Introduction

The Cormack-Lehane (C-L) classification is the standard system for grading laryngeal views during direct laryngoscopy. Among its four grades, C-L Grade 3 airways—where only the epiglottis is visible—and Grade 4—where no glottic structures are visible—pose significant challenges to successful tracheal intubation1.

These high-grade airways are clinically significant due to their association with lower intubation success rates, particularly in high-acuity settings such as emergency departments or during rapid sequence intubation. Less experienced providers are especially vulnerable, often operating under time constraints, with suboptimal patient positioning and limited support—factors that increase the likelihood of failure.

To address these challenges, simulation-based training has become a cornerstone of airway education, enabling clinicians to rehearse difficult scenarios and evaluate adjunctive techniques in a safe and reproducible environment2. Notably, Cormack and Lehane themselves introduced the gum-elastic bougie to facilitate intubation in C-L Grade 3 airways, emphasizing the importance of device-specific strategies1.

The incidence of C-L Grade 3 airways varies widely depending on the clinical setting—ranging from 1.2% to 6.9% in operating rooms3,4,5,6, and 2.7% to 20.5% in emergency and intensive care units7,8. This variation likely reflects differences in patient acuity, procedural urgency, operator proficiency, and equipment availability. In some cases, suboptimal technique or poor positioning—particularly by novice intubators—may inadvertently result in C-L Grade 3 airways, with reported intubation success rates as low as 44.7% in these circumstances9.

A Cormack-Lehane (C-L) Grade 3 view during videolaryngoscopy can result from both anatomical and technical factors. Anatomical contributors—such as an anteriorly positioned larynx, short thyromental distance, high BMI, and limited neck mobility—impede vocal cord visualization10,11,12. Procedural issues, including suboptimal blade positioning, misapplied pressure in the vallecula, or excessive insertion depth, may fail to tension the hyoepiglottic ligament, preventing epiglottic elevation13. In some cases, excessive force may stretch or displace the ligament, sustaining a Grade 3 view despite VL use. Inadequate patient positioning—often overlooked by novice operators—can further contribute to failed intubation attempts. Awareness of these factors is essential for optimizing airway strategies.

Two simulation-based studies by Hung et al. investigated intubation strategies for Cormack-Lehane (C-L) Grade 3 airways—where only the epiglottis is visible—using direct laryngoscopy on modified manikins14,15. These studies demonstrated that mechanically lifting the epiglottis with a stylet-equipped endotracheal tube significantly improved first-pass success and reduced intubation time compared to preloaded bougie techniques. This suggests that epiglottic lift may be a valuable adjunct during direct laryngoscopy when glottic exposure is limited. In contrast, videolaryngoscopes—particularly those with remote cameras and hyperangulated blades—can often bypass the need for such maneuvers by providing improved glottic visualization without requiring direct line-of-sight. Therefore, videolaryngoscopy offers a strategic advantage in Grade 3 airways by allowing visualization beyond what is achievable through traditional direct lifting techniques.

Videolaryngoscopy (VL), with a distal camera, improves glottic visualization and may help overcome these challenges. Studies show VL enhances C-L grading compared to direct laryngoscopy7,16,17. In our previous study, both the videostylet (VS) and channeled VLs provided better glottic views and faster intubation than standard geometry VLs (SGVL), even in complex cases like Ludwig’s angina18. VL has also been linked to reduced failed intubation rates in difficult airways7,8,9,13,16.

Since the COVID-19 pandemic, VL adoption has grown, underscoring the need to refine its application. Given the link between failed intubation and poor outcomes, identifying the most effective VL device for C-L Grade 3 airways is crucial—especially when Plan A (direct laryngoscopy or SGVL) fails and clinicians must rely on Plan B or C.

