The Safety and Feasibility of Laryngeal Mask Airway (LMA) Management in Non-intubated Subxiphoid-subcostal Thoracoscopic Thymectomy: Preliminary Results

Intubated general anesthesia and single-lung ventilation are considered mandatory for conventional thoracoscopic surgery. Non-intubated thoracoscopic thymectomy is technically challenging. The aim of this article was to present the initial results of non-intubated subxiphoid-subcostal thoracoscopic thymectomy (NI-STT) under LMA management for patients with thymic tumor or myasthenia gravis (MG) and to investigate the feasibility and safety of the procedure. A retrospective analysis of patients undergoing NI-STT for thymic tumor or MG at our department from January 2017 to January 2020 was performed. The clinical characteristics and perioperative outcomes of the patients were reviewed and analyzed. A total of 61 patients were received NI-STT in this analysis, of which 19 patients with MG undergone an extended thymectomy and the rest (n=42) undergone a partial thymectomy. The anesthetic induction duration, surgical duration and global operating room duration were 24.83±12.27 min, 118.75±32.49 min and 173.51±41.80 min, respectively. The lowest SpO2 and peak EtCO2 during operation were 96.15±2.93 mmHg and 41.79±7.53 mmHg, respectively. The mean duration of chest drainage and postoperative hospital stays were 1.87 days, and 2.91 days, respectively. Three cases had sore throat and irritable cough and two cases suffered nausea and vomiting occurred. one patient suffered from an atrial brillation, two patients experienced pneumonia, and one patient suffered wound infection, respectively. There were no phrenic nerve paralysis and mortality occurred in the study group. The postoperative pain was low on 1,3, 7, 14, 30, 90 and 180 postoperative days. NI-STT was a technically safe and feasible approach for treating thymic tumors or MG. It could be an alternative to intubated single-lung ventilation for thymectomy in selected patients. anesthesia techniques for non-intubated VATS described . All fasted for 8 hours The pulse oxygen saturation (SpO 2 ), electrocardiogram (ECG), noninvasive arterial blood pressure, bispectral index (BIS) and end-tidal carbon dioxide (EtCO 2 ) were routinely monitored. Fiberoptic bronchoscope and endotracheal intubation equipment should be provided before anesthesia, and tracheal intubation could be performed immediately if necessary. The operation and positioning of the disposable Supreme TM laryngeal mask airway (SLMA) were performed by experienced and professional anesthesiologists. The correct position of LMA was judged by lung and neck auscultation, airway pressure waveform, the capnography waveform, and beroptic bronchoscope. If LMA was not in the correct position, it should be removed and reinserted.


Introduction
Thymectomy plays an important role in the management of thymic tumors and MG. Video-assisted thoracoscopic surgery (VATS) for thymectomy has been widely used in selected patients with the bene ts of better cosmetic results, less trauma, fewer complications, shorter hospital stays and equivalent e cacy compared with conventional open thymectomy 1-2 . However, previously lateral approach VATS thymectomy still has drawbacks, such as poor visualization of neck region and contralateral phrenic nerve, intercostal nerve impairment leading to intercostal numbness and pain. Hence recently, subxiphoid approach VATS thymectomy has been successfully applied in clinical practice and favored by surgeons during the past decade. The major advantages of this modality are as follows: broader visual eld of bilateral phrenic nerves and neck region, less pains, and better cosmetic results [3][4][5][6] .
Traditionally, double-lumen endotracheal intubation and endobronchial blocker for one-lung ventilation are considered as s mandatory and indispensable part for thoracoscopic surgery, including lateral approach VATS thymectomy. Nonetheless, complications and adverse effects after tracheal intubation and general anesthesia could not be neglected, such as barotraumas, vocal cords injury, sore throat, postoperative dysphagia, irritating cough and increased sputum and risk of pneumonia [7][8][9][10] . For the above disadvantages of intubated anesthesia, non-intubated spontaneous ventilation thoracoscopic surgery was established and adapted by thoracic surgeons with encouraging results. And now it has gradually become another option for various thoracoscopic surgeries, such as bullectomy, wedge sections and lobectomy of lung, and mediastinal tumor resection [7][8][9][10]. However, as the diaphragm movement under non-intubated spontaneous ventilation anesthesia will interfere with the surgical instruments and further increase the di culty of surgery, there are few reports of subxiphoid thoracoscopic thymectomy under non-intubation spontaneous ventilation anesthesia 11 .
In view of this, we adopted a LMA for non-intubated anesthesia and mechanically controlled ventilation to avoid the side effects of tracheal intubation and extubation. In the present retrospective study, we reported our preliminary experiences of NI-STT with mechanically controlled ventilation under LMA management to evaluate its safety and feasibility.

