A randomised controlled trial comparing of the efficacy and safety of left lung isolation for minimally invasive direct coronary artery surgery using video-imaging double-lumen endobronchial tube with a bronchial blocker paced through a video-imaging single-lumen tracheal tube

Abstract

Background and objectives

Minimally invasive coronary artery bypass grafting (MIDCAB) requires effective left lung isolation, yet evidence comparing video-imaging single-lumen tracheal tubes with bronchial blockers (VSLT + BB) and video-imaging double-lumen endotracheal tubes (VDLT) remains limited. This randomized controlled trial aimed to: (1) quantitatively compare time efficiency for device placement and lung isolation between VSLT + BB and VDLT; (2) evaluate perioperative airway complications; and (3) assess differential impacts on postoperative recovery, including postoperative sore throat, hoarseness, and Quality of Recovery-15 (QoR-15) scores.

Methods

In this single-blind randomized controlled trial, 97 MIDCAB patients were allocated to VSLT + BB or VDLT groups. Primary outcomes were tube positioning time; secondary outcomes included total intubation time, oxygenation parameters, hemodynamic variables, and postoperative complications.

Results

VDLT exhibited shorter tube positioning time (128 ± 37 vs. 159 ± 58 s; p < 0.001) but longer total intubation time (192 ± 40 vs. 159 ± 58 s; p < 0.001). VSLT + BB demonstrated higher PaO₂ at 10-min post-OLV (226.0 vs. 168.0 mmHg; p = 0.035) with lower airway pressures (p < 0.05). Postoperative sore throat (33% vs. 13%; OR 2.84, 95% CI 1.04, 7.71; p = 0.018) and hoarseness at 48 h (55% vs. 31%; OR 2.45, 95% CI 1.08, 5.59; p = 0.018) were higher with VDLT. Lung collapse quality, hypoxemia rates, pulmonary complications, and QoR-15 scores showed no significant differences (all p > 0.05).

Conclusions

In MIDCAB surgery, VDLT demonstrated significantly shorter tube positioning time compared with VSLT + BB. However, VSLT + BB exhibited shorter total intubation time than VDLT by avoiding postoperative tube exchange. Both techniques provided clinically acceptable lung isolation with comparable lung collapse quality and hypoxemia incidence. VSLT + BB exhibited lower airway pressures and higher oxygenation indices during early OLV, whereas VDLT was associated with higher rates of minor airway complications (sore throat, hoarseness). No significant differences were observed in pulmonary complications, hemodynamic stability, or recovery quality.

Trial registration

ChiCTR2300072124, 3/6/2023.

Introduction

The global incidence of coronary artery disease (CAD) has shown a persistent upward trend, establishing it as one of the leading causes of mortality worldwide [1]. Coronary artery bypass grafting (CABG) remains an essential revascularization strategy for patients with complex coronary anatomy or elevated SYNTAX scores [2]. With continuous advancements in CABG techniques, mortality rates at major cardiac centers have decreased to below 0.6%, prompting a research shift toward optimizing long-term outcomes through minimally invasive approaches while ensuring surgical safety [3]. This paradigm evolution emphasizes the safe implementation of minimally invasive techniques to achieve optimal long-term therapeutic efficacy [4]. Comparative studies demonstrate that minimally invasive coronary artery bypass grafting (MIDCAB) yields surgical outcomes comparable to conventional off-pump CABG (OPCABG) via median sternotomy [5, 6]. Notably, exhibits significant advantages in acute postoperative recovery parameters, including reduced mechanical ventilation duration, decreased chest tube drainage volume, lower transfusion requirements, and shorter hospital stays, without increasing risks of postoperative complications [7, 8]. However, the technical challenges posed by restricted surgical exposure in anterior thoracic regions and constrained operative spaces inherent to MIDCAB necessitate enhanced surgical precision and specialized anesthetic management [9].

The successful implementation of one-lung ventilation (OLV) constitutes a critical prerequisite for optimal surgical field exposure. Current clinical practice establish left-lung isolation for surgical exposure through two primary modalities: a double-lumen endotracheal tube (DLT) or a single-lumen endotracheal tube (SLT) combined with a bronchial blocker (BB), each offering distinct technical advantages. Irrespective of the intubation technique selected, accurate positioning of the endotracheal tube or blocker necessitates verification via fiberoptic or electronic bronchoscopy to ensure effective lung isolation. Recent innovations in airway management technology have introduced visual endotracheal tubes, characterized by integrated high-definition micro-cameras at their distal tips, enabling continuous real-time visualization of tube placement. This advancement substantially enhances intraoperative monitoring efficiency and operational convenience for anesthesiologists. However, the integration of visual components increases the outer diameter of both video-imaging single-lumen tracheal tubes (SLTs) and double-lumen endotracheal tubes (DLTs) compared to conventional counterparts of equivalent sizes, potentially exacerbating hemodynamic instability and airway mucosal injury during intubation [10].Existing comparative studies between DLTs and bronchial blockers (BBs) primarily focus on thoracoscopic surgeries, typically evaluating standard or video-imaging DLTs against conventional SLTs paired with BBs [11–13].To date, no investigations have directly compared video-imaging SLT with BB (VSLT + BB) and video-imaging DLT (VDLT) in patients undergoing minimally invasive cardiac surgery (MICS) for coronary artery bypass grafting (CABG). This study represents the first comparative analysis of these two advanced airway management strategies in MIDCAB, aiming to establish evidence-based recommendations for optimizing perioperative ventilation protocols in this specific surgical patients [14]. Therefore, we hypothesized that the VSLT + BB strategy would demonstrate advantages in reducing total intubation time and minimizing airway trauma while providing comparable lung isolation efficacy. We further postulated enhanced postoperative recovery with this approach. The present randomized controlled trial aimed to: (1) quantitatively compare time efficiency for device placement and lung isolation between VSLT + BB and VDLT; (2) evaluate perioperative airway complications; and (3) assess differential impacts on postoperative recovery, including postoperative sore throat, hoarseness, and Quality of Recovery-15 [QoR-15] scores—ultimately providing evidence-based guidance for airway management in MIDCAB.

