Intraoperative extracorporeal life support in tracheal surgery: a 20-year single-center experience

Abstract

Background

Despite advancements in perioperative management, patients undergoing complex tracheal surgery still face increased risks of morbidity and mortality. Extracorporeal life support (ECLS) can be utilized to maintain hemodynamic and respiratory stability during surgical procedures, but there is currently no evidence-based practice protocol in place. This study seeks to evaluate the safety and benefits of various ECLS modes in critical tracheal surgery and establish a practical protocol for the application of ECLS.

Methods

Single center, retrospective, cohort study. This retrospective study collected data on patients who underwent tracheal procedures on ECLS between January 2000 and December 2020. Patients were divided into three groups according to configuration: cardiopulmonary bypass (CPB), modified CPB and extracorporeal membrane oxygenation (ECMO). Baseline characteristics, in-hospital complications, ECLS details, adverse events, and overall survival were assessed and compared between three groups.

Results

Over twenty years period, from 462 primary tracheal disease patients, 67 patients were identified from our institutional medical registry as having received CPB, modified-CPB or ECMO during the procedure (26.9%, 22.4% and 50.7% respectively). Heparin free (n = 30, 44.8%) or low dose heparin (n = 17, 26.6%) therapy was frequently utilized in ECMO and modified-CPB group. The Median pump duration of CPB, modified-CPB, VV, and VA ECMO was 43, 64, 139, and 127 min. Anastomotic complications were significantly lower in the VV-ECMO group (7.1%), while the duration of mechanical ventilation was significantly longer in the VA-ECMO group (66, [2–178]). Surgical bleeding was significantly higher in the CPB group (460 ± 73 ml). More patients in the CPB group (9/18) required blood transfusions; however, there was no significant difference in the volume of packed red blood cells between the groups.

Conclusion

Despite intraoperative ECLS being rarely used in tracheal and carinal disease, programmed extracorporeal life support with VV-ECMO demonstrated high effectiveness with an acceptable mortality and morbidity rate in high-risk tracheal surgery.

Background

The prevalence of critical tracheal disease is low, and advancements in perioperative management, anesthesiology, and surgical techniques have played a crucial role in enhancing patient outcomes. However, patients undergoing complex tracheal surgery face elevated risks of morbidity and mortality [1]. In cases of central airway neoplasms, there is a potential for severe airway obstruction, leading to difficulties in ventilation or intubation. This condition carries a risk of acute respiratory decompensation throughout the stages of anesthesia and the procedure [2, 3]. Extracorporeal life support(ECLS) can be administered through either cardiopulmonary bypass (CPB) or extracorporeal membrane oxygenation (ECMO) to uphold hemodynamic or respiratory stability during surgical procedures [2, 4]. CPB has been the traditional approach primarily utilized in the initial stages of tracheal surgery, whereas ECMO was introduced at a later stage [4, 5]. ECMO is a prolonged form of cardiopulmonary bypass that is commonly utilized in critically ill patients experiencing refractory cardiorespiratory failure as a vital life-saving support method [6]. While numerous case reports have detailed the utilization of the ECLS technique, which includes CPB and ECMO, in patients with central airway or carinal tumors, there is a lack of standardized protocols to guide perioperative procedures in such cases [7–9]. Given the scarcity of evidence-based practice guidelines, this study aimed to investigate the indications, complications, and outcomes of ECLS use in tracheal disease patients treated at the largest medical center for thoracic surgery in China over the past 20 years. In this research, we aim to conduct a retrospective assessment of the utilization of CPB, modified-CPB, and ECMO to explore the utility of ECLS in severe airway obstruction and to assess the safety and benefits of different ECLS modes in critical tracheal surgery by anastomotic complications, anticoagulant strength, surgical bleeding and other postoperative complications.

Methods

Study population

The retrospective study was carried out at a university-affiliated hospital that serves a population of more than 20,000 thoracic cases annually with an ECMO clinical service providing extracorporeal support for refractory respiratory or circulatory failure for nearly two decades. The local ethical review board approved the protocol (approval no. KSY28713). The medical data were collected while ensuring the protection of personal information and privacy, and obtaining informed consent was not deemed necessary. The analysis included all patients with critical tracheal disease who received treatment with extracorporeal life support between January 1, 2000, and December 31, 2020. Patients with tracheal stenosis caused by mediastinal tumor compression are excluded. A retrospective chart review of medical records was undertaken and intraoperative ECLS mode, blood flow, sweep gas flow, arterial blood gas, anticoagulation strategy, complications, and other perioperative data were recorded.