Study Design and Patients
We performed the subxiphoid-subcostal approach thoracoscopic surgery since 2013 and non-intubation thoracoscopic surgery since 2015. Thereafter, we performed NI-STT under LMA management in some cases in 2016. After reviewing the patients' medical records and discussing the procedures with the patients, the surgeons and anesthesiologists selected patients who were considered for NI-STT. All operations were performed by the same surgical team and anesthesia team using the same clinical protocols, care patterns, and perioperative orders. A series of consecutive patients with anterior mediastinal tumors or MG were admitted and received NI-STT between January 2017 and January 2020 in our department. The data of these patients were retrospectively collected and analyzed through patient medical notes and computerized records. Informed consent about operative and anesthesia techniques and data use agreement were obtained from all patients prior to surgery. All methods were carried out in accordance with relevant guidelines and regulations. This retrospective descriptive study was reviewed and approved by the institutional review board (IRB) of Tongji Hospital. The primary outcome of present study was the feasibility and safety of NI-STT.
The inclusion criteria for NI-STT were: (I) Diagnosis of thymic tumors by using chest computed tomography or magnetic resonance imaging; (II) American Society of Anesthesiologists (ASA) classi cation≤III. (III) Aged 18 to 65 years. Patients with obvious tumor invasion of surrounding organs, previous thoracic surgery, potential pleural adhesions, overweight (body mass index (BMI) >30 kg/m 2 ), sleep apnea, deranged preoperative arterial blood gases analysis, impaired preoperative forced expiratory volume in 1 second (FEV 1 ) and forced vital capacity (FVC), a heart failure, bleeding disorder or unfavorable features of airway, chest wall or spine were excluded. The patients combined with MG were discussed with neuroscientists through a multidisciplinary committee and treated with pyridostigmine or prednisone, for the purpose of determining the indications for surgery, and more importantly, determining the appropriate timing of surgery based on the stability of MG symptoms to reduce the risk of postoperative myasthenia crisis.
Anesthetic setting, induction, and maintenance The anesthesia techniques for non-intubated VATS were described previously [7][8][9][10][11] . All patients fasted for at least 8 hours before surgery. The pulse oxygen saturation (SpO 2 ), electrocardiogram (ECG), noninvasive arterial blood pressure, bispectral index (BIS) and end-tidal carbon dioxide (EtCO 2 ) were routinely monitored. Fiberoptic bronchoscope and endotracheal intubation equipment should be provided before anesthesia, and tracheal intubation could be performed immediately if necessary. The operation and positioning of the disposable Supreme TM laryngeal mask airway (SLMA) were performed by experienced and professional anesthesiologists. The correct position of LMA was judged by lung and neck auscultation, airway pressure waveform, the capnography waveform, and beroptic bronchoscope. If LMA was not in the correct position, it should be removed and reinserted.

Surgical technique
Following the induction and maintenance of general anesthesia and mechanically controlled ventilation with LMA, all patients underwent subxiphoid thoracoscopic thymectomy, which was similar to that reported previously 12 . Patients were placed in a supine position with legs spread out and shoulders raised by a cushion. During the operation, the surgeon stood between the patient's legs, while the assistant was on the right side and operated the camera. The nurse stood on the left side of the patients and the monitor was placed at the cranial side. A 3.0-cm longitudinal skin incision was made rst to expose the xiphoid and serve as observation hole. The retrosternal space was created and enlarged by blunt dissection with oval forceps or ngers. Under the guidance of ngers, two 0.5-cm subcostal holes were created at the midclavicular lines under bilateral costal arches and served as operation holes. Thereafter, a 10 mm, 30-degree angled thoracoscope was placed into the retrosternal space through the observation hole. Arti cial pneumothorax was created by carbon dioxide (CO 2 ) insu ation under a positive pressure of 8cm-H 2 O, which could expand the retrosternal space and ensure a su cient surgical eld. The procedure for patients without MG was a partial thymectomy, which was de ned as complete removal of the entire thymic gland along with the surrounding adipose tissues. While the surgery for patients combined with MG was an extended thymectomy, which was de ned as removal of entire thymus and adipose tissues from mediastinal, pericardiophrenic, aorta-pulmonary window and cervical area. The bilateral mediastinal pleurae were rstly dissected with an ultrasonic scalpel to expose the bilateral phrenic nerves. Along the front of the bilateral phrenic nerves to the neck, the lateral margin of thymus and its surrounding fat and adipose tissues at bilateral cardiophrenic angle were dissected. Great attention should be taken to the protection of bilateral phrenic nerves during the operation. The posterior border of the thymus was meticulously separated from the anterior edge of the pericardium to con rm the superior vena cava and bilateral innominate veins. The thymic veins were identi ed prior to draining into the innominate vein and meticulously sealed using an ultrasonic scalpel. By gently grasping the thymus, the bilateral superior poles of thymus were conveniently dissected by using an ultrasonic scalpel.
Finally, the whole thymus and the surrounding adipose tissues excised en-bloc from pericardium, cardiac diaphragmatic angle, innominate veins, brachiocephalic artery, trachea, and cervical area were placed in a plastic bag and removed from the subxiphoid incision ( Figure 1). A 20-Fr chest drainage tube was inserted into the mediastinum through subxiphoid incision the after the surgical eld was evaluated. The bilateral subcostal incisions were sutured. The LMA was removed immediately after operation.
Postoperative management and data collection Dynamic monitoring of ECG and SpO 2 were performed in the early postoperative period. For patients with MG, the same medications used preoperatively were restarted postoperatively under the guidance of the neurologists. Chest radiography was performed at the rst day postoperatively and after the extubation of chest drainage tube. Blood detection was performed on the rst and third postoperative days. Early removal of pleural drainage tube was encouraged postoperatively, if total drainage was less than 200 mL/day with no air leakage and chest X-ray chest radiographies and vital signs were normal. All patients were discharged without a pleural drainage tube. In general, no postoperative analgesics were used.
Unless the pains were intolerable, low-dose diclofenac sodium could be used. The visual analog scale (VAS) was used to evaluated the pain on 1, 3, 7, 14, 30, 90, and 180 postoperative days (POD). The demographic data, duration of hospital stay and chest drainage, operative and anesthetic results, side effects and complications were obtained from the institutional database, anesthesia and surgical notes, and the medical and nursing records.

