Comparison of adverse events between video and direct laryngoscopes for tracheal intubations in emergency department and ICU patients–a systematic review and meta-analysis

Tracheal intubation is a primary lifesaving procedure for emergency department (ED) or intensive care unit (ICU) patients associated with respiratory dysfunction or decreased airway protection. Under the urgent situation, however, airway management can be challenging due to decompensated cardiopulmonary physiology, inadequate provision of tools and skilled staffs, unfasted state, simultaneous performance of cardiopulmonary resuscitation (CPR) or other medical procedures, difficult access to the patient head, and various anatomic features of difficult airways [1]. It has been reported that failure rate of urgent tracheal intubations in the ED and ICU is significantly higher compared to tracheal intubations performed in the operating room [2]. Furthermore, the results of the Fourth National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society shows that at least one in four major adverse airway events in a hospital are likely to occur in the ED or ICU and adverse airway events leading to death or brain damage are approximately 30-fold and 60-fold more frequent in the ICU and ED than in the operating room [1]. Thus, the strategies or alternative methods to reduce the occurrence of adverse events associated with urgent tracheal intubations in the ED and ICU patients are needed.

Traditionally, tracheal intubation is performed using direct laryngoscope (DL), which requires the alignment of the oral, pharyngeal and laryngeal axes to allow direct visualization of the glottis. Thus, tracheal intubation with DL is generally regarded as a difficult skill to acquire and maintain. Video laryngoscope (VL) is a new device that contains a miniaturized camera at the blade tip to indirectly visualize the glottis. Because of proven advantages of a fast learning curve, an improved laryngeal visualization and an increased success rate, VL has been widely used for tracheal intubation in emergency and critical situations [3‐22]. There have been several meta-analyses comparing VL with DL for tracheal intubation in the ED and ICU patients, but they have mainly focused on the success rate of tracheal intubation and provided different outcomes [23‐29]. Most important, it is unclear whether the use of VL can reduce the adverse events associated with tracheal intubations in the ED and ICU patients. Thus, this systematic review and meta-analysis was carried out to determine whether VL compared with DL could reduce the adverse events associated tracheal intubation in the ED and ICU patients.

This systematic review and meta-analysis was conducted following the recommendations of The Cochrane Handbook for Systematic Reviews of Interventions and reported according to the PRISMA statement [30] (www.​prisma-statement.​org). The protocol was registered on the PROSPERO (http://​www.​crd. york.​ac.​uk/PROSPERO, ID: CRD42018100562).

This systematic review and meta-analysis is focused on adverse events of tracheal intubation using VL compared with DL in the ED and ICU patients by including both RCTs and observational studies. Pooled analysis showed that the use of VL reduced the rate of EI; as to the other adverse outcomes, however, no significant differences were identified.

The rate of EI was chosen as the primary outcome in this analysis, as even a single episode of recognized EI is significantly associated with desaturation, increased risk of aspiration and cardiac arrest. It has been shown that patients with EI have higher incidences of aspiration (6.1 times), dysrhythmia (6.4 times), hypotension (3.1 times), and hypoxemia [3]. In our analysis, however, a significant lower rate of EI by using VL did not result in any significant difference for other adverse events. This is probably due to small sample size, leading these results inconclusive, as the TSA has proved. Even though, a lower trend for the incidences of hypoxemia and aspiration with VL can still be identified. Inexperienced operators benefited more from using VL, with significant lower incidences of EI and hypoxemia. For inexperienced operators who have not performed the tracheal intubation with VL and DL, visualization of the airway on VL screen can allow their supervisors to directly assist them in completing tracheal intubation themselves, thus improving the success rate [42]. In contrast, experienced operators with extensive training and experience on the tracheal intubation using DL might overshadow the benefits of VL, especially when patients have a normal airway or there are some VL-related difficult scenarios like secretions or blood in the airway or certain VL design-related deficiencies which may bate their benefits [19, 22, 43, 44].

