Endotracheal tube-mounted camera-assisted intubation versus conventional intubation in intensive care: a prospective, randomised trial (VivaITN)

In comparison with patients undergoing elective surgery, critically ill patients requiring endotracheal intubation are at increased risk of life-threatening complications, mostly related to hypoxaemia during airway management [1]. During endotracheal intubation, expedited airway management reduces the rate of complications by shortening the time from the onset of anaesthesia-induced apnoea to the establishment of an airway and the beginning of artificial ventilation. However, in critically ill patients this is complicated by an increased incidence of difficult airways that is estimated at 10% [2, 3] and is approximately twice as high as in patients undergoing elective surgery [4]. Normally, endotracheal intubation is facilitated by direct laryngoscopy (DL) but this approach may fail in patients with difficult airway conditions increasing the rate of life-threatening complications from approximately 30 to 50% [5].

Because of the high incidence of airway-related complications, several methods have been proposed to improve the safety and reliability of endotracheal intubation in intensive care medicine. One of these methods is videolaryngoscopy (VL) [6]. With this technique, the larynx is visualised indirectly by a camera or a fibreoptic system. Although initial data have pointed to an increase in the first-attempt success rate [7, 8], this finding was challenged by the latest meta-analyses indicating no difference when used for unselected intubations in intensive care [9, 10].

Recently, an endotracheal tube with an integrated camera at its tip (VST) has been introduced that permits continuous visualisation of the tube’s position on a monitor connected to the camera (VivaSight™-SL, ETView Ltd., Misgav, Israel) [11]. The camera is laminated onto the anterior wall of the tube. In theory, the intubation with the advancement of the endotracheal tube through the vocal chords into the trachea should be visible on the connected monitor. We hypothesised that endotracheal intubation may be improved with this camera-aided intubation and that the tube placement could be confirmed immediately on the monitor by visualisation of tracheal cartilages. We assessed the first-attempt success rate and the number of attempts to achieve successful intubation against DL in a prospective, randomised, controlled study.

From June 2016 to January 2018, a total of 54 patients receiving endotracheal intubation were randomised to either camera-assisted intubation with the VST or conventional intubation under DL in a 1:1 ratio (see Fig. 1). Patients’ baseline characteristics are shown in Table 1. No patient presented with an anatomical condition likely to be associated with difficult intubation. Patients were intubated by attending physicians or fellows. The median professional experience was 14 years (14–15) in the VST group and 14 years (7–14) in the DL group. Intubation conditions were similar between groups (see Table 2). One patient in the VST group was excluded due to a cannot ventilate – cannot intubate situation caused by severe laryngeal oedema, and received emergency cricothyrotomy.

The first-attempt success rate was 96% in the VST group and 93% in the conventional group (not statistically significant (n.s.) with one patient in the VST group and two patients in the conventional group requiring a second intubation attempt (average number of attempts 1.0 vs. 1.1, n.s.). An overview of the outcome parameters is given in Table 3. The mean procedure duration from mouth opening to confirmed intubation was 34 s (28–39) in the VST and 35 s (28–40) in the conventional group (n.s.). Time to successful intubation on the first attempt, the rate of hypoxia, and the rate of arterial hypotension did not differ between the groups. Vomiting or aspiration did not occur in any patient. In 13 of 26 patients in the VST group, the camera image became blurred during intubation due to lens contamination by secretions, thus substantially hindering intubation by camera guidance.

In this prospective randomised trial comparing endotracheal intubation assisted by a tube-mounted camera (VST) with intubation by DL, we found no difference in the first-attempt success rate or the average number of attempts at successful intubation. Concerning other tested variables, no difference was detected between the two groups.

Thus far, the VST has shown promising results for endotracheal intubation via supraglottic airway devices [11, 17], in mannikins [18‐22] and in a cadaver study [23]. Furthermore, the guidance of percutaneous dilatational tracheostomy by VST has been evaluated and shown to be feasible [24, 25].

In critically ill patients, airway management is a hazardous procedure due to the high rate of patients with respiratory insufficiency and pre-existing hypoxaemia. Accordingly, the most prevalent complications during endotracheal intubation in the intensive care unit are hypoxaemic complications in approximately 30% of patients receiving endotracheal intubation, with 2% of patients even requiring cardio-pulmonary resuscitation [26, 27]. Because multiple intubation attempts are independently associated with complications [28], it is crucial to increase the first-attempt success rate to prevent complications.