Preoperative assessment

During the preoperative multidisciplinary risk assessment, the decision to utilize ECLS was based on an assessment of the risks entailed in tracheal intubation, ease of ventilation, surgical hemorrhage risks, and hemodynamic stability. All patients underwent preoperative chest X-ray and computed tomography (CT) to evaluate the aspects such as the lumen diameter at the narrowest part of the stenotic segment, the lumen diameter of the normal-width segment nearest to the stenotic segment, and the length of the stenotic segment. Most patients underwent flexible (and/or rigid) fiberoptic bronchoscopy (FOB) for careful evaluation of the extent of tumor involvement and quality of tissue. Given the potential risk of tracheal or tumorous rupture during FOB evaluation, the insertion of a flexible fiber-optic bronchoscope was not feasible in certain patients with near occlusion of the cervical trachea.

Anesthesia

A standard monitor during anesthesia (continuous pulse oximetry, noninvasive arterial blood pressure, heart rate, and electrocardiography) was set up before induction of anesthesia. Central venous catheterization and radial artery catheterization were conducted before induction. The patient received preoxygenation with 100% oxygen via a facemask, while maintaining spontaneous ventilation.

We applied ECLS under local anesthesia and mild sedation prior to induction in patients with a high risk of airway complications. For patients who could be safely intubated, tracheal intubation utilizing an oral tube was guided by bronchoscopy to prevent tracheal damage and tumor rupture. General anesthesia was induced with fentanyl (10–30 μg/kg) and supplemented with sevoflurane and/or propofol. Subsequently, anesthesia was maintained by TCI of propofol 2 mcg/mL plus remifentanil 2 to 5 ng/mL, with a BIS of 30 to 40. The tracheal anastomosis was assessed using bronchoscopy, and any blood clots or secretions were meticulously cleared upon completion of the procedure. Extubation was carried out quickly to minimise the strain on the sutures caused by positive-pressure ventilation.

Intraoperative ECLS strategy and management

Different ECLS models were used in different phases: 2000–2007 with CPB, 2008–2012 with modified-CPB, ECMO was first used in 2013 and became the most common mode of intraoperative ECLS (Fig. 1).

Fig. 1.

Distribution of different extracorporeal life support configurations by year. CPB: cardiopulmonary bypass; VV-ECMO: veno-venous extracorporeal membrane oxygenation; VA-ECMO: veno-arterial extracorporeal membrane oxygenation

The decision to proceed with ECLS was made by a multidisciplinary team comprising a surgeon, an anesthetist, and a perfusionist. ECLS was assisted with either local anesthesia before induction or after general anesthesia. A cardiac surgeon who was present in the operating theatre before induction proceeded with cannulation of the femoral or neck vessels in a supine position for preemptive ECMO cannulation. Ultrasound-guided Seldinger or open Seldinger technique was adopted for cannulation under the mode of femoral vein drainage and internal jugular vein return in VV-ECMO; while VA-ECMO and CPB, from femoral vein to femoral artery, were used usually when patients need circulatory support. Central cannulation was used for rescue extracorporeal support (VA-ECMO or CPB) when circulatory collapse or surgical bleeding occurs (Fig. 2). ECMO therapy was performed using the Rotaflow® system equipped with a Quadrox® oxygenator (both Getinge, Rastatt, Germany). The application of CPB or modified-CPB used a roller pump (Stockert S5, Livanova, Mirandola, Italia) and centrifugal pump (Bio-medicus 550 Bio-console, Medtronic, Minneapolis, MN, USA) respectively, as described in our previous article [10]. The three ECLS modes are shown in Fig. 2. The linear relationship between ECMO flow and oxygen transfer favors the use of large drainage cannulas (23-25Fr, Medtronic, Minneapolis, USA) to provide adequate oxygenation support.

Fig. 2.

Illustration of different extracorporeal life support and cannula configurations. CPB: cardiopulmonary bypass; m-CPB: modified cardiopulmonary bypass; VV-ECMO: veno-venous extracorporeal membrane oxygenation; VA-ECMO: veno-arterial extracorporeal membrane oxygenation

After ECLS initiation, VA-ECMO or CPB parameters were maintained as follows: Blood flow: 50–70 mL/kg; gas to blood ratio: 0.5–1:1; Fraction of sweep gas oxygen (F_GO₂): 50–70%. Blood flow and sweep flow were adjusted according to pulse oxygen saturation and blood gas results to maintain arterial oxygen tension (PaO₂) of 60–80 mmHg and a partial pressure of carbon dioxide of 35–45 mmHg with F_GO₂ 100% during VV-ECMO. Normothermia management was applied, with the nasopharyngeal and bladder temperature maintained at 36.5 ℃. Peripheral blood gas and pulse oxygen saturation should be routinely monitored through the right radial artery to detect differential hypoxia, while Near-Infrared Spectroscopy (NIRS) could detect cerebral hypoxia in all patients.

Patients in the supine position were transitioned from VA-ECMO or CPB support once a secure airway was established by inserting a sterile endotracheal tube into the distal airway in the field, or a double-lumen endotracheal tube into the bronchus, and ensuring satisfactory mechanical ventilation. VV-ECMO can obviate the need for an endotracheal tube or intermittent ventilation as it provides comprehensive oxygenation support to facilitate proper dissection and anastomosis, and patients were weaned off ECLS at the conclusion of the procedure.