Statistical analysis
All data analyses were conducted using SPSS version 22.0 (IBM SPSS Statistics, Chicago, USA). If the continuous variables had a normal distribution, it would be represented as mean ± standard deviation, otherwise it should be described as the median and range. Categorical variables were expressed as count and proportion.

Results
A total of 61 patients between January 2017 and January 2020 were received NI-STT under LMA management in this analysis, of which 19 patients with MG undergone an extended thymectomy and the rest (n=42) undergone a partial thymectomy. The demographic and clinical characteristics of the patient's cohort were summarized in Table 1 Table 2 showed that the mean intraoperative blood loss was 46.37ml. The LMA was removed immediately after operation in all patients. The mean duration of chest drainage was 1.87 days, and mean postoperative hospital stays was 2.91 days. The hospitalization expenses were 5.03±0.95 thousand US dollars (USD). Side effects of anesthesia occurred in three cases of sore throat and irritable cough and two cases of nausea and vomiting. One patient suffered from an atrial brillation and received treatment of amiodarone. Two patients experienced pneumonia, and one patient suffered wound infection. All the patients recovered well when discharged. There were no phrenic nerve paralysis and mortality occurred in the study group. Table 3 showed that the postoperative pain (VAS score) was low on POD 1,3, 7, 14, 30, 90 and 180. No tumor recurrence or metastasis was found during follow-up.