The first-attempt success rate of intubation depends on physicians’ experience and increases with physicians’ training [29‐31]. In our study, patients were intubated by attending physicians or fellows with several years of experience in airway management. Further, in our study, the first-attempt success rate was above 90% in both groups, which is congruent to data showing the first-attempt success rate for experienced physicians to be approximately 90% [30, 32]. However, many intubations in intensive care are performed by residents with less experience than the physicians in our study [30]. Therefore, our data may not be representative. Because mannikin studies with inexperienced operators have shown a benefit for the VST over DL with a higher success rate [18, 19], we cannot exclude the possibility that the VST performs better in the hands of inexperienced operators. However, this might not be a proof of VST being superior but a sign of lower quality DL in the hands of inexperienced operators. Nevertheless, in this case, the VST might be an alternative to conventional DL in resource-limited settings with the unavailability of a physician experienced in airway management; however, in our opinion this approach should only be chosen as the last option.

Videolaryngoscopy (VL) has been widely used for intubation in intensive care. It had been hypothesised that the first-attempt success rate could be increased over DL, but the latest data have shown no benefit when used in an unselected cohort of intensive care patients [9] and even inferiority when used pre-clinically [10]. When analysing patients in whom there was an expectation of difficulty in achieving intubation due to the condition of the airways, VL led to fewer failed intubations than DL. A potential benefit of VST in the hands of experienced operators managing anatomically difficult airways may be hypothesised. However, a direct comparison with alternative instruments such as VL is lacking. Although not an exclusion criterion, no patient in our study was expected to be difficult to intubate due to the anatomy of the airway and in analogy with VL, we did not detect any difference in the first-attempt success rate.

The procedure duration was identical in both groups. An intubation duration of approximately 30 s has been recorded for experienced anaesthesiologists using the VST and DL [8, 33, 34], although one study identified a difference of 5 s in favour of the VST [33]. On the other hand, the use of VL could prolong the time to intubation [8, 34]. Regarding our data and previously published data [33], there is no indication that intubation with the VST prolongs the duration of intubation. This may be beneficial in patients who require expedited airway management due to respiratory insufficiency with a diminished oxygen reserve and apnoea tolerance. Moreover, the usefulness of the device may increase with the passage of an individual learning curve. This is a factor that makes comparison of newly established methods with routinely applied standard methods difficult.

In our study, camera-guided intubation was substantially hindered in every second patient due to a contamination of the lens by secretions. The VST is equipped with a rinsing channel to clear the lens from secretions but in all cases, endotracheal intubation was completed by DL in favour of expedited establishment of the airway. The problem of a soiled airway is common in intensive care patients as opposed to mannikins and cadavers in studies in which the VST was evaluated for endotracheal intubation. For one VL model, severe lens contamination (defined as impossibility to visualise the larynx) occurred in 3.2% [35], while the same group of authors found an incidence of severe lens contamination for the same VL model in 0.5% of clean airways and in 1.3% of soiled airways [36]. In the latter study, the first-attempt success rate by VL dropped from approximately 90% in clean airways to approximately 80% in soiled airways compared with 75% to 65% by DL. As opposed to most videolaryngoscopes where the camera is recessed and therefore partially protected against secretions, the camera at the tip of the VST seems more vulnerable to lens contamination, decreasing its usefulness in the guidance of endotracheal intubation.

Our study has certain limitations. Not being a blinded study, the types of intervention compared were at risk of ascertainment bias, though quick and atraumatic intubation was the set goal in both groups. The first-attempt success rate was higher than anticipated and therefore the sample size may not have been sufficient to detect a difference between the groups. Due to lens contamination, half of the patients randomised to the VST group basically received conventional intubation, highlighting the decreased usefulness of VST in patients with a high incidence of soiled airways. Therefore, our sample size is possibly not large enough to detect significant differences between the two groups. Because of the requirement to obtain informed consent, only patients receiving elective or urgent intubation could be included. Therefore, no data are available on the performance for emergency airway management. None of the patients who underwent randomisation had a difficult airway and it would be of interest to establish whether the VST would be beneficial under these circumstances.

In patients requiring elective or urgent endotracheal intubation in the intensive care unit, no additional benefit was demonstrated for the VST over DL. The performance of the VST in patients with an expected difficult airway should be evaluated separately. Airways soiled with secretions may reduce the usability of the VST due to lens contamination. Because of a high first-attempt success rate in both groups, we suggest that the VST should be evaluated in a larger cohort of patients and should be evaluated when applied by physicians with less expertise in intubation.