Anticoagulation

Unfractionated heparin (UFH) was used to ensure full heparinization (activated clotting time, ACT > 480 s) during CPB. A singular dose of UFH (< 2500 IU) was administered for cannulation on ECMO, and no additional heparin was given, establishing a heparin-free approach. Administering additional heparin to achieve an ACT of 160–200 s throughout the support period represents a standard-heparin strategy. ACT is measured every 30 min during the procedure. Excessive heparinization is antagonized by the same amount of protamine (1:1) in CPB mode.

Statistical analyses

Normally-distributed continuous variables were expressed as mean ± standard deviation (SD) and compared with t-test. Non-normally distributed continuous variables were expressed as median with interquartile range (IQR) or range, and compared with Wilcoxon rank-sum test. Categorical variables were expressed as percentages (n, %). Odds ratios for postoperative complications were estimated with univariate logistic regression using robust standard errors. Results are reported as odds ratio with 95% confidence intervals (95% CIs). Statistical analysis was performed with SPSS software version 21(IBM Corp, Armonk, NY), and p < 0.05 was statistically significant.

Results

During the twenty years, 462 patients underwent the tracheal procedure, 67 (14.5%) of whom received intraoperative ECLS (18 CPB, 15 modified-CPB, 28 VV-ECMO, and 6 VA-ECMO). After 2013, VV-ECMO was the preferred option for all patients; however, six patients opted for VA-ECMO due to comorbid heart disease, and three patients selected CPB because of the need for reoperation or tumor proximity to major vessels (Fig. 1). Patients’ ages ranged from 14 to 74 years (37 females, 30 males). The most common comorbidities were hypertension and diabetes mellitus. The predominant symptoms were cough (n = 34, 50.7%) and shortness of breath (n = 34, 50.7%), followed by progressive dyspnea (n = 29, 43.3%), wheezing or stridor (n = 23, 33.3%), and hemoptysis (n = 18, 26.9%). The median time interval between the onset of symptoms and the operation was 13.6 months (IQR, 8–22). 60 (88.1%) patients demonstrated tracheal stenosis exceeding 90%. Clinical presentation, patient comorbidities, and diagnoses are depicted in Table 1. The most common surgical approaches were posterolateral thoracotomy (61.2%), and 56.7% (n = 38) of patients received carinal resection, while 6 of patients received carinal pneumonectomies (Table 1).

Table 1.