Among videolaryngoscopy (VL) devices, the hyperangulated videolaryngoscope (HAVL) stands out due to its sharply curved blade, which facilitates intubation with minimal lifting force7,19,20. Optimal performance depends on using a matching curved stylet, which positions the larynx in the upper half of the screen and prevents an overly anterior or close-up view that may paradoxically hinder intubation. This aligns with Kovac’s sign, which emphasizes maintaining adequate distance to allow space for maneuvering21.

In contrast, the videostylet (VS)—a malleable device with a distal camera—is advantageous in scenarios with restricted mouth opening18. Its flexible design enables navigation behind the epiglottis, making it a promising option for managing C-L Grade 3 airways.

Simulating C-L Grade 3 airways using mannequins remains suboptimal due to rigid structures, constantly patent airways, and non-elastic skin, which fail to reflect the dynamic properties of human tissue. One study found that mannequins replicated real airway conditions in only 32% of cases22. In contrast, frozen cadavers preserve airway collapsibility and soft tissue elasticity, offering a more realistic and reproducible model for simulating C-L Grade 3 airways18,23.

While experienced clinicians are better equipped to manage difficult airways, the unpredictable nature of C-L Grade 3 airways—often influenced by anatomy or poor positioning—can challenge even seasoned providers5. A stepwise approach, which adapts the choice of intubation device based on the glottic view obtained during the previous attempt, is often more effective than repeated use of a single technique.

This study aims to evaluate the efficacy of SGVL with a bougie, HAVL, and VS in a standardized C-L Grade 3 cadaveric model, with the goal of identifying the most suitable VL device for use as a backup when initial intubation efforts fail.

Method

This study employed a randomized crossover design and was conducted at the Tzu Chi Medical Simulation Center in Hualien County, Taiwan, on December 6, 2023. The study adhered to the CONSORT crossover trial guidelines. The research received ethical approval from the Institutional Review Board of Dalin Tzu Chi Hospital in Chiayi, Taiwan, under the approval number B11203025. Two cadavers were preserved using the fresh frozen dry method within eight hours post-mortem, a technique that maintains realistic tissue quality for simulation23. The male cadaver, measuring 166 cm in height with a 2.7 cm mouth opening, had a BMI of approximately 22.9. The female cadaver, slightly shorter at 159 cm with a 3 cm mouth opening, had a BMI of about 21.0. Written informed consent was obtained from all individuals prior to their participation.

Although the C-L classification was originally developed for direct laryngoscopy, it remains widely used in studies involving video laryngoscopy to describe glottic visualization. Its continued use facilitates comparison across different airway devices and studies by employing a familiar, standardized, and well-validated grading scale7,8,9,18.

Based on our previous study comparing intubation techniques on a mannequin with a C-L grade 3 airway, we performed a sample size calculation. Assuming an odds ratio of 3.10 for successful intubation in the epiglottis manipulation group to estimate hyperangulated videolaryngosocpes or videostylet group, compared to the direct layngoscopy group to estimate the standard geometric videolaryngosocpe group15, we determined that 17 participants per group would provide 80% power at a significance level of 0.05. To mitigate the impact of within-participant correlations due to repeated measures, we plan to include 30 participants in total. This study enrolled 33 s-year resident physicians from various medical specialties who were participating in airway management training at the Cadaver Lab. All participants had fewer than 30 prior intubation experiences, reflecting limited clinical exposure to advanced airway management. The training program was meticulously designed, offering in-depth instructions and specific learning objectives for each intubation tool: Standard Geometric Video-Laryngoscope (SGVL), SGVL augmented with a bougie, High-Angle Video Laryngoscope (HAVL), and a videostylet (VS), as shown in Fig. 1. The training sessions were organized into one-hour blocks, each dedicated to mastering a particular device, and involved practice on two cadavers that were later utilized for simulation.

Fig. 1
Fig. 1
Full size image

The intubation equipment set includes a (A) standard geometric videolaryngoscope (SGVL), (B) SGVL augmented with a bougie, a (C) hyperangulated videolaryngoscope, and a (D) videostylet.

