Predictors of difficult videolaryngoscopy with GlideScope^® or C-MAC^® with D-blade: secondary analysis from a large comparative videolaryngoscopy trial Free

Abstract

Videolaryngoscopy is widely used in the management of patients with presumed difficult airway. It offers an improved laryngeal view compared with direct laryngoscopy and increases the likelihood of successful intubation in patients for whom direct laryngoscopy is anticipated to be difficult.^1,2 For these reasons, videolaryngoscopy is now established as a rescue technique for failed direct laryngoscopy.^3–5

Although videolaryngoscope design differs significantly between devices, acute-angle blades are specifically intended for better visibility of the relatively anterior airway or when cervical motion is limited.^6,7 Such devices include, but are not limited to, the GlideScope^® (Verthon, Bothell, WA, USA), C-MAC^® with D-blade (Karl Storz, Tuttlingen, Germany), and McGrath^® Series 5 (Aircraft Medical, Edinburgh, UK). When using such devices, viewing of the glottic opening relies on video images captured by the blade's distally positioned camera. Placement of the tracheal tube is thus a video-guided procedure, which is fundamentally different from direct laryngoscopy. It is therefore conceivable that predictors of difficult direct laryngoscopy may not predict difficult acute-angle videolaryngoscopy.

Published predictors of difficult direct laryngoscopy include limited cervical motion, small mouth opening, short thyromental distance, and high Mallampati score.⁸ Although acute-angle videolaryngoscopes are desiged to overcome some of these limitations, their use may introduce other challenges, such as difficult blade insertion or difficult tube passage despite adequate videoscopic laryngeal view. Not surprisingly, although a highly successful airway management strategy, tracheal intubation using videolaryngoscopy is associated with its own set of challenges.^9,10 However, despite widespread use, the specific predictors of difficult acute-angle videolaryngoscopy remain unknown.

We therefore conducted a secondary analysis on an existing database from a large prospective multicentre randomized non-inferiority trial aimed at identifying the specific predictors of difficulty when using such devices. The primary analysis of the data from this trial trial revealed important new information about the comparative performance of the two tested acute-angle videolaryngoscopes (GlideScope^®vs C-MAC^® with D-blade) that included results about first-attempt intubation success and largyngeal view achieved for all 1100 patients and for certain subgroups (different patient and provider groups).¹¹ As part of the trial design, a comprehensive data set was collected for each patient that allowed a secondary analysis aimed at identifying predictors of difficult acute-angle videolaryngoscopy. Owing to the size and quality of the data used (complete data set of all relevant patient variables, surgical approach, head positioning, and provider characteristics), to our knowledge it provides the most reliable data set currently available for such analysis.

Methods

This is a secondary analysis of a randomized controlled non-inferiority trial of acute-angle videolaryngoscopy in 1100 patients with predictors of difficult direct laryngoscopy.¹¹ Institutional review board approval was obtained and written informed consent collected from all patients for randomization and subsequent data analysis, including secondary analyses such as the one presented in this report.

Data set

The trial design that resulted in the data set analysed in this study is described in great detail in an earlier publication.¹¹ In brief, the data originate from an 1100 patient, single-blinded, two parallel arm, randomized controlled non-inferiority trial comparing two videolaryngoscopes (GlideScope^® videolaryngoscope with a number 4 blade vs C-MAC^® videolaryngoscope with a D-blade) in patients with predictors of difficult direct laryngoscopy across three academic institutions between May 2013 and June 2014. The two treatment arms were not different in size (GlideScope, n=552; C-MAC, n=548), patient charcteristics, or bedside predictors of difficult airway. The study included patients undergoing general anaesthesia with neuromuscular block and a single-lumen orotracheal tube who presented with one or more of the following predictors of difficult direct laryngoscopy: (i) Mallampati classification score of III or IV (assessed in the sitting position with the tongue protruding and without phonation); (ii) reduced mouth opening (<3 cm, measured in the midline from the upper to lower teeth or gum), or (iii) large neck circumference (>40 cm for males and 38 cm for females, measured at the level of the cricoid cartilage). In addition, the patients were assessed for neck mobility (normal, limited, or severely limited), jaw protrusion (normal, limited, or severely limited), and thyromental distance (with neck in full extension).

Participating laryngoscopists included attending anaesthetists, anaesthesia residents, certified registered nurse anaesthetists, student registered nurse anaesthetists, anaesthesia assistants, and anaesthesia assistant students, all of whom were required to have both experience with videolaryngoscopy and at least 6 months of clinical anaesthesia experience. Non-anaesthesia providers (resident rotators), medical students, and other trainees were excluded from participation in the study. Before the study launch, all participating providers received formalized didactic instructions about the use of both videolaryngoscope devices and had ample opportunities for hands-on application (manikin). All study personnel also participated in a formalized online review of airway assessment (e.g. Mallampati and Cormack–Lehane scoring). These study personnel were then required to take and pass a test to validate their learning before study launch. After enrolment of 50% of the patients, all study personnel were required to take an online refresher module and pass another knowledge test before continuing in the study.