Baseline and surgical characteristics

Group	Total study population (n = 67)	CPB (n = 18)	m-CPB (n = 15)	VV-ECMO (n = 28)	VA-ECMO (n = 6)	P value
Male gender, n (%)	30 (44.8)	8	7	12	3	0.71
Age (years), median (range)	47.3 (14–74)	44.5 (32–69)	46.7 (47–74)	49.2 (14–65)	51.6 (44–63)	0.53
Weight (kg), median [IQR]	62.3 [43.2–79.7]	58.4 [43.2–71.1]	63.7 [51.8–68.3]	66.5 [54.2–65.8]	63.7 [57.2–79.7]	0.67
BMI (kg/m2) [IQR]	24.4 [19.8–27.9]	25.1 [23.6–26.3]	25.8 [23.7–27.9]	24.6 [19.8–27.6]	25.3 [23.6–27.4]	0.54
FEV1% [IQR]	40.2 [28.3–63.6]	33.6 [28.3–48.6]	35.8 [29.6–52.3]	36.5 [31.6–63.6]	35.9 [28.4–50.7]	0.43
Stenosis rate > 90%	60(88.1)	16(88.9)	13(86.7)	27(96.4)	4(66.7)	0.41
Clinical presentation, n (%)
Cough	34(50.7)	8(44.4)	7(46.7)	15(53.6)	4(66.7)	0.28
Hemoptysis	18(26.9)	7(38.9)	5(33.3)	4(14.3)	2(33.3)	0.49
Hoarse voice	12(17.8)	4(22.2)	2(13.3)	6(21.4)	0(0)	0.53
Wheezing or stridor	23(34.3)	8(44.4)	6(40.0)	7(25.0)	2(33.3)	0.41
Progressive dyspnea	29(43.3)	9(50.0)	7(46.7)	10(35.7)	3(50.0)	0.38
Shortness of breath	34(50.7)	10(55.6)	9(60.0)	11(39.3)	4(66.7)	0.47
Comorbidities, n (%)
Diabetes	11(16.4)	2(11.1)	3(20.0)	4(14.3)	2(33.3)	0.58
Coronary artery disease	6(8.9)	2(11.1)	2(13.3)	1(3.6)	1(16.7)	0.34
Hypertension	26(38.9)	6(33.3)	5(33.3)	12(42.9)	3(50.0)	0.43
Cerebrovascular disease	3(4.5)	1(5.6)	0(0. 0)	2(7.1)	0(0.0)	0.38
COPD	8(11.9)	2(11.1)	3(20.0)	3(10.7)	0(0.0)	0.52
Peripheral vascular disease	4(6.0)	0(0.0)	0(0.0)	3(10.7)	1(16.7)	0.49
Previous surgical history	3(4.5)	2(11.1)	1(6.7)	0(0. 0)	0(0. 0)	0.43
Diagnosis, n (%)
Carcinoma	21(31.3)	5(27.8)	3(20.0)	8(28.6)	5(83.3)	0.62
Carcinoid	13(19.4)	2(11.1)	4(26.7)	7(25.0)	0(0.0)	0.65
Tracheal fistulas	12(17.9)	4(22.2)	3(20.0)	4(14.3)	1(16.7)	0.45
Tracheal stenosis	18(26.9)	5(27.8)	4(26.7)	9(32.1)	0(0.0)	0.46
Others	3(4.5)	2(11.1)	1(6.7)	0(0.0)	0(0.0)	0.51
Surgical approach, n (%)
Median sternotomy	17(25.4)	6(33.3)	3(20.0)	6(21.4)	1(16.7)	0.21
Posterolateral thoracotomy	41(61.2)	11(61.1)	10(66.7)	16(57.1)	5(83.3)	0.32
VATS	9(13.4)	1(5.6)	2(13.3)	6(21.4)	0(0.0)	0.14
Resected site, n (%)
Only primary trachea	29(43.4)	7(38.9)	6(40)	14(50)	2(33.3)	0.49
Carina	38(56.7)	11(61.1)	9(60.0)	14(50.0)	4(66.7)	0.56
Type of resection, n (%)
Carinal resection with neocarina	32(47.8)	8(44.4)	9(60.0)	11(39.3)	4(66.7)	0.58
Carinal pneumonectomies	6(8.9)	3(16.7)	0(0.0)	3(10.7)	0(0.0)	0.49
Circular resection/anastomosis	24(35.8)	6(33.3)	4(26.7)	12(42.9)	2(33.3)	0.64
Other	5(7.5)	1(5.6)	2(13.3)	2(7.1)	0(0.0)	0.34

CPB cardiopulmonary bypass, m-CPB modified cardiopulmonary bypass, VV-ECMO veno-venous extracorporeal membrane oxygenation, VA-ECMO veno-arterial extracorporeal membrane oxygenation, BMI body mass index, FEV₁% percentage of forced expiratory volume in one second, COPD chronic obstructive pulmonary disease, VATS video-assisted thoracic surgery

Postoperative events were occurred in 18 (26.9%) patients, with a particular emphasis on anastomotic complications, which were significantly more prevalent in the CPB group (27.8%). Serious complications included bronchopleural fistula (n = 6, 9.0%) and respiratory failure postoperatively (n = 4, 6.0%). 43 (64.2%) patients were extubated in the operation room, and the ECMO group (71.4%) had a higher percentage. The median postoperative assisted ventilation time was 24 h, and both the CPB and VA-ECMO (32 vs 66 h) groups had a longer intubation time. Nine patients required prolonged mechanical ventilation for pneumonia or bronchopleural fistula. The surgical bleeding (CPB: 460 ± 73 ml), pleural drainage (276 ± 116 ml), and percentage of patients receiving blood (50.0%) tended to be higher in the CPB group, but perioperative transfusion of pRBC appeared no difference. There were 6 (9%) deaths reported in this study following the operation. Two patients developed a bronchopleural fistula and died of pyothorax after 32 and 44 days of VV-ECMO support. One patient died of right heart failure following pulmonary embolism, and three patients died of infection and ARDS leading to multiorgan dysfunction. The mean observed ICU and hospital length of stay remained similar (Table 2). 61 patients survived to hospital discharge without complications and returned to their baseline quality of life. All surviving patients upon discharge had a one-year survival rate of 100%, and none of them experienced readmission due to surgical complications.

Table 2.