During the experiment stage, the instructors closely monitored each resident. Their goal was to ensure that the residents could skillfully execute intubations on a cadaver using a variety of devices: The devices used during the intubation procedures included: a standard geometry videolaryngoscope (SGVL; Macintosh size 3 blade, Trachway, Grandmedical Enterprise LTD., Taichung, Taiwan); SGVL combined with a 5.0 mm bougie (Sumi, Sulejówek, Poland); a hyperangulated videolaryngoscope (HAVL; D-blade, Storz C-MAC, Tuttlingen, Germany) used with its dedicated guide stylet (model 8401DS); and a video stylet (VS; model TVI-4102, Trachway, Grandmedical Enterprise LTD., Taichung, Taiwan). Participants were required to successfully perform a minimum of three intubations using these devices.

After a one-hour break, the residents were randomly assigned to perform intubations on simulated cadavers using the four settings previously mentioned, as illustrated in Fig. 1. The order of the settings had been randomized prior to conducting the study, using the National Cancer Institute Clinical Trial Randomization Tool (https://ctrandomization.cancer.gov/tool/). To simulate a consistent Cormack-Lehane Grade 3 airway—defined by visualization of the epiglottis only—we first transected the hyoepiglottic ligament, which normally facilitates indirect epiglottic elevation during laryngoscopy. To further restrict epiglottic mobility and prevent visualization of glottic structures, the epiglottis was then adhered to the vallecular region using medical-grade adhesive. This dual modification ensured that the epiglottis remained positioned over the glottic inlet, preventing visualization of the vocal cords and thereby replicating the limited view characteristic of a real-world Cormack-Lehane Grade 3 scenario. All simulated laryngoscopic views were recorded on video and reviewed to ensure consistency and reproducibility of the Grade 3 airway presentation prior to each intubation attempt. This approach allowed us to verify that the glottic view remained consistent across all trials and was not affected by prior manipulations.

Intubation attempts were digitally recorded from the moment the devices entered the cadaver’s mouth. Participants had the opportunity to continue until successful intubation, marked by device removal. Any attempt surpassing 90 s was classified as a failure12,20,21,22. After each intubation, participants assessed the difficulty using a Visual Analog Scale (VAS), rating their perceived ease or challenge of the intubation with the device. The video recordings from the study were analyzed to confirm the Cormack-Lehane (C-L) grade of 3 during laryngoscopy, ensuring consistency with the study’s objective. Additionally, the recordings were used to evaluate the duration and success rate of each intubation attempt.

Data analysis

This study utilized data from airway training sessions conducted with second-year residents. The dataset included participants’ profiles, intubation device types and their order of use, cadavers used, Cormack-Lehane grades, intubation durations, success rates, and visual analogue scores (VAS).

Kaplan-Meier survival curves were plotted to analyze the time to successful intubation and identify significant trends. To account for repeated measures from multiple attempts by the same participants, Cox regression models with a robust sandwich covariance matrix estimate were used. Hazard ratios (HRs) for successful intubation were calculated, adjusting for gender, intubation sequence, and cadaver type.

Multivariate logistic regression was applied to estimate the odds ratios (ORs) for intubation success within 90 s, with similar adjustments to ensure robust findings. Linear regression explored the relationship between intubation duration and VAS scores, while logistic regression assessed the ORs of success rates based on VAS scores.

Subgroup analyses evaluated HRs for successful intubation across individual cadavers using Cox regression models, adjusting for gender and intubation sequence. To address participant clustering in logistic and linear regression models, the generalized estimating equation (GEE) method was applied.

Statistical analyses were performed using SAS version 9.4 and STATA version 17.0. A two-tailed p-value of < 0.05 was considered statistically significant.

