Complications of Video-Assisted Thoracic Surgery: A Five-Year Experience

Transcription

1 Complications of Video-Assisted Thoracic Surgery: A Five-Year Experience Ren6 Jancovici, MD, Loic Lang-Lazdunski, MD, Francois Pons, MD, Louis Cador, MD, Antoine Dujon, MD, Marcel Dahan, MD, and Jacques Azorin, MD Departments of Thoracic Surgery, H6pital du Val de Grace, Paris; Clinique du C6dre, Rouen; H6pital Purpan, Toulouse; and H6pital Avicenne, Bobigny, France Background. Although thoracoscopy was originally described in 1910, recent developments in video-assisted surgical techniques and endoscopic equipment has expanded the application of video-assisted surgical procedures in the field of thoracic surgery. Methods. In an effort to define both high-risk patients for video-assisted thoracic procedures and high-risk video-assisted thoracic surgical procedures, we reviewed the experience of four surgical institutions from June 1991 through May We looked specifically at complications resulting from the 937 video-assisted thoracic procedures performed during this period. Results. Perioperative incidents or complications occurred in 35 patients (3.7%), and 116 procedures (12.4%) were converted to a thoracotomy. The in-hospital mortality rate was 0.5%, and death occurred principally in patients operated on for malignant pleural effusion. The overall incidence of postoperative complications was 10.9%, and the most prevalent complications were prolonged air leak (6.7%) and pleural effusion (0.7%). Conclusions. The incidence of complications was acceptable and, except for that of prolonged air leak, did not differ significantly from that resulting from analogous open procedures. Video-assisted thoracic surgery appears safe and particularly useful for some indications. However, the possibility of dramatic life-threatening perioperative complications requiring emergency conversion to thoracotomy justifies the fact that only trained thoracic surgeons should perform video-assisted thoracic surgical procedures. (Ann Thorac Surg 1996;61:533-7) he rapid development of endoscopic surgery and T particularly of video-assisted thoracic surgery has prompted thoracic surgeons to perform video-assisted surgical procedures with increased frequency and to enlarge their indications, sometimes without data demonstrating benefit. Six of us, French thoracic surgeons from different institutions but all members of the Thorax Group created in September 1990 in France, report here our 5-year experience with incidents and complications encountered in more than 900 video-assisted thoracic surgical (VATS) procedures to define high-risk procedures, inappropriate indications for VATS intervention and patients at high risk for the procedures. Material and Method A standardized data collection form on each patient having a thoracoscopic or VATS procedure was completed at the time of discharge in each institution. The form covered demographic data, indications, type of surgical procedure, type of anesthesia, type of analgesia, and perioperative or postoperative complications. From June 1991 through May 1995, 937 forms were completed and received in our department, where they were ana- Accepted for publication Oct 23, Address reprint requests to Dr Jancovici, Department of Thoracic Surge~, H6pital du Val de Grace, 74 blvd de Port-Royal, Paris Cedex 05, France. lyzed by three of us. This analysis forms the basis of this report. Of the 937 patients, 70.1% were male. Average age at operation was 40.6 years (range, 16 to 88 years). Various VATS procedures were performed in 896 patients (95.6%) and video-assisted lobectomy, in 41 patients (4.4%). The indications for operation are listed in Table 1. The VATS procedures were performed with trocars or ports and sometimes incorporated very small incisions never exceeding 3 cm in length. Video-assisted lobectomy procedures included one trocar opening for visualization and one incision in the fifth intercostal space or on the midaxillary line. The length of the incision was 5 to 7 cm, but the intercostal space was usually opened from 10 to 14 cm. Video-assisted lobectomies were performed exclusively by one of us (J.A.). The surgical procedures are shown in Table 2. The majority of patients were operated on in a lateral position, and most operations were performed with general anesthesia and a double-lumen endotracheal tube, good positioning of which was confirmed bronchoscopically. Transcutaneous oximetry and end-tidal carbon dioxide tension were measured continuously in most patients. A small number of procedures were done using a single-lumen endotracheal tube or local anesthesia in patients with very poor cardiopulmonary status. All patients were taken from the operating room with one or two chest tubes, depending on the surgical procedure, in place by The Society of Thoracic Surgeons /96[$15.00 Published by Elsevier Science Inc SSDI (95)