Perioperative complications and outcome

Group	Total study population (n = 67)	CPB (n = 18)	mini-CPB (n = 15)	VV-ECMO (n = 28)	VA-ECMO (n = 6)	P value
Survival to discharge, n (%)	61(91.0%)	17(94.4)	14(93.3)	26(92.9)	4(66.6)	0.57
OR extubation, n (%)	43(64.2%)	10(55.6)	10(66.6)	20(71.4)	3(50.0)	0.04
Postoperative morbidity, n (%)	18(26.9%)
Anastomotic complications	11(16.1%)	5(27.8)	3(20.0)	2(7.1)	1(16.7)	0.04
Atelectasis	8(11.9)	2(11.1)	2(13.3)	4(14.3)	0(0.0)	0.47
Respiratory infection	7(10.4)	1(5.6)	2(13.3)	3(10.7)	1(16.7)	0.58
Chylothorax	3(4.5)	2(11.1)	0(0.0)	2(7.1)	0	0.72
Bronchopleural fistula	6(9.0)	3(16.7)	1(6.7)	1(3.6)	1(16.7)	0.63
Respiratory failure	4(6.0)	2(11.1)	0(0.0)	1(3.6)	1(16.7)	0.65
Operation time (min)	119 ± 38	142 ± 59	146 ± 63	133 ± 43	138 ± 28	0.61
Surgical bleeding (ml)	290 ± 126	460 ± 73	220 ± 78	210 ± 55	330 ± 47	0.03
Blood transfusion n (%)	16(23.9)	9(50.0)	4(26.7)	3(10.8)	2(33.3)	< 0.01
pRBC Transfusion (ml)	180 ± 34	280 ± 57	170 ± 43	140 ± 38	110 ± 45	0.43
Pleural drainage (ml)	193 ± 88	276 ± 116	172 ± 51	164 ± 49	130 ± 42	0.02
Ventilation time (hour), median [IQR]	24[1–233]	32[3–233]	17[11–46]	14[1–121]	66 [2–178]	< 0.01
ICU length of stay (hour), median [IQR]	51[0–427]	56[18–320]	27[17–59]	51[0–427]	100[12–258]	0.14
Hospital length of stay (days), median [IQR]	8.5 [4-40]	7.9 [5–26]	6.8 [4–12]	11 [5–22]	9.1 [4–40]	0.74

CPB cardiopulmonary bypass, m-CPB modified cardiopulmonary bypass, VV-ECMO veno-venous extracorporeal membrane oxygenation, VA-ECMO veno-arterial extracorporeal membrane oxygenation, OR operating room, pRBC packed red blood cell, ICU intensive care unit

The use of ECLS was urgent in 19.4% of cases and prophylactic in 80.6% of cases. Peripheral cannulation was utilized in 89.6% of patients, and it was carried out in all patients receiving ECMO. Heparin-free and standard-heparin therapies were most prevalent in the VV-ECMO (92.9%) and modified-CPB (86.7%) groups, respectively. Group CPB had the highest total heparin dose (26,578 ± 6816 IU) and ACT (655 ± 196 s). There were no significant differences in other ECLS data (Table 3). Postoperative complications were most closely associated with previous surgical history, CPB, total heparin dose, ACT-max, and FEV₁% < 30%, and BMI > 25 kg/m² (Table 4).

Table 3.

Cardiopulmonary support (CPS) run details and adverse event

Group	Total study population (n = 67)	CPB (n = 18)	mini-CPB (n = 15)	VV-ECMO (n = 28)	VA-ECMO (n = 6)	P value
Planned, n (%)	54(80.6)	11(61.1)	13(86.7)	25(89.3)	5(83.3)	0.02
Cannulation, n (%)
Peripheral cannulation	60(89.6)	12(66.7)	14(93.3)	28(100.0)	6(100.0)	0.03
Central cannulation	7(10.4)	6(33.3)	1(6.7)	0(0.0)	0(0.0)	0.02
Pump duration (min)	95.4 ± 55.1	43 ± 18.2	64 ± 23.7	139 ± 45.6	127 ± 34.2	< 0.01
Mean pump flow (ml/kg)	52 ± 15.6	38 ± 19.7	36 ± 21.3	66 ± 14.7	35 ± 14.8	< 0.01
Sweep gas flow (L/min)	1.8 ± 0.3	1.9 ± 0.2	1.9 ± 0.2	3.4 ± 0.4	2.1 ± 0.2	0.16
Heparin free, n (%)	30(44.8)	0(0.0)	0(0.0)	26(92.9)	4(66.7)	< 0.01
Total heparin dose (UI)	10,255 ± 10,701	26,578 ± 6816	7213 ± 867	2725 ± 364	4033 ± 361	< 0.01
SaO2-R (%)	\	\	\	88.4 ± 2.7	\	< 0.01
PaO2-R (mmHg)	140.1 ± 68.8	187.4 ± 45.1	192.5 ± 33.2	67.5 ± 8.6	206.2 ± 31.2	0.00
PaCO2-R (mmHg)	42.7 ± 6.5	42.8 ± 6.6	41.9 ± 6.42	43.3 ± 7	42.4 ± 5.3	0.48
SO2 (pre-oxygenator)	78.3 ± 6.2	76.8 ± 9.5	74.4 ± 8.4	68.5 ± 5.4	75.0 ± 6.5	0.01
Lac-max (mmol/L)	2.2 ± 0.8	2.3 ± 0.8	2.4 ± 0.7	2.2 ± 0.7	1.9 ± 0.5	0.57
ACT-max (seconds)	321 ± 230	655 ± 196	250 ± 40	169 ± 18	209 ± 17	0.04
Major cannulation site bleeding, n (%)	9(13.4)	3(16.7)	2(13.3)	3(10.7)	1(16.7)	0.63
Limb ischemia, n (%)	1(1.5)	1(5.6)	0(0.0)	0(0.0)	0(0.0)	0.67
Vessel injury, n (%)	2(3.0)	1(5.6)	1(6.7)	0(0.0)	0(0.0)	0.72

