Prospective, randomised, controlled study comparing Videolaryngoscopy versus Direct Laryngoscopy for Double-Lumen Endotracheal Tube Intubation in thoracic surgery

Methods: A clinical, prospective, randomised, controlled trial was conducted with a total of 70 patients undergoing elective thoracic surgery and need for a DLT intubation. Finally, 65 patients were analysed (DL group [n = 31] vs. GVL group [n = 34]). Endpoints were time to intubate, subjective symptoms and objective trauma of the oropharynx and supraglottic space. Results: Mean intubation time in GVL group was 178 seconds ± 279 versus 84 seconds ± 62 in DL group [p = 0.044]. Videolaryngoscopy resulted in significantly improved visualisation of the larynx compared with Direct Laryngoscopy (DL). Cormack and Lehane (CL) grade of 1 was more frequent in GVL group (97%) than in DL Group (74%) [p = 0.008]. In analysis of the postoperatively completed questionnaires, no significant differences regarding the subjective symptoms could be determined. Objectifiable endoscopic examinations revealed significant differences in GVL group compared to DL group showing less hypopharyngeal haematoma [p = 0.048], red-blooded vocal cord [p = 0.01], vocal cord haematoma [p = 0.027] and vocal cord haemorrhage [p = 0.03]. Conclusions: study was the incidence of intubation-related injuries in both groups. Therefore, we performed two consecutive transnasal flexible endoscopic examinations of the oropharynx, of the supraglottic space and of the vocal cords, a follow-up survey by questionnaire and a dental examination to detect dental trauma. intubation

trauma. The different objective intubation traumata did not seem to affect subjective symptoms of the patients after a DLT intubation.

Background
Modern thoracic surgery is supported by one-lung ventilation, which can be achieved by intubation with a double-lumen endotracheal tube (DLT) or a standard single-lumen endotracheal tube and an additive bronchial blocking method (1). Compared to a single-lumen endotracheal tube, a DLT has a larger outer diameter, and is longer and more rigid. Therefore, the DLT intubation procedure is more challenging and the incidence of injuries to the pharynx, larynx (particularly to the vocal cords), and trachea during tube placement is higher than during placement of a standard single-lumen endotracheal tube combined with a bronchial blocker (2)(3)(4). Videolaryngoscopic intubation with a standard single lumen endotracheal tube using a GlideScope® causes less soft tissue damage compared to conventional MacIntosh laryngoscopy (5). The advantages of videolaryngoscopy for DLT intubation compared with traditional MacIntosh laryngoscopy (DL) is currently being discussed controversially in terms of differences in intubation times and the associated injuries (6,7). Moreover, some meta-data with moderate to low quality evidence exist, showing a higher success rate at first attempt, a higher incidence of malpositioned double-lumen tube and a lower incidence of oral, mucosal or dental injuries with videolaryngoscopy for DLT intubation (8). However, the benefit of using videolaryngoscopy for intubation with DLTs is still unclear.
GlideScope® videolaryngoscopy with a 60° hyper-angulated blade was developed in Canada in 2001 by Pacey (9,10). Videolaryngoscopy itself with high-resolution digital camera often enables a better view of the vocal cord (9)(10)(11)(12)(13). There is convincing data for using videolaryngoscopy in difficult airway management and reducing cervical spine movement (11,14,15). Current research shows an improvement in intubation conditions and first success rates with slightly longer intubation times through GlideScope® videolaryngoscopy (13). In 2014, the GlideScope®-Titanium video laryngoscope was launched. The GlideScope®-Titanium Single use blades are thinner than the previous blades of the GlideScope®. For GlideScope®-Titanium, single use blades are 3 mm (Size 3) and 2.7 mm (Size 4).
In our study, we investigated the effect of videolaryngoscopy on the DLT intubation and especially the potential benefit of one particular videolaryngoscope (GlideScope® Titanium single use blade), which had a blade that was nearly 3 mm thinner than the currently used GlideScope® videolaryngoscope. A thinner blade design should be useful during the intubation of patients with a small oral cavity or limited mouth opening capacity. Presumably this device offers more space in the pharynx during DLT intubation. In contrast to the previous studies, we tried to increase the quality of our follow-up, detecting subjective symptoms and objectifiable injuries after DLT intubation. Therefore, in addition to a patient questionnaire, we used two consecutive investigator-blinded nasal endoscopic examinations for follow-up. The video files of these endoscopies were evaluated by three independent investigators. The main exclusion criterion was patients aged less than 18 years. In addition, non-fasting patients, pregnant women and patients with gastro-oesophageal reflux disease, which were indicated for rapidsequence induction, were excluded from the study. Other exclusion criteria were the following: contraindication to a left or right DLT; contraindication to one-lung ventilation; and an abnormal physical status of the Cervical spine (e.g., after C-spine trauma, Bechterew's disease).

