner checked the intubation drugs and the vital signs, the patient received preoxygenation (most through non-invasive ventilation) and roles were assigned to each member of the team. We found the lowest adherence to the protocol for the items concerning the use of the checklist to verify supply and equipment stocks and the assessment of anatomical signs suggestive of difficult airway (Table 5).

In our study, the first-pass success rate (60%) was similar to the rates reported in other studies conducted in the PICU setting, including studies of large registries such as the National Emergency Airway Registry for Children (NEAR4KIDS).1,5,7 We did not identify any patient characteristic or indication for intubation associated with intubation failure, which was consistent with some previous studies.5 However, studies of larger scale, like a recent report of the NEAR4KIDS registry,7 have found that younger age, a previous history of difficult airway or signs of upper airway obstruction were associated with an increased number of attempts.

The experience of the physician performing the procedure has been identified as a relevant factor that contributes to the first-pass success rate.9–13 In our study, the percentage of first attempts of TI performed by attending physicians was higher compared to other studies, in which medical residents in intensive care are usually the providers responsible for the first attempt.1,5 Since Spain does not have an official curriculum for the paediatric intensive care speciality, medical residents in Spain spend less time training in this subspeciality of paediatrics compared to residents in other countries, which may partly explain our results. However, other studies have also described a decrease in the number of procedures performed by residents in recent years,14 and the development of non-invasive ventilation techniques has also decreased the frequency with which patients require intubation,15 which in turn is reducing frequency with which PICU providers practice intubation in the clinical setting. In addition, given that urgent intubation is frequent in the PICU, the most experienced physician may be tasked with carrying out the procedure to guarantee success and minimise risks to the patient. Paradoxically, in our study, contrary to the findings of the NEAR4KIDS registry,7 we did not find differences in the first-pass success rate based on the experience of the physician performing the procedure, which could be due to selection bias, as attending physicians may have taken charge of intubating unstable patients or patients with suspected difficult airway from the outset, restricting residents to attempting intubation in patients in whom it was less likely to fail.

Intensive care patients admitted to intensive care units often require multiple attempts to achieve intubation.16,17 The incidence of difficult intubation in our study (defined as the need of 3 or more attempts) was slightly higher (12.4%) compared to other studies (7% in the NEAR4KIDS report).5,18

Several factors, such as lack of preparation of the patient or the equipment, deviation from protocols and teamwork problems (poor team composition or communication) can also play a role in failed intubation.17,19,20 The use of cognitive aids such as RSI and difficult airway protocols, TI checklists19,21 and training with simulation of TI scenarios, as was done in our unit, have been associated with improved intubation outcomes,22–24 although we were unable to analyse the effect of these measures in our study, as we had not collected data prior to their introduction.

Some studies have found that the use of video laryngoscopy achieves a greater intubation success rate in intensive care units,3,17,25 although others have found contradicting results.26–28 At present, the conventional laryngoscope continues to be the preferred device,29 while the video laryngoscope is considered the first-line device in the case of difficult airway or failed intubation with direct laryngoscopy.3,30 In our study, we did not find a difference in the success rate with the use of conventional versus video laryngoscopy, but the latter has only been introduced recently, and as is the case in every novel technique, training is crucial for its implementation and success, so that structured training programmes need to be developed, including practice on manikins in initial trainings and simulation of clinical scenarios.31

Adverse events of TI are frequent, occurring in up to 20% of critically ill paediatric patients,1, and are associated with poorer patient outcomes, longer duration of mechanical ventilation and an increased mortality.32 In our study, we found an incidence of adverse events of 12.5% with an incidence of serious TIAEs of 8.3%, figures that were similar to those observed in other studies.1,32 Several studies have found a higher frequency of TIAEs in mixed intensive care units that provide care to cardiac patients, such as our own, since cardiac patients are at increased risk of adverse events.33–37 In our study, although a high percentage of intubations corresponded to children with heart disease (41.7%), we did not find a greater frequency of TIAEs in this subgroup of patients.

Cardiac arrest can occur in 7% of intubation events and death in 1.6% of intubation events in the cardiac PICU setting.38 In our study, there was one case of cardiac arrest (2%) and none of the patients died from complications of intubation.

Desaturation events are the main complication and most frequent cause of discontinuing intubation.39 Nishisaki et al.1 and Lee et al.7 have reported a frequency of desaturation during intubation (defined as SpO2 < 80%) of 13.5% and 27%, respectively, similar to the frequency found in our study. Desaturation is more frequent and severe in children compared to adults, as children have a lesser functional residual capacity and increased metabolic oxygen demands.6 Desaturation increases with the number of attempts7 and is associated with haemodynamic adverse events (cardiac arrest, hypotension or hypertension requiring intervention and arrhythmia).36,40 Preoxygenation and apnoeic oxygenation strategies may prolong the time to desaturation and facilitate successful intubation in the first attempt.3,17,41,42 However, the current evidence is contradictory as regards the best oxygenation approach.43,44

In our study, neither any of the patient-related factors nor the experience of the physician were associated with TIAEs. In opposition, the NEAR4KIDS registry has identified several factors associated with TIAEs, such as difficult airway and haemodynamic instability.1,7 Some studies have found that intubation performed by medical residents may be associated with an increased incidence of adverse events.1 Branca et al. found a significant increase in the first-pass success rate and decrease in TIAEs with increasing experience in paediatric critical care medicine fellows throughout their training period, which indicates that there is a progressive and significant acquisition of skills.9

As regards the TI procedure, there was a history of difficult bag-mask ventilation in 10% of cases, which was associated with the development of TIAEs in the univariate analysis, in agreement with the findings of the NEAR4KIDS registry.37 The performance of multiple intubation attempts has also been identified as an important factor associated with adverse events.7 Lee et al. found that the incidence of TIAEs increased from 10% in the first attempt to 29% in the second and 38% in the third and subsequent attempts, while the incidence of serious TIAEs also increased from 5% to 8% and 9%, respectively.

The use of cognitive aids and the application of protocols and the airway verification checklist have been associated with a decreased incidence of complications of TI.1,3,8,22,45 In our study, we found considerable adherence to the RSI protocol, and adherence was independently associated with a lower frequency of TIAEs, which supports the use of these aids to improve intubation outcomes in PICUs.

There are several limitations to our study. The data were verified rigorously by the researchers, but were initially self-reported by the physicians that performed the intubations, so there is a risk of recall bias or Hawthorne effect (participants in a study may change their behaviour because they are aware of being observed). On the other hand, our study was conducted in a single centre, in a small sample and in a PICU with a high percentage of patients with complex congenital heart disease, which may limit the generalization of our findings to PICUs with different characteristics. The multivariate analysis may have been affected by the small sample size. Notwithstanding, our study may serve as a model for the implementation of a quality improvement strategy and the analysis of intubation outcomes in the PICU setting.