CPB cardiopulmonary bypass, m-CPB modified cardiopulmonary bypass, VV-ECMO veno-venous extracorporeal membrane oxygenation, VA-ECMO veno-arterial extracorporeal membrane oxygenation, OR operating room, SaO₂-R oxygen saturation of the right radial artery, PaO₂-R oxygen partial pressure of the right radial artery, SO₂ oxygen saturation, Lac-max intraoperative maximum lactate value, ACT-max intraoperative maximum activated clotting time

Table 4.

Univariate analysis of factors associated with operative morbidity

Factor	Operative morbidity
	OR	95% CI	P value
Female gender	1.328	0.7438–1.482	0.458
Age > 60 years	1.017	0.859–1.263	0.325
BMI > 25 kg/m2	1.872	1.039–3.421	0.031
FEV1% < 30%	1.25	1.11–1.24	0.022
Urgent support	0.964	0.902–0.914	0.024
Progressive dyspnea	1.217	0.783–1.576	0.548
Diabetes	18.376	2.481–149.275	0.338
Hypertension	1.328	0.634–2.758	0.461
Previous surgical history	1.852	1.435–2.691	0.009
VATS	1.213	0.876–1.436	0.491
CPB	0.987	0.918–1.032	0.012
Pump duration	2.67	1.735–221.138	0.644
Heparin free	2.38	2.54–6.74	0.521
Total heparin dose	1.73	1.03–2.64	0.011
ACT-max	4.93	1.29–14.58	0.015
SaO2-R	2.55	0.32–18.37	0.104
PaO2-R	0.85	0.55–1.47	0.095

OR odds ratio, %95 CI confidence interval, BMI body mass index, FEV₁% percentage of forced expiratory volume in one second, VATS video-assisted thoracoscopic surgery, CPB cardiopulmonary bypass, SaO₂-R oxygen saturation of the right radial artery, PaO₂-R oxygen partial pressure of the right radial artery

Discussion

Airway surgery involving the trachea or main stem bronchi presents a unique challenge to effective airway management and perioperative oxygenation. Critical airway surgery is highly risky and very difficult, because most conventional airway procedures require non-ventilation to be performed for a limited time, and the surgical field is obstructed by the endotracheal tube inserted into the distal trachea [11]. Airway management strategies may include apneic oxygenation, high-frequency jet ventilation (HFJV), and a surgically placed bronchial tube during a tracheal procedure. However, these methods are often inadequate for oxygenation when used alone and can also result in excessive hypercapnia [12, 13], particularly in obese patients or those with chronic obstructive pulmonary disease [14]. As a cardiopulmonary support technology, ECLS may provide safe conditions and an unobstructed surgical field (especially if peripheral cannulation is used), as well as adequate time for precise dissection and reconstruction. This helps to avoid anastomotic tension and facilitates careful preservation of the tracheobronchial blood supply [15]. Furthermore, intraoperative ECLS may facilitate the adoption of video-assisted thoracoscopic surgery (VATS) in tracheal surgery, minimizing surgical trauma. This was demonstrated in our study of nine patients. In recent years, a growing amount of literature has confirmed the use of periprocedural ECLS in high-risk thoracic surgery, including lung transplantation, airway procedures, and mediastinal surgery [16]. Over the last twenty years, our department has used CPB, modified CPB and ECMO, with ECMO becoming the standard configuration over the last decade. This paper reviews our institutional experience of using ECLS in tracheal surgery. We report an operative mortality rate of 9 In recent years in 67 patients undergoing tracheal surgery with intraoperative ECLS over the last 20 years.