Different from above non-fatal adverse outcomes, an increased trend was otherwise shown for incidence of new-onset cardiac arrest and short-term mortality by using VL. However, it should be noted that among the studies reporting the results of new-onset cardiac arrest and short-term mortality, only one was the study in which  emergency tracheal  intubation was performed by experienced operators [14]. The longer duration of intubation with VL by inexperienced operators may result in an increased risk of severe life-threatening adverse outcomes [19]. In addition, the pooled result of the short-term mortality was mainly from a CPR study [8]. Although the participants were randomized, the baseline clinical characteristics for both groups was not comparable, with more ischemic heart diseases in the VL group, leading the patients in this group to be at a higher risk of severe adverse outcomes. It has been reported that the use of a VL can reduce the chest compression interruptions for both experienced and inexperienced operators [11, 18]. However, visualization of VL can be unfavorably compromised by increased amount of secretions and emesis in the upper airway, which are common during the CPR, leading to a longer duration of intubation which may worsen the patient’s prognosis [16, 22]. Besides traditional airway suctioning, other techniques to decrease influences of massive secretions and emesis on airway visualization and intubation procedure, such as intentional esophageal intubation (IEI) [45] and suction-assisted laryngoscopy and airway decontamination (SALAD) [46], have been described. Furthermore, the use of airway decontamination technique provides improved intubation conditions with a VL. However, these methods appear mostly in case report or simulated mannequin study and have not been generalized in clinical practice. Nowadays, even for skilled operators like trained anesthesiologists, DL might still be the first choice for urgent tracheal intubation during the CPR, especially in patients without difficult airways.

Our study included RCTs and observational studies. The overall risk assessment of bias for the included RCTs was classified as low risk. Although blinding was not adopted in the most RCTs, we judged “no blinding” as low risk, as it seems impossible to blind personnel in the urgent situations at times. The overall risk assessment of bias for the included observational studies was classified as moderate or severe risk, mainly due to the confounding factors. For most outcomes, the level of heterogeneity was low, but subgroup and sensitivity analyses based on some potential clinical heterogeneous factors had been performed in our analysis. Furthermore, the TSA ensured the credibility of the result of primary endpoint from RCTs.

There are some limitations in our analysis that deserve special attentions. First and foremost, the inclusion of observational studies inevitably introduces selection bias, leading the possibility that patients in the VL and DL groups differ significantly in terms of operator’s expertise, anesthesia methods, certain difficult airway characteristics, clinical scenarios, and even basic characteristics of patients [4‐6, 8‐10]. Besides the disparities between groups, a reporting bias might be also present due to self-reported or recall property of the data in the observational studies, leading to the inaccuracy of data collection. Second, some of the included studies were quality-improvement process within a before–after study [4, 9]. The better result may probably not only from the switch from DL to VL, but also from the improved patient management in quality-improvement studies. Third, during the searching process, 12 observational studies were found to be carried out in a same ED or ICU from a same hospital with an overlapped study period [3, 5, 10, 47‐55]. Only the latest 3 large studies [3, 5, 10] covering the majority of other 8 studies were decided to be included. This might lose some validated participants while avoiding repeated enrolment. Fourth, definition of the expertise used in our analysis was somewhat arbitrary. A Cochrane review defines an experienced operator as a clinician with more than 20 tracheal intubations with each device and thence obtains a fewer failed intubations when using VL [56]. However, it is suggested that the number of procedures required to achieve proficiency with tracheal intubation using DL in an controlled environment like OR is approximately 50 [57, 58]. Given the challenges of limited cardiopulmonary reserve, difficulty on airway assessment, few chances of tracheal intubation practice in the ED or ICU, it is even hard to know how many procedures are required to achieve a definitive competence with tracheal intubation in this patient population. Anyhow, having one VL available for all urgent tracheal intubation could potentially offer a major safety advantage. Fifth, there might be still some other heterogeneous factors in our analysis, such as different types of VLs used and different in-hospital setting, i.e. ED, ICU, or general ward. About one third of our included studies used more than one type of VL [3‐5, 8, 10, 17, 19], making subgroup analysis on this factor difficult. Although ED and ICU are two main in-hospital settings for urgent tracheal intubation, indications, intubation conditions, and severity of patients’ illness would not differ significantly, at least not as much as the difference between pre-hospital and in-hospital setting. Lastly, some studies used the episode of intubations or attempts instead of participants [3, 4, 12]. When calculating incidence of adverse outcomes, this would be inappropriate. However, sensitivity analysis by excluding these studies did not change the result of primary endpoint.

This systematic review and meta-analysis reveals that VL can reduce the risk of EI during urgent tracheal intubations in the ED and ICU patients, but does not provide significant benefits on other adverse events associated with tracheal intubation. Further studies are needed to demonstrate whether severe adverse events like cardiac arrest or mortality are significantly different between two devices. Furthermore, well-designed RCTs are further needed to focus on a specific scenario and should stratify some other prognostic indicators such as length of hospital stay and cost.