Methods

Study design and ethics

This single-blinded, randomized controlled trial was conducted at Peking University Third Hospital (Beijing, China) with approval from the Medical Scientific Research Ethics Committee of the hospital (No. IRB00006761-M2023282). Written informed consent was obtained from all participants or their legally authorized representatives prior to surgical intervention. The study was prospectively registered with the Chinese Clinical Trial Registry (ChiCTR 2300072124) before patient enrollment and followed Consolidated Standards of Reporting Trials (CONSORT) guidelines.

Patients

Eligibility criteria required participants to: (1) be 18–80 years old; (2) undergo MIDCAB with multivessel revascularization via sequential anastomosis techniques; (3) be classified as American Society of Anesthesiologists (ASA) physical status Ⅲ to Ⅳ; and (4) provide written informed consent. Exclusion criteria included: (1) tracheobronchial anatomical abnormalities or tumors; (2) history of prior main bronchus surgery; (3) morbid obesity (BMI > 35 kg/m²); (4) severe systemic comorbidities (such as sepsis, severe chronic obstructive pulmonary disease); (5) clinically suspected or confirmed difficult airway; (6) inability to comply with study protocols; (7) anticipated to require lung isolation for a duration of < 30 min or > 6 h; (8) alterations in surgical approaches; (9) occurrence of life-threatening intraoperative complications; (10) patients planned for fast-track extubation.

Randomisation and blinding

Randomization was performed after operating room arrival but prior to anesthesia induction. Using SPSS (Version 26.0; IBM Corp.), a third-party independent statistician generated a stratified block randomization sequence via random number tables, allocating participants 1:1 to VSLT + BB or VDLT groups. This sequence was implemented through sequentially numbered, opaque sealed envelopes to maintain allocation concealment until intervention. While performing, anesthesiologists were unavoidably unblinded due to device-specific techniques, outcome assessors and follow-up personnel remained blinded to group assignments throughout data collection, with patients similarly blinded during the study period.

Anesthesia methods and surgical procedure

All intubations were performed with patients in the supine position. In the VSLT + BB group, video-imaging single-lumen tracheal tubes (Zhejiang Baiyi Medical Technology Co., Ltd., China) were used with inner diameters (ID) of 7.5 mm (outer diameter [OD] 11.3 mm) for males and 7.0 mm (OD 11.5 mm) for females. The endoscopic camera was integrated at the distal tip of the video-imaging tracheal tube, providing unobstructed anterograde visualization of the tracheobronchial anatomy. Following glottic visualization using a video laryngoscope (Zhejiang UE Medical Equipment Co., Ltd., China), the endotracheal tube was advanced until its distal tip was positioned approximately 3–4 cm superior to the carina. The bronchial blocker (Jiangsu Kechuang Medical Supplies Co., Ltd., China) was then directed into the left main bronchus using the tube’s built-in camera for direct visualization. Optimal placement was confirmed when the upper edge of the inflated blocker cuff rested about 0.5 cm over the carinal. The technical configuration of airway management devices employed in this trial is detailed in Fig. 1.

Fig. 1.

Airway management tools for OLV. (a) The video laryngoscope used in this study. (b) The video-imaging single-lumen tracheal tube and bronchial blocker with their monitor. (c) VSLT + BB, the working status of a video-imaging single-lumen tracheal tube combined with a bronchial blocker. (d) VDLT, the working status of a visual double-lumen endotracheal tube and its monitor