We believe that ECLS can be advantageous for carefully selected patients with complex airway tumors and, in some cases, for emergent life-threatening complications during anesthesia induction or surgery [7, 17]. Tracheal tumors are characterized by slow progression and a late onset of obvious symptoms of irritation and airway obstruction. Most patients only seek medical care once they have developed dyspnea or respiratory distress [18]. Increased airway stenosis may worsen dyspnea and cause wheezing and stridor [19]. In our series, the most common symptoms were cough (50.7%), breathlessness (50.7%), and progressive dyspnea (43.3%). Thorough preoperative assessments involving history taking, physical examinations, and imaging studies, with imaging tests considered very important for identifying patients at risk of airway complications [20]. A chest CT and FOB can demonstrate tracheal obstruction and permit accurate measurement of airway diameters. In addition, they help to determine the precise location and extent of tumor obstruction. All patients in our study underwent preoperative CT and at least one FOB and 88.1% of patients had a tracheal stenosis rate greater than 90%. For patients with difficult airways, including those with tumors invading the distal bronchi and carina or severe airway stenosis, where conventional endotracheal or fiberoptic intubation may be impossible or dangerous, prophylactic ECLS can prevent hypoxia and asphyxia before anesthesia is induced [21, 22]. Although studies on the use of ECLS before anesthesia induction to prevent airway and hemodynamic complications are limited to a few case reports, the prophylactic establishment of a femoro-femoral bypass under local anesthesia has been safely performed as an elective procedure [23]. Acute cardiopulmonary decompensation may be triggered by changes in position, loss of muscle and diaphragm tone, and alterations in lung compliance and chest wall structure during anesthesia and surgery [24]. Even with a primed circuit and prepared team, at least 5 to 10 min are required for cannulation and establishing adequate support. Such delays can lead to adverse neurological consequences [25]. In our series, ECLS was used prophylactically in 80.6% of cases, and VV-ECMO was more common than CPB (89.3% vs 61.6%). For all patients at high risk of asphyxia or difficult cross-field intubation, we began introducing VV-ECMO prophylactically. Otherwise, we inserted deep venous catheters into the right internal jugular vein and the femoral vein to facilitate urgent ECMO support after 2013. Based on this study and previous reports [11, 26], Based on our experience, we recommend establishing ECLS for prophylactic purposes before the induction of anaesthesia in patients classified as'unsafe'or'uncertain'when the following conditions occur: Severe positional symptoms (SaO₂ < 80% or tachycardia) due to airway obstruction; FEV₁ < 30%; stenosis rate > 90%; tracheal lumen < 5 mm; lesion adhesions; cross-field ventilation not feasible; and risk of tumorous rupture or tracheobronchomalacia.

Despite growing experience and progressive technology, tracheal resection and reconstruction remain technically challenging procedures that carry a risk of significant morbidity and mortality in cases of complex tracheal disease [27, 28]. Bibas et al. reported the rates of overall complications, anastomotic complications and non-anastomotic complications to be 44.6%, 21% and 23.2% respectively [29]. Anastomotic complications, including granulation tissue, restenosis and anastomotic fistulae accompanied by wound infections, are not uncommon following tracheal resection [20]. A study by Amoros et al. reported a rate of 20.9% for minor complications and 4.1% for major complications in patients undergoing tracheal resection [30]. In the present study, overall complications (26.9%) and anastomotic complications (16.1%) were kept at low levels in critical patients (Table 2). Table 4 shows that obesity, concomitant diabetes, declining lung function and anticoagulation carry significant additional risk for operative morbidity. These complications were also related to the application of ECLS and its configuration. After 2013, only six patients opted for VA-ECMO due to comorbid heart disease and multiple other health issues. Two of these patients died postoperatively from multiple organ failure and infections, respectively, which contributed to the higher mortality rate in the VA-ECMO group.

VA-ECMO provides full respiratory support as well as circulatory stability during tracheal surgery, but as we know, it does not come without potential risks [31]. Complications of VA-ECMO include vascular injury and hemorrhage at the cannula site, thromboembolism, ischemia of the limb and Harlequin syndrome, which is secondary to intraoperative ventilatory arrest [32, 33]. Such an acute desaturation episode was related to a persistent desaturated cardiac blood flow across nonventilated lungs, resulting in a clinically significant differential hypoxia [34, 35]. At our center, we routinely perform intraoperative right radial artery pulse oximetry and near-infrared spectroscopy (NIRS) monitoring to identify upper body hypoxia, which is very important in patients with VA-ECMO or peripheral CPB. Fatal brain ischemia may be avoided by altering the ventilator settings or an additional venous cannula providing arterialized blood to the right heart (VAV-ECMO mode) [36]. Due to the recirculation phenomenon in VV-ECMO, systemic arterial saturations do not increase to 100% [37, 38]. Patient oxygenation can be improved by increasing pump flow and oxygenating a greater proportion of the cardiac output. In practice, ECMO blood flow should be around more than two-thirds of the patient’s cardiac output to provide acceptable arterial saturation [39].