Methods
Patients were pre-medicated with 3.75-7.5 mg oral midazolam 45 minutes before surgery. On arrival in the operating room (OR), all participants were blinded, randomly assigned to either conventional laryngoscopy (MacIntosh) or GlideScope® videolaryngoscopy (GlideScope; Verathon Inc., Bothell, WA) by sealed envelope randomisation. In the OR, every patient received pre-oxygenation with 100% oxygen through a mask over 5 minutes. After pre-oxygenation, anaesthesia was induced with 0.3 µg/kg sufentanil and 2 mg/kg 1% propofol, followed by a total intravenous anaesthesia (TIVA) according to the in-house standard operation procedure (SOP) in anaesthesia for thoracic surgery with propofol (4-6 mg/kg bodyweight/h) and remifentanil (15-25 µg/kg bodyweight/h). Muscle relaxation was induced with 0.6 mg/kg rocuronium bromide followed by at least 3 minutes of bag-mask ventilation. The depth of anaesthesia was monitored by measuring derived EEG signals using Bispectral Index monitoring (BIS). The neuromuscular monitoring was performed by a relaxometry Train of Four (TOF). DLT intubation was not performed until there was a sufficient depth of sleep and full relaxation of the patient, which was checked by the full relaxation status (TOF 0/4) and a BIS value below 60. Intubation was performed using a MacIntosh blade (size 3 or 4) in the MacIntosh group or with the GlideScope®-Titanium Single-Use-blade with hyper-angulation (size 3 or 4) in the videolaryngoscopy group. At the start of our study, there was no rigid stylet (GlideRite® Guide) for double-lumen tubes (DLT) available. Therefore, we used DLT with the original inner stylet in both groups. In the GVL group, we used the original rigid stylet GlideRite® for the single-lumen endotracheal tube as a template to shape the inner stylet and to replicate the curve of the GlideScope according to the hyper-angulation videolaryngoscopy-blade with the 60 degree angle. An appropriately sized DLT (Rüsch Bronchopart; Teleflex Medical GmbH, Dublin, Ireland) was chosen for DLT intubation by the experienced physician. All DLT intubations were performed by three experienced senior physicians in the department. All three physicians who took part in this study were specialists who regularly perform thoracic anaesthesia and regularly use the GlideScope® for tracheal intubation. All three of the physicians have routinely performed more than 1000 intubations with a DLT using direct laryngoscopy. In addition, they also had a lot of experience in videolaryngoscopy with GlideScope® (primary videolaryngoscope established as a routine in the department since 2010) for a single-lumen endotracheal tube. Nevertheless, all three physicians had little experience using the GlideScope® for DLT insertion and no exercise with the new GlideScope® Titanium single use blade before the investigation started.
The primary outcome was duration of endobronchial DLT intubation regardless of the number of attempts. The intubation time was defined as: blade passes mouth opening → positive capnography (three visible CO 2 curves at the capnograph). In this context, we also reported the number of intubation attempts, the assessment of difficulty and any complications during DLT intubation in both groups.
The secondary endpoint of this clinical study was the incidence of intubation-related injuries in both groups. Therefore, we performed two consecutive transnasal flexible endoscopic examinations of the oropharynx, of the supraglottic space and of the vocal cords, a follow-up survey by questionnaire and a dental examination to detect dental trauma.
The first endoscopic examination was under general anaesthesia before extubation, while the second endoscopic examination took place 24 h after extubation under local anaesthesia. Both endoscopic examinations were performed by the main investigator.
All endoscopic examinations were anonymised and blinded to the technique as video files and evaluated by three independent investigators (2 anaesthesiologists and 1 ENT specialist, investigatorblinded). Both endoscopic examinations of each patient were compared. The hypopharynx, the vocal cords and the arytenoid cartilage were evaluated on the basis of various criteria. The different criteria were scored from zero to three depending on the degree of the findings (0 = not assessable, 1 = without pathological findings, 2 = minor injuries, 3 = severe injuries). The results were averaged for further analysis. Finally a physician of oral and maxillofacial surgery (investigator blinded) performed a dental examination after DLT intubation in all study patients, examining the patient for lip and dental trauma, such as enamel fractures.
Our sample size calculation was based on a previous study from 2012 (7), which reported a mean (SD) time of 46 (11) s for DLT placement with videolaryngoscopy. Based on these results and factoring in a 20% drop-out rate, we calculated that 35 patients were required in each group to detect a 20% difference in the time taken for DLT intubation. Our sample size was calculated for a data analysis using a non-parametric Mann-Whitney U test. The a priori power analysis was performed with a (p) of 0.80 and a significance level α = 0.05.
The values for descriptive statistics were expressed as using means, standard deviations and percentages with the Student's t-test as appropriate. Non-parametric data were analysed using the Mann-Whitney U-test, with p < 0.