Results

The research involved 33 resident physicians, examining their success rates for intubation within a 90-second timeframe using different devices. The intubation process was recorded and reviewed to confirm that each glottic view, particularly with SGVL and HAVL, was classified as C-L grade 3 airways, ensuring consistency with the study objectives. The success rates observed were 57.6% for the Standard Geometric Video-Laryngoscope (SGVL), 60.6% for SGVL combined with a bougie, 63.6% for the Hyperangulated Video-Laryngoscope (HAVL), and 87.9% for the Video-Stylet (VS), as detailed in Table 1. The mean intubation times, along with standard deviations and 95% confidence intervals, were as follows: 63.3 ± 4.6 s (95% CI: 54.2–72.4) for SGVL, 69.5 ± 3.4 s (95% CI: 62.8–76.2) for SGVL with bougie, 63.8 ± 4.2 s (95% CI: 55.5–72.1) for HAVL, and notably shorter at 47.6 ± 4.3 s (95% CI: 39.1–56.0) for the videostylet (VS) (Table 1; Fig. 2).

Table 1 In the comparative analysis of intubation techniques, success rates and the duration of intubation were evaluated across four different methods.
Full size table
Fig. 2
Fig. 2
Full size image

A Kaplan-Meier failure analysis was employed to evaluate the success rate of intubation across different time intervals. Intubation attempts lasting longer than 90 s were categorized as failures. In this analysis, “VS” refers to the use of a video stylet (Trachway), “SGVL” denotes the employment of a standard geometric video-laryngoscope, and “HAVL” represents the use of a hyperangulated videolaryngoscope.

Hazard ratio analysis revealed that VS significantly outperformed SGVL in terms of intubation success, with a hazard ratio of 2.5 (95% CI: 1.4–4.4, p = 0.003), as shown in Table 2. The use of a bougie with SGVL or the HAVL did not significantly increase the odds of successful intubation compared to SGVL alone.

Table 2 The logistic regression analysis was employed to investigate the success rates of intubation across various devices, with the standard geometric videolaryngoscope (SGVL) serving as the reference point.
Full size table

Visual Analog Scale (VAS) scores were also evaluated and found to correlate significantly with intubation outcomes. Each 1 cm increase in VAS was associated with a decrease in the odds ratio for success (OR 0.8) and an increase of 2.4 s in intubation time (Table 2).

Glottic visualization was assessed using the Cormack-Lehane (C-L) grading system. The VS provided the best glottic view with a median grade of 1, followed by HAVL with a median grade of 2. SGVL and SGVL with bougie consistently presented grade 3 views, aligning with the study’s simulated model.

Additionally, factors such as participant gender, instructor identity, and the order of device usage did not significantly affect intubation success (Table 2).

Discussion

A Cormack-Lehane (C-L) grade 3 view, where only the epiglottis is visible, poses significant challenges for both experienced and novice practitioners1,4,5,6. Intubation under such conditions often involves blind techniques, increasing the risk of failure. In the era of advanced videolaryngoscopes (VLs), determining the most effective Plan B device for difficult airways is crucial7.

To ensure consistent and anatomically relevant Cormack-Lehane (C-L) Grade 3 airway conditions, we surgically transected the hyoepiglottic ligament and adhered the epiglottis to the vallecula. This approach aimed to prevent passive or inadvertent elevation of the epiglottis during laryngoscopy, a limitation often encountered in manikin-based simulations where participants are simply instructed not to lift the epiglottis14,15. Compared to these models, our cadaveric modification offers improved fidelity by replicating a fixed, reproducible, and visually obscured glottic inlet, consistent with clinical C-L Grade 3 presentations. Independent video review confirmed that the simulated airway view remained stable across all trials, supporting the internal validity of the model.

To minimize bias from practitioner preference and varying skill levels, this study enrolled novice participants who underwent a standardized airway training course. Participants developed proficiency with four devices: Standard Geometry Video-Laryngoscope (SGVL), SGVL combined with a bougie, Hyperangulated Video-Laryngoscope (HAVL), and Video-Stylet (VS).