In the VDLT group, left-sided video-imaging double-lumen endotracheal tubes (VDLTs; Guangzhou Welllead Medical Equipment Co., Ltd., China) sized 37 Fr (bronchial lumen outer diameter [OD] 11.5 mm) for males and 35 Fr (bronchial lumen outer diameter [OD] 10.5 mm) for females were selected based on preoperative computed tomography (CT) measurements of tracheal diameter. Within the video-imaging double-lumen endobronchial tube, the micro-camera was embedded in the interluminal septum at the distal terminus, positioned flush with the orifice of the right bronchial lumen to enable simultaneous surveillance of both luminal openings and the carina. The insertion depth was determined using the following formula: Depth (cm) = 12.5 + 0.1 * height (cm) [15]. Under video laryngoscopic guidance, the VDLT was carefully advanced until the bronchial cuff traversed the glottis. Continuous video monitoring facilitated further advancement into the target bronchus, with final positioning confirmed when bronchial cuff inflation aligned its proximal margin with the carina. This process frequently required specialized rotational maneuvers for optimal bronchial lumen alignment, occasionally necessitating controlled withdrawal/re-advancement under real-time imaging feedback. In contrast, VSLT + BB placement permitted direct, atraumatic bronchial blocker advancement through the main lumen under uninterrupted visualization, inherently minimizing carinal contact. Critically, both techniques leveraged continuous anatomical feedback to avoid forcible manipulation throughout placement. All airway management procedures were exclusively performed by one of two dedicated consultant anesthesiologists, each possessing over 5 years’ specialized experience in both DLT and BB techniques. This operator standardization protocol minimized inter-operator variability and technical bias throughout the study. In both study groups, bilateral pulmonary auscultation was systematically performed following mechanical ventilation initiation to verify correct tube positioning. Following documentation of tube positioning times, both groups underwent two-lung ventilation with 100% oxygen (FiO₂ =1.0) for 3 min to facilitate atelectasis. The ventilator circuit was then disconnected for 60 s, exposing the airway to atmospheric pressure. Under continuous video-imaging guidance, the bronchial cuff/blocker balloon was inflated was inflated before surgical incision, and lung isolation was achieved. Surgical lung collapse quality was assessed by the operating cardiac surgeon immediately after pleural incision using a standardized 3-point Likert scale (3 = excellent collapse with optimal surgical exposure, 2 = moderate collapse requiring minor adjustments, 1 = poor collapse interfering with surgical access) [16]. The surgeon was blinded to the specific lung isolation device used, as both techniques were concealed beneath sterile drapes and the assessment was made based solely on the surgical field visualization. The procedures were performed by a dedicated team of five cardiac surgeons. Surgical access was achieved through a left mini-thoracotomy in the fifth intercostal space. Under direct vision, the left internal mammary artery (LIMA) was harvested as a pedicled graft, while the great saphenous vein (GSV) was endoscopically harvested from the lower limb. Following systemic heparinization, pericardiotomy was performed. Using a cardiac stabilizer under direct vision, the LIMA was anastomosed to the left anterior descending artery (LAD). A side-biting clamp was then applied to the ascending aorta, and an aortotomy was created using an aortic punch for proximal anastomosis of the GSV. The heart was stabilized to facilitate sequential or individual anastomoses of the GSV to target vessels (obtuse marginal branch, posterior descending artery, or other coronary arteries), achieving complete revascularization as clinically indicated [17]. Throughout the procedure, patients were maintained in the supine position with operating table adjustments (including Trendelenburg, reverse Trendelenburg, or lateral tilt) as required to optimize surgical exposure. The positions of the VDLT and bronchial blocker were continuously monitored in real time using the integrated video display, and any device displacement was promptly corrected and recorded.

Following the completion of surgical hemostasis, standardized bilateral lung ventilation was initiated. For patients managed with VSLT combined with bronchial blocker (VSLT + BB), the bronchial blocker was removed before transfer to the Intensive Care Unit (ICU) without exchanging the endotracheal tube. In contrast, patients receiving VDLT underwent protocol-directed exchange to single-lumen endotracheal tubes under video laryngoscopic visualization prior to ICU admission. BBs adjustment and tube exchange were performed only after heparin reversal confirmed by normalized ACT values. All study participants underwent extubation in the intensive care unit setting [18].

All candidates for elective MIDCAB underwent protocol-directed pre-anesthesia evaluations 24h preoperatively, incorporating systematic anesthetic strategy reviews, evidence-based risk stratification, and multimodal analgesic planning. Standardized informed consent procedures with biometric verification were conducted in a dedicated preoperative suite, adhering to institutional ethics protocols while preserving patient autonomy. Intubation-related procedures were standardized and performed exclusively by designated anesthesiologists to minimize operator variability. Preoperative management included 8-hour fasting, fluid restriction, and structured inspiratory muscle training. Invasive arterial monitoring was established via ultrasound-guided radial artery cannulation before anesthesia induction, followed by ultrasound-assisted central venous access placement.

A protocolized anesthetic regimen was implemented: induction with sufentanil (≥ 1 µg·kg⁻¹), etomidate (0.1–0.2 mg·kg⁻¹), and cisatracurium (0.15–0.2 mg·kg⁻¹), supplemented by propofol titration as clinically indicated. Maintenance anesthesia consisted of continuous sufentanil infusion, sevoflurane inhalation (0.8–1.2 MAC), and cisatracurium boluses guided by neuromuscular monitoring (Train-of-Four ratio < 0.9). All patients underwent mechanical ventilation in a volume-controlled mode during both two-lung and one-lung ventilation. Airway pressures were measured and obtained directly from the anesthesia machine (WATO EX-65 Pro, Mindray Bio-Medical Electronics Co., Ltd., China) and continuously recorded during the procedures. Mechanical ventilation parameters were standardized using predicted body weight (PBW) calculations: two-lung ventilation (TLV) maintained at FiO₂ 100%, tidal volume 8 mL/kg PBW, respiratory rate 11–13 breaths/min, and inspiratory-to-expiratory (I: E) ratio 1:2; one-lung ventilation (OLV) parameters included reduced tidal volume (6 mL/kg PBW), sustained FiO₂ 100%, elevated respiratory rate (14–18 breaths/min), modified I: E ratio (1:1.5), and 5 cmH₂O positive end-expiratory pressure (PEEP). End-tidal carbon dioxide (PETCO₂) was maintained at 35–45 mmHg through continuous capnography monitoring. Hemodynamic optimization targeted mean arterial pressure ≥ 65 mmHg and heart rate ≤ 100 bpm to mitigate perioperative cardiovascular complications. After surgical hemostasis confirmation and prior to thoracic closure, multimodal analgesia was initiated with ultrasound-guided intercostal nerve blockade. Ventilation strategy transitioned from OLV to protective bilateral lung ventilation (VT 6–8 mL/kg, PEEP 5 cmH₂O). Postoperative pain management employed continuous intercostal nerve block analgesia via patient-controlled infusion pumps, standardized across all study participants. Arterial blood sampling was performed by the attending anesthesiologists involved in this study. All samples were immediately transported to the designated blood gas analyzer by anesthesia nurses not otherwise involved in the research. To maintain consistency, a standardized arterial blood sampler (BD Preset, Becton, Dickinson and Company, UK) was used for all samples. Immediate analysis was performed using the same blood gas analyzer (ABL90 FLEX, Radiometer Medical ApS, Denmark) located in the operating suite, in strict accordance with the manufacturer’s protocols.