For each patient, pre-induction position was at the discretion of the laryngoscopist and recorded as ‘ramped’, ‘supine sniffing’, or ‘supine neutral’ according to the following predefined criteria: the ‘ramped’ position was defined as alignment of the sternal notch to the external auditory meatus in the horizontal plane using positioning devices or changes in the bed orientation (reverse Trendelenburg, elevation of the head of the bed, or bed ‘flexion’); the ‘supine sniffing’ position was defined as placement of a device/pillow/headrest under the head beyond a standard gel rest/pillow to elevate the head so that flexion of the neck and extension of the head were achieved; and the ‘supine neutral’ position was defined as the head on a standard gel/pillow headrest with no further manipulation.

Induction of anaesthesia was at the discretion of the attending anaesthetist but always included preoxygenation and the use of neuromuscular block. After confirmation of adequate relaxation, videolaryngoscopy was performed. Certified registered nurse anaesthetists, anaesthesia assistants, anaesthesia residents, and students were all supervised by an attending anaesthetist. During induction of anaesthesia, both a C-MAC device and a GlideScope were available. A rigid stylet (GlideRite^®; Verathon, Bothell, WA, USA) was used with all tracheal tubes, except when a bougie was used.

Outcome measures

We retrospectively analysed the data from the randomized controlled trial (n=1100). From the original trial, the primary outcome was successful intubation at first attempt. For the present study, the primary outcome determining ‘difficult acute-angle videolaryngoscopy’ was defined as first attempt intubation failure or first attempt intubation time >60 s. A successful intubation attempt was defined as tracheal tube placement as confirmed by persistent end-tidal carbon dioxide. Removal of the laryngoscope from the mouth or manipulation of the laryngoscope by anyone other than the primary laryngoscopist constituted a failure of the attempt. Intubation time was defined as the interval between blade insertion into the mouth and first recording of end-tidal carbon dioxide.

Statistical analysis

In order to make a univariate comparison of difficult vs not difficult laryngoscopy for categorical variables, a χ² test or Fisher's exact test was used. Normality of the continuous data was tested using the Shapiro–Wilks test. In order to make a univariate comparison of continuous variables between the two groups, a two-sample Student's t-test was used for normally distributed data, and Wilcoxon rank sum test was used when the normality assumption was not satisfied. Those covariates with univariate P-values <0.25 were included in the multivariate model as potential covariates. Multivariate logistic regression models were fitted along with stepwise model selection techniques to find the smallest set of covariates associated with the difficult laryngoscopy. Models were compared based on their Akaike's information criterion. Statistical significance of the two-way interaction terms for the included four main effect terms were examined. Those covariates with P-values of <0.05 were reported in the final multivariate model. The P-values were not adjusted for multiple comparisons. Data analyses were performed in SAS software 9.3 (SAS Institute Inc., Cary, NC, USA).

Results

The secondary analysis of the primary data set (n=1100) identified 301 patients who matched the definition of ‘difficult acute-angle videolaryngoscopy’ (time to first successful tracheal intubation exceeded 60 s). Amongst these, in 244 patients the first intubation attempt took longer than 60 s, in 27 patients more than one intubation attempt was required, and in the remaining 30 patients, intubation took both longer than 60 s and required more than one attempt.

Table 1 presents patient, surgical, provider, and treatment arm characteristics for those patients with ‘difficult’ vs ‘not difficult’ videolaryngoscopy; note that these two groups of patients were similar based on most characteristics. According to the univariate analysis, only limited mouth opening (P=0.018), type of surgical procedure (P=0.004), laryngoscopist training (P=0.001), and patient positioning at induction (P=0.008) were associated with ‘difficult videolaryngoscopy’.

As explained in ‘Methods’, those covariates with univariate P-values of <0.25 were included in the multivariate model selection process to examine their independent associations with the ‘difficult videolaryngoscopy’ in the presence of other covariates in the model. Table 2 reports the results of the final multivariate logistic regression model after the model selection; note that all four univariately associated covariates stayed in the final model, and none of the two-way interaction terms for the included four main effect terms was significant.