Results
After written informed consent, 70 patients were recruited. Out of the 70 patients 65 completed the study and were included in the final analysis (Fig. 1). Four patients in the MacIntosh group and one patient in the GlideScope® group were excluded from the final analysis. In two participants randomised to the conventional MacIntosh group, the conventional DLT intubation failed and the experienced examiner changed the method using videolaryngoscopy. Finally intubation with a single endotracheal tube and a bronchial blocker had to be performed in these two cases because of impossible DLT intubation, with both devices conventionally with DL and with videolaryngoscopy (GVL). Two participants randomised to the conventional MacIntosh group refused postoperative nasal endoscopic examination and one participant in the GlideScope group needed long-term postoperative ventilation on the intensive care unit and was lost to follow-up. All five participants were excluded from the final analyses due to relevant study protocol violation (Fig. 1).
Both groups had no significant differences in bio-metric data and pre-operative airway assessments (Table 1). In almost all cases, a left-sided DLT was used. It was noticeable that all three experienced specialists in thoracic anaesthesia who performed the DLT intubations in the study used smaller blades and tended towards smaller tube sizes in the GVL group. The difference between the two groups was statistically significant [p < 0.05] ( Table 1).   (Table 2).
At endoscopic control a correct DLT position directly after successful endobronchial intubation was reported in 77% of the MacIntosh group and only 53% of the GVL group. The difference observed between the two groups was statistically significant [p = 0.041] ( Table 2).
There was no other significant difference in terms of direct complications under DLT intubation between the two groups. All parameters like oxygen saturation (SpO 2 < 85%), bronchospasm and arrhythmias which occurred during DLT intubation did not differ significantly between the GVL and MacIntosh groups. Blood on device after successful endobronchial intubation and carina traumata endoscopically examined directly after successful endobronchial intubation were similar between the two groups (Table 2). Furthermore, lip and dental trauma, as well as enamel fractures examined by the dental follow-up, were not significantly different in both groups (Table 3).