Discussion
With a combination of subxiphoid-subcostal access and non-intubated anesthesia with mechanically controlled ventilation under LMA management, our study showed that NI-STT was safe and feasible and resulted in excellent perioperative outcomes and short hospital stays. Thoracoscopic surgery without endotracheal intubation avoids intubation-related complications, leading to relief of symptoms, a lower rate of complications and a fast recovery.
Over the past several decades, various surgical techniques have been advocated for thymectomy in treating anterior mediastinal tumors and MG. Although trans-sternal surgery remains the gold standard of thymectomy, with the improvements of thoracoscopic instruments and techniques, VTAS surgery has been an alternative to thymectomy 1-2 . There are different approaches to VATS thymectomy, which includes the unilateral, bilateral, transcervical, robotic, subxiphoid or combination approach, and the main choice is the lateral approach 13,14 . Although there were efforts to reach a consensus regarding the standard VATS thymectomy, but no consensus has been available now 14 . The choice of surgical approach is usually based on the experience and preference of the surgeon. However, the lateral VATS approach is associated with di culties in identifying the contralateral phrenic nerve, narrow operative eld in neck region and persistent intercostal nerve paralysis or neuralgia 15 . Hence, in recent years, Hsu et al. 16 and Suda et al. 4,5,15 previously reported a thoracoscopic thymectomy via the subxiphoid approach without any intercostal incision, which has several advantages 4,5,12,15,16 . First, the excellent surgical eld from subxiphoid approach is similar to the view of median sternotomy and helps con rm the superior poles of thymus and bilateral phrenic nerves. The adequate visualization allows the maximum resection of thymus and surrounding fat tissues, reducing the chance of accidental blood vessels or phrenic nerve injury. Second, with the help of arti cial pneumothorax created by CO 2 insu ation instead of lifting of the sternum, the mediastinum is compressed and the bilaterally ventilated lungs are shifted away, greatly providing the operative vision and allowing for safer dissection. In our experience, low pressure CO 2 insu ation into the mediastinum does not cause the changes in hemodynamic status of the patients. In addition, subxiphoid approach thymectomy does not cause intercostal nerve injury by avoiding any intercostal incision. Therefore, the postoperative pain was signi cantly relieved, and there was no abnormal chest wall paresthesia in the study. Less pains mean less usage of analgesics. Also, there is no visible scar in the neck and upper chest area, which is also a clear cosmetic advantage. Furtherly, in our experiences, subxiphoid VATS thymectomy has short drainage duration, postoperative hospital stays and low hospitalization expenses, comparable to other studies 15,17 .
Double-lumen intubation and single-lung ventilation general anesthesia technique were needed for conventional thoracoscopic surgery, which met the need of intraoperative lung ventilation and provided the surgeon with a broad surgical eld. However, many complications are caused by intubation and extubation procedures, such as airway trauma, residual neuromuscular blockade, and ventilator-induced lung injury. To overcome the complications associated with endotracheal intubation, non-intubated VATS was recently described in selected patients with thoracic disease, including pulmonary and mediastinal diseases, and even minimally invasive esophagectomy and pleural empyema 7-11,18−22 . Non-intubated anesthesia may avoid the potential complications that the intubation may result in, and may be helpful to judge the reason of postoperative irritable cough, sore throat, and hoarseness. One major concern about using the non-intubated anesthesia for VATS thymectomy was how respiratory function is maintained in a patient with an arti cial pneumothorax under the management of LMA. To prevent respiratory failure, we selected patients with good cardiopulmonary function for this study. Only one patient developed persistent hypoxia and was converted to intubated general anesthesia nally. In most of the nonintubated patients, SpO 2 was satisfactorily maintained above 90% during the whole operation.
Nonetheless, patients with poor cardiopulmonary function should be carefully examined and selected before attempting this procedure.
The other major concern is whether it is feasible and safe to perform a surgical procedure during nonintubated thoracoscopic thymectomy. Dissection of blood vessels, and bilateral phrenic nerves are more technically demanding because of the diaphragmatic movement during spontaneous breathing, especially when the respiratory rate was more than 20 breaths per minute. The surgeons must adjust the instruments according to the rhythm of the breathing and diaphragm movement to make a precise dissection, just like beating heart coronary bypass surgery. We performed the NI-STT since 2016. In our experiences, the diaphragmatic movement during spontaneous breathing would interfere with the surgical procedures, resulting in an increase in the di culty of the operation and a prolonged learning curve. Therefore, in our patient cohort, we tried to preform NI-STT without spontaneous breathing under LMA management. The use of muscle relaxants inhibited the movement of diaphragm and spontaneous breathing, so that it did not interfere with surgical procedures. Besides, additional intrathoracic vagal blockade was not required to abolish the cough re ex by using the muscle relaxants during non-intubated thymectomy in our study group 22 . Finally, epidural anesthesia or paravertebral block was used traditionally for non-intubated spontaneous ventilation anesthesia, however, it may lead to side effects such as hypotension, bradycardia, urinary retention, epidural hematoma, or infection. In this present study, Intercostal nerve block and rectus abdominis sheath block were performed with 0.4% ropivacaine instead of epidural anesthesia or paravertebral block. And the results revealed that postoperative pain was mild in almost all cases, indicating that local anesthesia was also feasible and effective in this procedure.
However, the NI-STT also has disadvantages. Even with the presence of arti cial pneumothorax, an excessively high BMI would lead to a narrow surgical eld and further make surgery di cult. In addition, for patients with tumors invading surrounding organs or large vessels, this method may be not suitable, and open surgery should be recommended. In addition, we acknowledged that the present study was limited by its single-center retrospective nature and a small number of patients, and thus, intrinsic bias may exist. Besides, the lack of a control group who received endotracheal intubated anesthesia makes it di cult to con rm the speci c bene ts of NI-STT. However, the low conversion rate of non-intubated to intubated anesthesia and low complication incidence indicate that NI-STT could be safely performed in selected patients. Further research is encouraged to clarify the applicability and bene ts of NI-STT for speci c patient groups.

Conclusion
In summary, the present study showed that NI-STT under LMA management was technically feasible and clinically safe for treatment of anterior mediastinal tumors and MG. NI-STT avoided complications related to endotracheal intubation and intercostal incision, leading to relief of symptoms, lower complication rate, and faster recovery. Although true bene ts of NI-STT should be further veri ed in larger prospective randomized controlled studies, it could be an attractive alternative in selected patients with thymic tumor or MG.