Baseline data and postoperative follow-up information were collected in the ward, while other data were obtained in the operating room.

Outcomes

The primary outcome, tube positioning time, was defined as the time interval from the initial insertion of the endotracheal tube under video laryngoscope guidance to the achievement of left lung isolation, confirmed by video images provided through the visible tube and bilateral lung auscultation. Secondary outcomes comprised four domains: (1) Intubation-related metrics including total intubation time (group-specific definitions: VSLT + BB = positioning time; VDLT = positioning time + postoperative exchange time), number of insertion attempts (≥ 3 attempts constituting failed intubation), and intraoperative repositioning incidence triggered by predefined criteria: sustained hypoxemia (SpO₂ ≤ 92% for > 60 s during OLV); plateau pressure > 30 cmH₂O; video-assisted detection of > 1 cm tube migration; or clinician-determined necessity; (2) Respiratory parameters encompassing surgical lung collapse scoring (3-point Likert scale: 3 = excellent, 2 = moderate, 1 = poor) assessed immediately post-pleural incision, dynamic airway pressures, arterial blood gas values at 10/30-min OLV intervals, and hypoxemic episodes (SpO₂ ≤ 92%); (3) Hemodynamic variables (HR, SBP, MAP) recorded at standardized phases: T₀(pre-induction), T₁(post-induction/pre-intubation), T₂(post-intubation), T₃(post-positioning), T₄(5-min stabilization); (4) Postoperative outcomes incorporating vocal morbidity (hoarseness/sore throat) at 4/24/48 h post-extubation, 48-hour pulmonary complications (PPCs: pneumonia, effusion, atelectasis, respiratory failure, bronchospasm, aspiration, pneumothorax), and QoR-15 recovery scores at 24/48 h, with all outcomes adjudicated per CONSORT-endorsed definitions [19–22].

Statistical analyses and sample size calculation

The sample size was recalculated using PASS version 15.0 (NCSS, LLC) based on updated clinical data demonstrating mean tracheal tube positioning durations of 128 ± 54 s for double-lumen endotracheal tubes versus 168 ± 70 s for single-lumen endotracheal tubes with bronchial blockers [23]. With a two-tailed α = 0.05, statistical power = 80%, and an anticipated attrition rate of 20%, the required sample size was 98 participants (49 per group). Ultimately, 102 consecutive patients undergoing MIDCAB under general anesthesia were actually enrolled at Peking University Third Hospital between June 2023 and April 2024, meeting the sample size requirement.

Statistical analyses were performed using SPSS (Version 26.0; IBM Corp.) with normality of continuous variables assessed via the Shapiro-Wilk test (W statistic) and homogeneity of variances evaluated using Levene’s test (F statistic). Parametric data are presented as mean ± standard deviation (SD) and analyzed using independent Student t-tests. Nonparametric data are reported as median (interquartile range, [IQR]) and compared with the Mann-Whitney U test. Categorical variables are expressed as absolute frequencies (proportions, %) and analyzed with Pearson’s χ² test or Fisher’s exact test, the latter applied when expected cell frequencies fell below five. All statistical tests were two-tailed, with significance set at α = 0.05 unless otherwise adjusted.

Results

Baseline characteristics

Between June 14, 2023, and April 30, 2024, a total of 102 consecutive patients (aged 33–79 years, ASA physical status III–IV) scheduled for minimally invasive cardiac surgery (MICS) coronary artery bypass grafting (CABG) under general anesthesia were prospectively enrolled in this study. Four cases were excluded due to intraoperative protocol deviations (conversion to alternative surgical approaches post-induction), with one additional exclusion for 30-day postoperative mortality. Consequently, 97 patients completed the study protocol (Fig. 2: CONSORT flow diagram). Comparative analysis revealed no statistically significant differences in demographic parameters, comorbidities, or preoperative hemodynamic indices between the VDLT and VSLT + BB cohorts (all p > 0.05). Furthermore, intergroup comparisons of induction agents (sufentanil, etomidate, cisatracurium) demonstrated comparable dosing regimens without clinical or statistical differences (detailed in Table 1).

Fig. 2.

CONSORT flow diagram. VDLT, visual double-lumen tube. VSLT + BB, Video-imaging single-lumen tracheal tube combined with bronchial blocker

Table 1.

Baseline characteristics, preoperative data and anesthesia induction drugs of patients in VDLT and VSLT + BB groups