Cormack-Lehane (C-L) Grade 3 airway is an anatomically challenging airway often necessitates the use of specialized or adjunctive devices to achieve successful tracheal intubation. VLs, particularly HAVL and VS, leverage advanced camera placement and angulation to provide improved visualization, addressing the limitations of direct laryngoscopy and SGVL8,16. This study evaluates the effectiveness of SGVL, SGVL with a bougie, HAVL, and VS in a simulated cadaveric setting. It also addresses the ongoing debate regarding the use of a bougie1,15, aiming to identify the most reliable device for managing C-L grade 3 airways.

Our study demonstrates that VS, which require specific techniques unlike SGVL or HAVL, enable novice intubators to attain high success rates (87.9%) with minimal training period. VS led to the shortest average intubation time of 42 s, predominantly achieving C-L grade 1 (Table 1). This is in contrast to the longer intubation times associated with SGVL, both with and without a bougie, and HAVL (Fig. 2). The principal challenges were due to inadequate inline technique and sequencing. This finding emphasizes the importance of adequate training to successfully navigate and identify structures during intubation, especially before considering VS as an alternative approach in C-L Grade 3 scenarios. Hazard ratio analysis revealed that VS outperformed SGVL in terms of success rates, with a ratio of 2.5 and a 95% confidence interval ranging from 1.4 to 4.4, showing statistical significance with a p-value of 0.003, as shown in Table 2.

The Cormack-Lehane (C-L) grading system, originally developed for direct laryngoscopy (DL) training and decision-making, is now routinely applied in anesthetic documentation for sharing critical information. The frequency of encountering difficult airways, classified as C-L grade 3 or 4 airways, varies widely, ranging from 1.2% to 6.9% in operating rooms4,5,6 and 2.7% to 20.5%7]– [8 in emergency and intensive care settings. For challenging airways, first-pass success rates with DL are notably lower, at 44.7% for C-L Grade 3 cases, but improve to 66.7% with alternative techniques9.

Driver et al.‘s study demonstrated a higher success rate for first-attempt intubation in patients with difficult airways using a bougie (98%) as opposed to a stylet (87%), and this was consistent even with no significant differences in intubation time or hypoxemia rates24. The study didn’t specifically focus on C-L grade 3 airways but included a subgroup analysis comparing 27 patients with C-L grade 3 airways who used a bougie to 23 patients who used a stylet, with no significant differences in outcomes observed. Cormack and Lehane suggested that employing a tracheal introducer, commonly known as a bougie, with a 30-degree upward curve could improve the success rate of first attempt intubations and might be an effective strategy for intubating airways classified as Cormack-Lehane grade 31,14.

In our study focusing on C-L grade 3 airways, we observed that bougie use frequently resulted in misdirection toward the esophagus due to restricted glottic visualization. Our analysis revealed a slight improvement in intubation success rates using the SGVL with a bougie (60.6%) compared to the SGVL alone (57.6%). However, both the hazard ratio (0.9) and odds ratio (1.1) suggest that this difference does not represent a significant enhancement in outcomes (Table 2; Fig. 2). Kovac et al. conducted a study comparing the use of a videostylet and a gum elastic bougie on a manikin model25. Their findings indicated that when encountering a Cormack-Lehane grade 3 view, the first attempt success rate dramatically decreased with the gum elastic bougie, from 99% to 6%, whereas the videostylet maintained high success rates, dropping only slightly from 90% to 87%. Additionally, the mean time to achieve successful intubation increased with the gum elastic bougie, from 31 s to 45.6 s, compared to a minimal increase from 29.2 s to 31 s with the videostylet. Therefore, in cases of C-L grade 3 airways, incorporating a bougie with SGVL did not markedly improve outcomes.