According to multivariate model results, in comparison to a ‘supine neutral’ position at induction of anaesthesia, the ‘supine sniffing’ position was associated with an increased risk of ‘difficult videolaryngoscopy’ {odds ratio (OR) 1.63, 95% confidence interval (CI) [1.14, 2.32]}. Anaesthesia residents who intubated under supervision of an attending anaesthetist were less likely to encounter difficult videolaryngoscopy than anaesthetists who performed the intubation themselves (OR 0.55, 95% CI [0.34, 0.87]). ‘Difficult videolaryngoscopy’ was more common in patients undergoing otolaryngologic surgery or cardiac surgery than in patients undergoing general surgery (OR 6.13, 95% CI [1.85, 20.37] and OR 1.89, 95% CI [1.19, 3.01], respectively). Finally, ‘difficult videolaryngoscopy’ was more likely in patients with reduced mouth opening (OR 1.18, 95% CI [1.02,1.36]).

Table 3 provides the association of laryngeal view with difficult laryngoscopy. It summarizes the frequency of Cormack–Lehane scores achieved by providers using acute-angle videolaryngosocpy in the primary study. The statistical analysis indicates a significant association between higher Cormack–Lehane scores and ‘difficult videolaryngoscopy’.

Discussion

The important findings of this analysis are that acute-angle videolaryngoscopy is more likely to be difficult when patients: (i) are in ‘supine sniffing’ position; (ii) have limited mouth opening; (iii) present for otolaryngologic or cardiac surgery; or (iv) are intubated by attending anaesthetists. Furthermore, the laryngeal view achieved during ‘difficult acute-angle videolaryngoscopy’ was worse than the view achieved with ‘not difficult’ acute-angle videolaryngoscopy, suggesting that difficult videolaryngoscopy using these types of devices is attributable not only to difficult tube passage, but also to inadequete laryngeal view.

Direct laryngoscopy and acute-angle videolaryngoscopy are entirely different airway management approaches, and it is likely that each has unique predictors of difficulty. Difficult direct laryngoscopy has been shown to be more likely in patients with increased BMI, large neck circumference, and high Mallampati score.^8,12,13 In contrast, the same covariates have not been identified as predictors of difficult videolaryngoscopy.

Our analysis identified four independent predictors for difficult intubations when using acute-angle videolaryngoscopes; two of them are modifiable (‘provider level’ and ‘positioning’), and the others are not (mouth opening and type of surgery), and none should be.

First, based on the multivariate logistic regresson model, the ‘supine sniffing’ position was identified as an independent predictor of an increased risk for difficult videolaryngoscopy compared with the ‘supine neutral’ position. The sniffing position has traditionally been considered the preferred position for direct laryngoscopy despite a paucity of data on the matter, and even despite contradictory evidence. The sniffing position was found to reduce the risk for difficult direct laryngoscopy in obese patients and in patients with limited neck extension (n=456).¹⁴ However, benefits for direct laryngoscopy may be restricted to male patients <50 yr old (n=200),¹⁵ and the sniffing position may be inferior to other positions in select patient populations.^16–18 The effect of head positioning on the success of acute-angle videolaryngoscopy has not been investigated before. Our results suggest that, in fact, the sniffing position hinders successful videolaryngoscopy-based orotracheal intubation. Our data are not suited to determine the cause of such an association, but reasons may include more challenging insertion of an acute-angle videolaryngoscopy blade in this position, because neck flexion may narrow the angle between the sternum and chin. In contrast, placing the patient in a ‘neutral’ or ‘ramped’ position was not associated with a higher incidence of difficult videolaryngoscopy. Given that the ‘ramped’ position, when compared with the regular supine position, prolongs time to desaturation in the obese patient,^19,20 this position seems better suited for acute-angle videolaryngoscopy in this population.

Second, our analysis also identified an effect of the provider background on the incidence of difficulty with acute-angle videolaryngoscopy. Unexpectedly, the data suggest that attending anaesthetists have a higher likelihood of encountering difficulties using these devices compared with residents. As for the other factors, the data do not allow us to identify the reason for these associations, and it remains unclear why residents would have a smaller risk for difficult videolaryngoscopy than providers with more experience. As residents tend to perform intubations more frequently than attending anaesthetists, our results might reflect this constant day-to-day clinical routine with laryngoscopy. Another advantage a resident might have is the additional eyes on the video screen to guide the intubation, whereas attending anaesthetists, working with their own patients, might have been intubating without any assistance. It is also possible that attending anaesthetists, when working with trainees, gravitated toward intubating the more difficult-appearing patients themselves. This would have resulted in bias toward increased difficulties for these providers. Given that provider background appears to influence the likelihood of encountering difficulties with videolaryngoscopy-based airway management, further focused research may be able to illuminate such relationships and test potentially effective interventions, such as training.