	VDLT group (N=49)	VSLT + BB group (N=48)	p Value
Age, mean (SD), years	66.3±8.7	64.8±8.8	0.408
Gender, n (%)			0.888
Male	32(65.3)	32(66.7)
Female	17(34.7)	16(33.3)
BMI, mean (SD), kg.m−2	25.0±3.5	25.2±3.2	0.679
ASA physical status, n (%)			0.663
Ⅲ	46(93.9)	46(95.8)
Ⅳ	3(6.1)	2(4.2)
Mallampati score, n (%)			0.943
1	16(32.7)	16(33.3)
2	33(67.3)	32(66.7)
Maximal mouth opening, mean (SD), cm	4.3±0.5	4.3±0.5	0.514
Neck circumference, mean (SD), cm	39.1±3.7	38.8±3.9	0.699
Thyromental distance, mean (SD), cm	7.8±0.8	7.9±1.1	0.765
Preoperative HGB, mean (SD), g·L−1	132.4±18.0	131.6±17.2	0.814
LVEF, median (IQR)	68.0(57.0–70.0)	60.0(52.0–69.0)	0.058
PaO2, mean (SD), mmHg	78.7±10.5	80.9±10.5	0.319
PaCO2median (IQR), mmHg	38.0(35.7–40.0)	38.8(36.6–41.8)	0.066
Duration of operation, mean (SD), min	247.4±47.9	246.3±50.4	0.909
Duration of anesthesia, mean (SD), min	338.7±51.5	330.0±46.2	0.383
Preoperative pulmonary function index
FVC, mean (SD), L	3.0±0.8	3.0±0.7	0.993
FEV1, mean (SD), L	2.4±0.6	2.3±0.6	0.381
FEV1/FVC<70%, n (%)	6(12.0)	7(14.6)	0.735
Anesthesia induction drugs
Sufentanil, mean (SD), μg	79.0±23.4	80.5±22.8	0.760
Etomidate, mean (SD), mg	10.2±3.5	11.1±3.5	0.186
Propofol, median (IQR), mg	50.0(30.0–110.0)	50.0(50.0–78.8)	0.355
Cisatracurium, mean (SD), mg	15.4±5.7	15.6±5.9	0.896
Norepinephrine, median (IQR), μg	0.0(0.0-0.0)	0.0(0.0–8.0)	0.071
Phenylephedrine, median (IQR), μg	0.0(0.0-0.0)	0.0(0.0-0.0)	0.958
Ephedrine, median (IQR), mg	0.0(0.0-0.0)	0.0(0.0-0.0)	0.440

Abbreviations: BMI body mass index (calculated as weight in kilograms divided by height in meters squared), ASA American Society of Anesthesiologists, HGB hemoglobin, IQR interquartile range, LVEF left ventricular ejection fraction, FVC forced vital capacity, FEV1 forced expiratory volume in 1 s, FEV1/FVC ratio of forced expiratory volume in 1 s to forced vital capacity

Mean outcomes

Table 2 compares endotracheal intubation outcomes between the VDLT and VSLT + BB groups. Tube positioning time, the primary outcome, was significantly longer in the VSLT + BB group than in the VDLT group. In contrast, total intubation time (positioning time plus tube exchange time in the VDLT group) was shorter in the VSLT + BB group. No significant between-group differences were observed in intubation attempts, lung collapse scores, or repositioning incidence.

Table 2.

Endotracheal intubation variables in VDLT and VSLT + BB groups

Outcomes	VDLT group (N = 49)	VSLT + BB group (N = 48)	p Value	Mean Difference/OR (95% CI)
Tube positioning time, mean (SD), s	128 ± 37	159 ± 58	< 0.001	−32(−52, −12) #
Total intubation time, mean (SD), s	192 ± 40	159 ± 58	< 0.001	32(12, 53) #
Intubation attempts, mean (SD)	1 ± 0	1 ± 0	0.550	−0.02(−0.09,0.04) #
Lung collapse scale, mean (SD)	3 ± 1	3 ± 0	0.472	−0.06(−0.22, 0.10) #
Need of tube repositioning during OLV, n (%)	6(12)	5(10)	0.776	0.83(0.24, 2.93) *

Abbreviations: OLV one-lung ventilation; OR odds ratios, CI confidence intervals

^# mean difference with 95% CI, ^* odds ratio with 95% CI

Secondary outcomes

At 10 min of one-lung ventilation, the VSLT + BB group demonstrated significantly higher intraoperative PaO₂ levels compared to the VDLT group. Concurrently, the VSLT + BB group exhibited a lower alveolar-arterial oxygen gradient. No statistically significant intergroup differences were observed in airway pressures or blood gas parameters at 30 min of one-lung ventilation. No significant between-group difference was found in the incidence of intraoperative hypoxemia (Table 3).

Table 3.

One-Lung ventilation variables in VDLT and VSLT + BB groups

Outcomes	VDLT group (N = 49)	VSLT + BB group (N = 48)	p Value
10 min
PIP, mean (SD), cmH2O	24 ± 3	22 ± 3	0.019
PP, mean (SD), cmH2O	25 ± 3	23 ± 3	0.019
PaO2, median (IQR), mmHg	168 (96–214)	226 (104–333)	0.035
PaCO2, mean (SD), mmHg	40 ± 6	42 ± 5	0.103
A-aDO2, mean (SD)	486.5 ± 96.8	431.0 ± 124.5	0.016
30 min
PIP, mean (SD), cmH2O	23 ± 3	22 ± 3	0.091
PP, mean (SD), cmH2O	24 ± 3	23 ± 3	0.060
PaO2, median (IQR), mmHg	138 (100–222)	185 (104–254)	0.332
PaCO2, mean (SD), mmHg	43 ± 6	43 ± 5	0.698
A-aDO2 mean (SD)	492 ± 89	460 ± 109	0.115
Intraoperative SpO2 ≤ 92%, n (%)	2(4)	1(2)	0.570

Abbreviations: PP peak pressure, PIP plateau inspiratory pressure, A-aDO₂ = (PB - PH₂O) × FiO₂ - PaCO₂/R - PaO₂, PB (barometric pressure) = 760 mmHg, PH₂O (water vapor pressure at room temperature) = 47 mmHg, R (respiratory quotient) = 0.8

No significant differences in MAP, HR, or SBP were observed between the VDLT and VSLT + BB groups at baseline (T₀) or post-induction (T₁) (p > 0.05). Comparative analysis of changes from T₀ to T₁ (ΔT₀–T₁) during T₂–T₄ revealed no statistically significant differences; however, transient numerical trends suggested marginally greater attenuation in hemodynamic fluctuations in the VDLT group. Both groups exhibited equivalent perioperative hemodynamic stability, with no clinically significant deviations in cardiovascular parameters (Table 4).