Initially, conventional CPB was used for intraoperative cardiopulmonary support in complex thoracic surgery. The venous reservoir in CPB enables suctioned blood to be re-transfused and allows drugs and fluids to be administered [40]. However, CPB has several disadvantages: direct blood-air contact strengthens coagulation and inflammation, a venous reservoir increases priming volume and hemodilution, and suctioned blood poses a risk of tumor scattering [40, 41]. Given the risk of significant blood loss, we still favour CPB for surgeries involving large mediastinal tumors that compress or invade the heart, major vessels, and tracheobronchial tree. In our case series, three patients still opted for CPB due to reoperation or tumor proximity to major vessels for quick volume resuscitation and perfusion of vital organs. Modified CPB has been shown to decrease the inflammatory response, reduce hemodilution, lower the requirement for anticoagulation and reduce the need for red blood cell transfusions. These advantages were reflected in a reduced need for mechanical ventilation and shorter ICU and hospital stays [42]. Since 2008, we have switched from CPB to modified CPB in tracheal surgery, achieving satisfactory clinical results. This transitional technique allows the procedure to be performed with less priming, lower anticoagulation and fewer complications [10]. In recent years, there has been an increasing trend towards the use of VV-ECMO at our center. This avoids cross-field intubation and differential hypoxemia while providing a clearer surgical view and more ventilator-free time. In this series, the longest VV-ECMO runtime was 198 min, while the longest period without mechanical ventilation was 124 min. Different ECLS configuration has their characteristics and adaptable scenarios. Different ECLS configurations have different characteristics and adaptable scenarios. A summary of the advantages and disadvantages of different ECLS configurations can be found in Table 5.

Table 5.

Comparison of advantages and disadvantages of different CPS

CPS	CPB	m-CPB	VA-ECMO	VV-ECMO
Hemodynamic support	full support	full support lower, but with risk of hemodynamic instability due to blood loss and insufficient drainage	full support lower, but with risk of hemodynamic instability due to blood loss and insufficient drainage	no, with risk of hemodynamic instability due to blood loss
Respiratory support	full support	partial support due to native cardiac output	partial support due to native cardiac output	full support
Recirculation	no	no	no	yes
Harlequin syndrome	no	yes	yes	no
Oxygen delivery	full	full	full	barely enough
Hemodilution	higher	moderate	lower	lower
Inflammatory activation	higher	moderate	lower	lower
Field blood saver and cardiotomy reservoir	yes	no	no	no
ACT	> 480 s	< 200 s	< 200 s	< 200 s
Bleeding	severe	moderate	mild	mild
Cannulation	central or peripheral	peripheral	peripheral	peripheral
Runtime	< 6 h	< 6 h	weeks and back to ICU	weeks and back to ICU
Cost	lower	moderate	higher	higher
Cancer dissemination risk	yes	no	no	no

CPB cardiopulmonary bypass, m-CPB modified cardiopulmonary bypass, VV-ECMO veno-venous extracorporeal membrane oxygenation, VA-ECMO veno-arterial extracorporeal membrane oxygenation, ACT activated clotting time

Our study has several limitations. Firstly, it is a retrospective, observational, single-center study. Secondly, there was no control group of patients who did not receive ECLS in similar clinical scenarios. Thirdly, despite being one of the largest contemporary series of tracheal surgeries involving ECLS, the small number of cases in our retrospective study limits its scientific value. We believe that each case must be evaluated individually, with careful consideration given to the technical challenges of the procedure and the risks associated with ECLS. These risks include systemic anticoagulation, hemodilution, inflammatory activation, differential hypoxia syndrome, cannulation complications, and longer operative times.

Conclusion

Despite intraoperative ECLS being rarely used in tracheal and carinal disease, programmed extracorporeal life support with VV-ECMO demonstrated high effectiveness with an acceptable mortality and morbidity rate in high-risk tracheal surgery. And our case series revealed that ECLS could be a feasible strategy in the individualized management of these patients. With advances in ECLS and further accumulation of experience, a more sophisticated protocol in tracheal surgeries can be derived.

Abbreviations

CPB

Cardiopulmonary bypass

ECMO

Extracorporeal membrane oxygenation

VV-ECMO

Veno-venous ECMO

VA-ECMO

Venoarterial ECMO

ECLS

Extracorporeal life support

pRBC

Packed red blood cells

F_GO₂

Fraction of sweep gas oxygen

PaO₂

Partial pressure of oxygen

ACT

Activated clotting time

UFH

Unfractionated heparin

HFJV

High-frequency jet ventilation

Authors’ contributions

Lin Sun and Xiaoyu Zhou: Conceptualization, Methodology, Investigation, Writing—Original Draft; Liping Xin: Data Curation, Writing-Original Draft & Editing Yuwei Qiu (Corresponding Author 2) and Lingfeng Xu: Resources, Writing—Review & Editing; Xin Chang: Visualization and Resources; Xiaoyu Zhou: Data Curation, Statistical Analyses and Writing—Original Draft; Zhen Guo (Corresponding Author 1): Conceptualization,Resources, Supervision, Writing—Review & Editing. All authors have read and agreed to the published version of the manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

All medical data in the article were obtained from the hospital's local database and approved by the Ethics Committee of Shanghai Chest Hospital. The medical data were collected while ensuring the protection of personal information and privacy. Data is provided within the manuscript or supplementary information files.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Shanghai Chest Hospital, Shanghai JiaoTong University School of Medicine (approval no. KSY28713). The medical data were collected while ensuring the protection of personal information and privacy, and obtaining informed consent was not deemed necessary.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.