Third, our analysis suggests an increased risk of difficulty when acute-angle videolaryngoscopy is attempted in patients with a small mouth opening. This is in contrast to other studies, which did not show an increased risk of difficult or failed intubation with such devices in this setting (n=2004 and n=400).^9,10 A small mouth opening might hinder the initial placement of the videolaryngoscope and subsequent guidance of the tracheal tube toward and through the glottis. It remains to be determined whether videolaryngoscopes can have a low enough profile to be effective in patients with a small mouth opening. Until then, the patient's ability to open their mouth remains an important consideration when planning videolaryngoscopy with the GlideScope or C-MAC with a D-blade.

Fourth, our analysis suggests a higher likelihood for intubation difficulty when acute-angle videolaryngoscopy is attempted in patients undergoing otolaryngologic or cardiac surgery compared with those undergoing general surgery. Patients presenting for otolaryngologic surgery are at higher risk for difficult intubation when direct laryngoscopy is used,²¹ and our data suggest that this is also true when acute-angle videolaryngoscopy is used. Previous studies have suggested predictors of difficult videolaryngoscopy, which include an abnormal upper lip bite test,¹⁰ the presence of neck pathology, limited cervical spine motion, and short thyromental distance;⁹ each of which, alone or in combination, could be observed in patients presenting for otolaryngologic surgery. Our results echo these findings. Despite improved intubation success with acute-angle videolaryngoscopy, it is likely that the underlying pathology in some otolaryngology patients may make laryngeal viewing or tracheal tube placement difficult, even with these modern devices. In contrast to patients presenting for otolaryngologic surgery, in patients presenting for cardiac surgery videolaryngoscopy has not been studied, and it remains unclear why acute-angle videolaryngoscopy carries a higher risk for difficulty in this patient population. It should be noted that the absolute number of cardiac surgery patients in the analysed database was low, suggesting that additional research is necessary to validate these findings. Nevertheless, our data suggest that airway management may be difficult when using acute-angle videolaryngoscopy in both the otolaryngologic and cardiac patient populations.

This study has several limitations. First, this is a secondary analysis of data that were obtained to test other hypotheses. Although the quality of the data is high, the study design was not powered towards this secondary analysis. Second, all patients were selected based on predictors of difficult direct laryngoscopy. These predictors included enlarged neck circumference (90.8%), higher Mallampati score (54.7%), and small mouth opening (1.2%), and a large number of patients enrolled were obese (mean BMI 38.9 kg m⁻²).¹¹ Therefore, the analysed data do not allow us to compare normal vs obese patients or normal vs large neck circumference. Thus, we were not able to determine whether obesity or a large neck circumference are independent predictors of difficult intubation when using acute-angle videolaryngoscopy. A third limitation is the practice setting of large academic institutions. Some of the findings, such as those related to provider background, may not be immediately applicable to non-academic practice settings. Fourth, the study was set exclusively in the well-controlled environment of the operating room. The results, therefore, may lack external validity to other practice environments where tracheal intubations are attempted using acute-angle videolaryngoscopy, including emergency medicine, intensive care, or prehospital emergency care. Fifth, there may be other factors associated with the difficult acute-angle videolaryngoscopy that were not examined in our analyses. Sixth, the results may not apply to other videolaryngoscopes, such as those with ‘channelled’ designs or those that do not feature an acute angle. Lastly, multiple variables analysed in this study, such as provider level and patient position, were not randomized, so these results simply reflect associations with, rather than causal factors of, difficult videolaryngoscopy.

In summary, this secondary analysis of an existing database originating from a large multicentre randomized trial (n=1100) identified four predictors of difficult acute-angle videolaryngoscopy (‘supine sniffing’ position, limited mouth opening, planned otolaryngologic or cardiac surgery, and intubation by an attending anaesthestist). Amongst these factors, avoiding the ‘supine sniffing’ position and altering providers when attempting acute-angle videolaryngoscopy can easily be implemented in day-to-day clinical practice. However, our data demonstrate associations, not causality, so altering these variables may not change the intubation outcome. Acute-angle videolaryngoscopy appears useful only in selected patients with a small mouth opening, and its indication in this patient population should be considered very carefully. Future research is necessary to gain a better understanding of the risks of acute-angle videolaryngoscopy in cardiac surgery patients and for different groups of providers.

Authors’ contributions

Study design and data analysis: M.F.A., E.O.B., M.M.T., A.M.B.Writing all drafts of the manuscript: M.F.A., E.O.B., M.M.V.T.

Writing of all but the first draft of the manuscript: A.M.B.

Writing of the final manuscript: M.M.T.

Contributors

Storz Against Glidescope Efficacy (StAGE) Investigators

Ron O. Abrons, MD

Davide Cattano, MD, PhD

David E. Swanson, MD

Carin A. Hagberg, MD

References

Author notes