Table 4.

Hemodynamic changes from baseline (T₀ and T₁) in VDLT and VSLT + BB groups

Outcomes	VDLT group (N = 49)	VSLT + BB group (N = 48)	p Value
MAP T0, mean (SD), mmHg	89 ± 10	92 ± 10	0.117
HR T0, mean (SD)	70 ± 10	68 ± 10	0.301
SBP T0, mean (SD), mmHg	126 ± 15	129 ± 16	0.244
Delta value from T 0
MAP, mmHg
T2, mean (SD)	15 ± 10	14 ± 9	0.602
T3, median (IQR)	14 (9–31)	15(7–26)	0.636
T4, median (IQR)	14 (9–20)	11(6–20)	0.164
HR
T2, median (IQR)	11 (7–18)	11(7–20)	0.994
T3, median (IQR)	12 (5–27)	12(4–20)	0.386
T4, mean (SD)	14 ± 10	11 ± 8	0.090
SBP, mmHg
T2, median (IQR)	16 (7–31)	10 (3–22)	0.165
T3, median (IQR)	13 (5–16)	10(2–14)	0.134
T4, median (IQR)	10(4–21)	9(4–15)	0.278
MAP T1, mean (SD), mmHg	78 ± 13	82 ± 13	0.149
HR T1, mean (SD)	57 ± 10	57 ± 9	0.951
SBP T1, mean (SD), mmHg	120 ± 21	124 ± 24	0.357
Delta value from T 1
MAP, mmHg
T2, median (IQR)	5(2–9)	5(1–8)	0.344
T3, median (IQR)	8(3–13)	6(2–13)	0.366
T4, median (IQR)	9(4–15)	7(3–14)	0.366
HR
T2, median (IQR)	1(1–5)	2(1–4)	0.321
T3, median (IQR))	5(3–9)	6.5(2.0–10.0)	0.565
T4, median (IQR)	3(2–5)	2(1–4)	0.178
SBP, mmHg
T2, median (IQR), mmHg	8(3–16)	6(2–13)	0.236
T3, median (IQR), mmHg	13(7–26)	9(6–21)	0.388
T4, median (IQR), mmHg	15(10–25)	12(6–25)	0.280

Abbreviations: MAP mean arterial pressure, HR heart rate, SBP systolic blood pressure

Post-extubation outcomes demonstrated significant differences between the VDLT and VSLT + BB groups. Patients in the VDLT group experienced a significantly higher incidence of sore throat at 4 h and more frequent hoarseness. In contrast, the rates of postoperative pulmonary complications and the QoR-15 scores at 24 and 48 h were comparable between groups (Table 5).

Table 5.

Post-extubation outcomes in VDLT and VSLT + BB groups

Outcomes	VDLT group (N = 49)	VSLT + BB group (N = 48)	p Value	Mean Difference/OR (95% CI)
4 h after extubation
Sore throat, n (%)	16(33)	6(13)	0.018	2.84(1.04, 7.71) *
24 h after extubation
Sore throat, n (%)	12(25)	6(13)	0.129	2.79(0.85, 6.45) *
QoR-15 score, mean (SD)	105(14)	108(13)	0.266	−3.4(−8.9, 2.1) #
48 h after extubation
Sore throat, n (%)	11(22)	5(10)	0.110	2.27(0.78, 6.65) *
Hoarseness, n (%)	27(55)	15(31)	0.018	2.45(1.08, 5.59) *
PPCs, n (%)	1(2)	2(4)	0.545	0.19(0.01, 4.04) *
QoR-15 score, mean (SD)	112 ± 12	115 ± 13	0.274	−2.9(−7.8, 2.0) #

Abbreviations: QoR-15 score quality of recovery-15 score

^# mean difference with 95% CI, ^* odds ratio with 95% CI

Discussion

Comparison of tube positioning time and airway safety

Our results indicated that tube positioning was procedurally more efficient in the VDLT group than in the VSLT + BB group. This observation is consistent with the findings of Zhang et al., who reported shorter placement times for double-lumen tubes compared with bronchial blockers (125 ± 60 s vs. 210 ± 120 s, p = 0.02) under bronchoscopic guidance [12]. The relatively faster placement of VDLTs can be attributed to the elimination of bronchoscopy-dependent steps required for bronchial blocker positioning in single-lumen or video-imaging single-lumen tubes [13]. It is worth noting that the VSLT + BB group in our study still achieved shorter positioning times than those reported in historical cohorts using conventional SLT + BB techniques, likely due to reduced reliance on bronchoscopy and improved maneuverability afforded by the larger lumen diameter of the VSLT.

In our institutional setting, patients undergoing minimally invasive cardiac surgery with a DLT are routinely exchanged to a single-lumen tube, while the bronchial blocker is removed in the BB group before ICU transfer to facilitate postoperative airway management, consistent with the approach described by Miranda Holmes et al. [24]. Although the initial tube positioning was quicker in the VDLT group, this modest temporal advantage was offset by the necessity of tube exchange at the conclusion of surgery. The resulting total intubation time—defined as the duration from laryngoscopy to final tube exchange—showed a statistically significant but clinically marginal prolongation in the VDLT group. More importantly, the VDLT group experienced significantly higher incidence of postoperative sore throat and hoarseness compared to the VSLT + BB group. These complications may result from multiple factors including the larger outer diameter of the VDLT device, the initial intubation trauma, and the additional airway manipulation during mandatory tube exchange. Evidence from thoracic surgery, where DLTs are used without postoperative tube exchange, similarly demonstrates higher airway complications with DLTs compared to BBs, suggesting that device characteristics contribute substantially to these adverse effects beyond the tube exchange procedure alone.