Discussion
We compared DLT tracheal intubation with videolaryngoscopy with a thin video-blade to the conventional MacIntosh laryngoscopy. We found significantly prolonged intubation times and no significant difference in intubation success despite significantly better intubation conditions compared to conventional laryngoscopy. In addition, there was a significantly higher incidence of malpositioned double-lumen tubes in the videolaryngoscopy group. Our complex investigator blinded nasal endoscopic follow-up after DLT intubation showed significant differences in only five of the seventeen objectifiable parameters, e.g. a reduced incidence of vocal cord redness, vocal cord haematomas, vocal cord bleeding and hypopharynx haematoma in the videolaryngoscopic group. The dental followup showed no significant differences in DLT intubation-related injuries between the two groups. These results of reported intubation-related injuries from two trans-nasal endoscopic examinations after DLT intubation do not seem to have an impact on patient wellbeing, since none of the subjective symptoms like sore throat, dysphagia, cough and hoarseness differed between the two groups.
Prolonged intubation times for GVL videolaryngoscopy for DLT intubation were shown in previous studies (6,16,17). Prolonged intubation times inevitably have a greater risk of hypoxia and could be harmful to patients with pulmonary comorbidities. In a prior study by Russell et al. in 2012, anaesthetists found that GVL was more difficult to use than the MacIntosh laryngoscope and ETI took longer. In their study all DLT intubations were performed by less experienced novice anaesthetists (6), whereas in our study, all DLT intubations were performed by three consultants of anaesthesiology well experienced with DLT intubations in thoracic anaesthesia and GlideScope videolaryngoscopy.
Nevertheless, we could also show prolonged intubation times with GVL. A limitation of our study was, that prior to present investigation our consultants had little experience using the Titanium single use blade GlideScope for DLT insertion. The first attempt success rate reported here using GVL for DLT intubation was 85%. Our reported failure rate of 15% at the first attempt using GVL is similar to most of the results reported by other groups (6,18,19). We were also unable to show a 100% first pass success rate with GVL, like the group of Hsu et al. (7). In the study of Hsu et al., all DLT intubations were performed by two experienced anaesthetists, who had both performed over 300 intubations using DLT with GVL. In addition, the BURP manoeuvre was not required for successful DLT intubations with GVL (7).
The process of advancing the DLT past the vocal cords seems to be the main sticking point for using hyper-angulated videolaryngoscopy (6,20). Our findings support that DLT tube delivery and advancement into the trachea is the most difficult step in the procedure using videolaryngoscopy with hyper-angulated non-channelled blades for DLT intubations which causes prolonged intubation times for GVL. This assumption is supported by our data. Our experienced anaesthetists needed a BURP manoeuvre in 32% of all DLT intubations with GVL and we reported significantly more malpositions of the DLT (in 47% of cases). Presumably caused by the rotation manoeuvre, which is needed for the advancement into the trachea, there is a higher incidence of main bronchus malposition of the DLT.
Caused by the required bending of the DLT-tube for hyper-angulated blades, the tip of the DLT often hits the ceiling when trying to insert the tube into the trachea past the vocal cords. Usually, a BURP manoeuvre is first performed to adjust the trachea, positioning it more posterior and more in line.
Second, a rotation manoeuvre could be necessary. Such a rotation manoeuvre was described by Bustamante and Hernandez (21,22). Rotation manoeuvre more often results in the incorrect position of the DLT. Liu et al. concluded that the use of videolaryngoscopes, especially with a hyper-angulated blade for DLT intubation, complicates the already complicated DLT intubation technique through rotation manoeuvres (8). Our data support the thesis that these sequential rotation manoeuvres are probably the reason why videolaryngoscopy increases the incidence of mispositioning of DLT. We could not confirm our hypothesis that a thinner hyper-angulated blade provides better visibility and in consequence more space for a rotation manoeuvre and therefore a lower incidence of sore throat and hoarseness. Hsu et al. were able to show a lower incidence of sore throat and hoarseness (7). There are controversial results in the literature. Russell et al. were unable to identify any significant differences in their study (6). Due to the controversial results in the current literature and our results, the question remains whether the questionnaires used are sensitive enough to record differences in subjective symptoms.
In regard to the incidence of dental trauma, a study by  comparing DL and VL showed that less force is exerted on the teeth of the upper jaw when using VL (23,24). This is in accordance with the results of the current study situation (6,17). Our dental follow up showed no significant differences in DLT intubation-related injuries. However, our study was not powered to detect that the incidence of dental trauma is different. Limitations: This study was randomised, but has some limitations. First, the operators were not blinded to the intubation device used; however, it is difficult to circumvent this problem when evaluating different intubation devices. Nevertheless, the patient and follow-up endoscopic examinations were anonymised and blinded. A second limitation might be that the operators were not equally experienced with both intubation devices. Videolaryngoscopy is used routinely in our department, but not for DLT intubation. Furthermore, the new GlideScope®-Titanium was used for the first time during this study in our hospital.
A further limitation of our study is the small number of patients with a supposed difficult airway (Mallampati 3 and 4, 13% in the DL group vs. 17% in the GVL group) and the low incidence of predicted difficult airways (CL 3 and 4, 3% in the DL group vs. 0% in the GVL group).
In addition, an appropriate rigid stylet for the DLT intubation with the GlideScope, like the GlideRite® Rigid Stylet, which is standardly used for the single lumen tube intubation, was not available at the start of the study (18). Instead, we used the original rigid GlideRite® stylet for the single-lumen endotracheal tube as a template to shape the inner stylet of the DLT. A technique like the one developed and described by Bussier et al. and Bustamante et al. was not mandatory for our anaesthetists (18,21).
DLT intubation with the GlideScope®-Titanium might be improved if the users are given some additional training and use the adequate rigid stylet for DLT, which keeps its shape and is better adapted to the hyper-angulated blade of the GVL.
There are currently many different videolaryngoscopes with varying designs and quality available on the market. For these reasons, our study results should not be generalised, and further investigation into VL and DLT intubation is needed.

Conclusions
In conclusion, DLT tracheal intubation using GlideScope Titanium videolaryngoscopy with an extra thin blade takes more time while allowing a better laryngeal view and causing less objectified trauma than the MacIntosh laryngoscope. In our study collective, this had no impact on subjective symptoms such as sore throat, dysphagia, cough, and hoarseness after DLT intubation. We conduct our study in compliance with recognized international standards, including the principles of the Declaration of Helsinki.
All experiments and data collection were performed with prior informed consent to participate.
Written consent to participate in the study were obtained from each participant.

Consent for publication
Not applicable Availability of data and material The data that support the findings of this study are available from the corresponding author. The datasets used and analysed during the current study are available from the corresponding author on reasonable request.