The HAVL, with its over 50-degree blade curve, enhances glottic visualization7,18. However, encountering Kovac’s sign, which occurs when the camera on a videolaryngoscope is too close to the glottis, can restrict the maneuverability of a stylet-equipped endotracheal tube21. This spatial limitation may lead to reduced success in intubation efforts. Our study reflects this: despite a median C-L view of grade 2, better than SGVL with or without a bougie (grade 3 view), there was no notable improvement in intubation success or time. The hazard ratio and odds ratio were 1.1 and 1.3, respectively, lacking statistical significance (Table 2; Fig. 2).

In videolaryngoscopy, indirect lifting refers to placing the blade tip in the vallecula to tension the hyoepiglottic ligament, whereas direct lifting involves advancing the blade tip beneath the epiglottis to elevate it. Clinical evidence suggests that direct lifting can substantially enhance glottic visualization. For example, in a prospective study using a standard geometric (Macintosh-style) videolaryngoscope, direct lifting improved modified Cormack–Lehane grades and POGO scores compared with indirect lifting (87.5% vs. 64.4%)26. More recently, Wünsch et al. (2023) analyzed 129 difficult Macintosh videolaryngoscopy cases and found that direct epiglottis lifting increased POGO by a mean of 49.7% (95% CI 41.4–58.0), which was non-inferior to switching to a hyperangulated videolaryngoscope (+ 43.7%; 95% CI 34.1–53.3)27. A combination of both maneuvers achieved the best glottic exposure.

In our cadaveric grade-3 model, we intentionally transected the hyoepiglottic ligament and fixed the epiglottis to the vallecula, and no direct lifting was performed with any videolaryngoscope. This design prevented view conversion via lifting and instead isolated each device’s performance when only the epiglottis was visible. As a result, our findings may underestimate the clinical utility of SGVL (and possibly some HAVL), where selective direct lifting could convert views and facilitate tube passage. Nevertheless, the consistently strong performance of the videostylet in our study underscores its robustness even when lifting maneuvers are unavailable. Practically, after encountering an epiglottis-only view with SGVL, clinicians may either attempt direct epiglottic lifting (when safe and anatomically feasible) or transition early to devices that decouple visualization from tube delivery, while balancing potential trade-offs such as mucosal trauma and blade ergonomics.

Limitation

First, although the study involved second-year resident physicians who underwent standardized training, their relative inexperience with difficult airway management may have influenced device performance outcomes. While this aligns with real-world settings where junior trainees often perform intubations, it limits the applicability of our findings to more seasoned clinicians. Second, the use of dry-preserved cadavers, while offering anatomical consistency and reproducibility, lacks several physiological features of live patients—such as mucosal pliability, secretions, reflex responses, and dynamic airway collapse—potentially oversimplifying the challenge of a true Cormack-Lehane grade 3 scenario. Third, our simulation of grade 3 views involved intentional surgical and adhesive modifications. While these methods achieved the intended visual obstruction, they may not fully capture the spectrum of natural anatomical variations or the unpredictability encountered in live clinical settings. Finally, as with any cadaveric or simulation-based study, the controlled and standardized environment does not reflect the time pressures, hemodynamic instability, or multi-tasking demands typical of real-world emergency airway management. These factors should temper direct extrapolation of our results to actual clinical practice.

Conclusions

In this cadaver-based simulation of Cormack-Lehane (C-L) Grade 3 airways, the videostylet (VS) demonstrated a higher first-pass success rate and shorter intubation time compared to the standard geometric videolaryngoscope (SGVL), SGVL with bougie, and hyperangulated videolaryngoscope (HAVL). While the bougie improved intubation success over SGVL alone, its performance was inconsistent, with occasional esophageal misdirection observed. The HAVL provided enhanced visualization but did not consistently translate into higher intubation success, possibly due to difficulty in maneuvering the endotracheal tube despite adequate glottic views. These findings suggest that the VS may offer certain advantages in simulated C-L Grade 3 scenarios; however, further studies in clinical settings are warranted before establishing it as the preferred backup strategy.