Therefore, the selection between these techniques should not be primarily based on procedural speed but rather on a comprehensive risk-benefit assessment that weighs the initial operational ease of video double-lumen tubes (VDLT) against the potential hazards associated with tube exchange. It is noteworthy that although the reoperation rate in our single-center MIDCAB cohort was 1.7% [17], the VSLT + BB technique may be particularly advantageous in such scenarios requiring rapid intervention and minimal tissue trauma. In our institutional setting, we have managed cases where patients initially managed with a DLT required reoperation. In such scenarios, patients must undergo re-intubation with a DLT for the second surgery, followed by another tube exchange to a single-lumen tube postoperatively. This effectively constitutes two separate surgical procedures involving four distinct intubation events, based on our actual clinical experience. In contrast, the VSLT + BB approach only requires repositioning of the bronchial blocker through the existing single-lumen tube. Some cardiac surgery centers may perform immediate extubation at the end of the operation in patients managed with either a double-lumen tube (DLT) or a bronchial blocker (BB), thereby avoiding the need for tube exchange. Although this strategy can also be implemented at our institution, it is subject to strict eligibility criteria. In the present study, all enrolled patients had multivessel coronary artery disease and were at higher surgical and anesthetic risk. According to our institutional management protocol, such patients are routinely transferred to the intensive care unit (ICU) with an endotracheal tube in place for postoperative recovery. Therefore, to facilitate postoperative airway management, patients intubated with a DLT were switched to a single-lumen endotracheal tube before leaving the operating room [18].

Quality of lung collapse and ventilation

Our study revealed comparable lung collapse scores between the VSLT + BB and VDLT groups, consistent with findings from Palaczynski et al.‘s meta-analysis of 25 trials (n = 1,636), which demonstrated equivalent lung collapse quality between DLTs and BBs during one-lung ventilation (OLV) [10]. Ender et al. similarly reported no significant difference in rates of inadequate lung collapse between the two techniques (DLT: 2% vs. BB: 2%, p > 0.05). In contrast, Zhang et al. observed superior lung collapse scores with BBs at 5 and 10 min during thoracoscopic esophagectomies, a finding potentially attributable to differences in surgical field exposure between thoracic and cardiac minimally invasive procedures [12]. This discrepancy is, however, tempered by evidence from other thoracoscopic studies indicating equivalent collapse quality between SLT + BB and DLT after pleural opening, supporting the view that both VSLT + BB and VDLT provide effective OLV in cardiac surgery [25]. Repositioning rates did not differ significantly between groups (12% vs. 10%, p = 0.776), a result that contrasts with the lower displacement rates reported by Ender et al. (DLT: 8% vs. BB: 2%) [13]. This variation may be explained by methodological improvements in our protocol, wherein continuous videoscopic monitoring allowed for the detection of subtle tube displacements that could be missed by conventional methods relying solely on hypoxemia or clinical signs.

Airway pressures and oxygenation during OLV

At 10 min after OLV initiation, the VDLT group exhibited significantly higher peak and plateau airway pressures than the VSLT + BB group (p < 0.05 for both), aligning with reports of elevated inspiratory pressures in conventional DLT cohorts. The lower pressures recorded in the VSLT + BB group are likely attributable to incomplete lung isolation due to anatomical microleaks around the bronchial blocker—an inherent characteristic of the technique that may allow partial ventilation of the non-ventilated lung. It is important to acknowledge that this physiological distinction represents a potential source of treatment bias, as the two techniques created fundamentally different ventilation conditions. The less complete isolation inherent to the BB technique, while appearing beneficial for certain parameters, constitutes a fundamental difference in the physiological environment between groups, which should be considered when interpreting comparative outcome.

This microleak phenomenon may explain the concurrent elevation in PaO₂ observed in the VSLT + BB group, as partial ventilation of the non-dependent lung through gas leakage around the blocker cuff improves the overall V/Q ratio. However, minute ventilation during one-lung ventilation was not recorded in this study, which could have provided direct supporting evidence for this interpretation as well as an explanation for the associated PaCO₂ levels. An alternative explanation may involve the narrower lumen profile of the BB, which could reduce alveolar oxygen washout during lung collapse, thereby enhancing oxygen diffusion efficiency compared to the wider lumen geometry of the VDLT. Together, these two mechanisms—leak-mediated partial ventilation and altered gas flow dynamics—may account for the early oxygenation advantage observed in the VSLT + BB cohort. Although this finding diverges from some reports that showed nonsignificant PaO₂ variations between conventional DLT and BB groups, it is consistent with thoracoscopic studies documenting improved oxygenation during BB-mediated OLV at lower airway pressures [26]. Although the intergroup pressure differences equilibrated by 30 min with ongoing surgical manipulation, the observed disparities in both airway pressures and PaO₂ at this timepoint failed to reach statistical significance (p = 0.063 for both). This non-significance may reflect a Type II error, wherein a smaller but clinically meaningful effect existed yet remained undetected due to the limited sample size of our study population.

Notably, the incidence of hypoxemia did not differ significantly between groups (4% vs. 2%, p = 0.570), which contrasts with the findings of Palaczynski et al., who reported higher rates of hypoxemia with DLT versus BB (13.5% vs. 6.0%, p = 0.02) [10]. This lower overall incidence in our cohort may reflect protocol-specific optimizations in MIDCAB, particularly the use of VSLT for real-time visualization and adjustment of airway device position.

Hemodynamic stability and postoperative recovery

Lu et al. observed significantly elevated SBP, DBP, and HR in left-sided DLT groups compared to BB groups during right-sided video-assisted thoracoscopic surgery at 1-minute and immediate post-intubation intervals (p < 0.05), contrasting with our observations [11]. This divergence may originate from anatomical distinctions: Lu et al.’s protocol leveraged the right mainstem bronchus’s acute bifurcation angle (~ 25°) for BB placement, whereas our left bronchial BB positioning necessitated rotational adjustments, increasing carinal stimulation risk.

Zheng et al. demonstrated that BBs for OLV in video-assisted thoracic surgery induced fewer hemodynamic perturbations compared with alternative techniques [26]. In our study, hemodynamic parameters (HR, MAP, SBP) at T₂–T₄ (relative to baseline T₀–T₁) showed no significant intergroup differences (p > 0.05), suggesting that cardiovascular stress with both intubation methods. This equivalence may reflect the shared use of visual tube technologies, which reduced intraprocedural tracheal manipulation. In contrast, Zheng et al. utilized conventional DLTs characterized by larger luminal diameters and extended profiles. This technical approach merits particular consideration given that thoracic surgical procedures typically involve reduced opioid administration compared with cardiac operations—a critical distinction that may potentiate hemodynamic instability during conventional endotracheal device placement.

No significant difference in PPCs incidence was observed between the VSLT + BB and VDLT groups, both rates being substantially lower than historical benchmarks. This favorable outcome may relate to our institutional preoperative inspiratory muscle training protocol, proven to reduce PPC risk [27]. Emerging evidence associates minimally invasive approaches like MIDCAB with enhanced pulmonary recovery, particularly beneficial for COPD patients [28]. All MIDCAB procedures in this cohort were performed off-pump, thereby circumventing pulmonary injury from extracorporeal circulation. Collectively, these protocol-specific factors likely contributed to the observed low PPCs rates.

The QoR-15 questionnaire assesses five patient-reported outcome domains—pain, physical comfort, physical independence, psychological support, and emotional state—via multiple subdomains to quantify postoperative recovery [22]. In this study, the sole intergroup variable between the VSLT + BB and VDLT cohorts was the airway management technique. Given the multifactorial nature of postoperative recovery influenced by diverse perioperative variables, isolated differences in intubation methodology were insufficient to significantly influence composite QoR-15 outcomes. Consequently, no statistically significant intergroup differences in QoR-15 scores were observed at 24 h (p = 0.63) or 48 h (p = 0.58) post-extubation.

This study has several limitations. First, owing to the nature of the interventions, a double-blind design was not feasible. However, this was a single-blind study, as the patients and all postoperative outcome assessors were blinded to the group allocation to minimize assessment bias. Second, we investigated multiple secondary outcomes with repeated measurements at different time points during OLV, which increases the risk of Type II error and may limit the reliability of interpreting non-significant findings for individual secondary endpoints. Third, although hemodynamic fluctuations were observed during tube positioning, our study had notable limitations in physiological monitoring. We were unable to perform comparative quantification of neuroendocrine stress responses due to the absence of serial stress biomarker measurements (serum cortisol or catecholamine levels) [29, 30]. Additionally, hemodynamic parameters during the postoperative VDLT-to-SLT conversion—a procedurally intensive period with potential for cardiovascular instability due to repeated airway manipulation—were not systematically recorded, precluding comprehensive assessment of the physiological impact of tube exchange. Finally, the clinical relevance of the ‘total intubation time,’ while statistically significant, may be limited as the absolute difference was small. The more critical consideration is the clinical risk associated with the tube exchange procedure itself, which this metric encapsulates.

Conclusion

In minimally invasive direct coronary artery bypass (MIDCAB) surgery, video-imaging double-lumen endobronchial tubes (VDLT) demonstrated significantly shorter tube positioning time compared to video-imaging single-lumen tracheal tubes with bronchial blockers (VSLT + BB). However, the VSLT + BB technique exhibited shorter total intubation time than VDLT by avoiding postoperative tube exchange. Both techniques provided clinically acceptable lung isolation with comparable lung collapse quality and incidence of hypoxemia. The VSLT + BB group exhibited lower airway pressures and higher oxygenation indices during the early phase of one-lung ventilation, whereas the VDLT group was associated with a higher incidence of minor postoperative airway complications, including sore throat and hoarseness. No significant differences were observed between groups in pulmonary complications, hemodynamic stability, or postoperative recovery quality.

Authors’ contributions

YF and MW are co-first authors who designed the study, drafted the manuscript, and analyzed and interpreted the data. YF prepared Figs. 1 and 2. YZ and YH revised the manuscript and approved the final version. All other authors were investigators involved in protocol review, study execution, and manuscript review. All authors approved the final version of manuscript to be published.

Funding

This study was funded by grants from Capital’s Funds for Health Improvement and Research (2024-2-40912), Wu Jieping Medical Foundation (320.6750.2023-08-5), Innovation and Transformation Project of Peking University Third Hospital (BYSYZHKC2022103).

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This single-center, randomized, controlled study got approved & registered by the Chinese Clinical Trial Registry (ChiCTR 2300072124, chictr.org.cn, 3/6/2023), passed the Ethics Committee review of Peking University Third Hospital (No. IRB00006761-M2023282), and followed the Declaration of Helsinki. Written informed consent was obtained from all patients or their legal representatives.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.