2 534 JANCOVICI ET AL Ann Thorac Surg COMPLICATION OF VATS 1996;61:533-7 Table 1. Indications for Operation ~ Indication Patients Percent Pneumothorax/bullous emphysema Pulmonary nodule/mass Lung biopsy Malignant pleural effusion Mediastinal mass/cyst Lung cancer (T1 or T2 NO) Hyperhidrosis of upper limbs Staging of lung cancer Empyema Trauma/hemothorax Bronchopleural fistula Reoperation for bleeding Pericardial effusion Diaphragmatic injury Pulmonary hydatid cyst Truncal vagotomy Chylothorax Esophageal lesion Total 937 Numbers in parentheses are percentages. Postoperative pain was managed by several techniques. Most common was continuous administration of narcotic agents through thoracic epidural analgesia or patient-controlled narcotic analgesia in select patients during the first 48 postoperative hours. Then, nonsteroi- Table 2. Thoracoscopic and Video-Assisted Thoracoscopic Surgical Procedures" Procedure Patients Percent Pleurodesis/pleurectomy ,5 Wedge resection Lung biopsy Drainage of malignant effusion/ talc pleurodesis Biopsy/resection of mediastinal mass/cyst VATS lobectomy Thoracic sympathectomy Nodal examination/biopsy Pleural drainage/hemostasis Drain empyema Bronchopleural fistula closure Pericardial drainage/biopsy/window Diaphragmation repair Hydratid cyst resection Truncal vagotomy Esophageal operation Total 937 a Numbers in parentheses are percentages. VATS - video-assisted thoracic surgical. dal antiinflammatory drugs or analgesic agents such as codeine or acetaminophen were administered orally. Chest tubes usually were removed on the second to the fifth postoperative day, depending on the surgical procedure, the daily output, and the postoperative chest roentgenograms. Results Mortality Five patients (0.5%) died during the postoperative period. Four had undergone operation for malignant pleural effusion and died during the first postoperative month. Causes of death were pulmonary embolism in 1, septic shock in 1, ventricular arrhythmia in 1, and respiratory insufficiency in 1. The fifth patient had been referred for a chest gunshot wound. He had been shot 3 weeks before the admission. Preoperative chest magnetic resonance imaging disclosed a bullet in the pericardium on the left side of the heart. We decided to remove it by a left-sided VATS procedure. At operation, the bullet was situated in the pericardium on the left side of the left atrium and was easily removed under thoracoscopy. No cardiac wound was diagnosed at operation, and there was no pericardial effusion. The patient died suddenly 36 hours after operation. No postmortem examination was done. We speculate this death was related to a sudden ventricular arrhythmia. Morbidity Perioperative incidents or complications occurred in 35 patients (3.7%), and 116 patients (12.4 /,,) required conversion to thoracotomy. There were several reasons for converting to an open procedure. The presence of a vascular wound was the cause in 18 patients. These vascular lesions were observed mainly during lobectomies for cancer and occurred either during vessel division or during stapling. In most cases, the lesion was a pulmonary artery branch wound that occurred during division of a mediastinal artery in proximity to an inflammatory adenopathy. One patient required conversion on an emergency basis; the endostapler cut but failed to staple a left pulmonary vein, and a massive hemorrhage occurred. In 3 patients, the reason for converting was the presence of very tight pleural adhesions and in 3 others, the impossibility of obtaining a correct lung exclusion. In 47 patients, the reason was either the inabili~ to locate a pulmonary nodule or the difficulty of resecting a mediastinal cyst or tumor. In 17 patients undergoing a wedge resection for a pulmonary nodule, the pathologic examination disclosed either a primary lung cancer, which necessitated a thoracotomy to perform an "open" lobectomy, or the presence of malignant cells in the stapling line, this necessitating a larger parenchymal excision or a lobectomy. Two tracheobronchial wounds occurred during videoassisted resection of bronchogenic cysts of the mediastinum. Thoracotomy was necessary to repair the trachea in 1 patient and to repair the left main bronchus in the other. Six patients with malignant mediastinal tumors

3 Ann Thorac Surg JANCOVICI ET AL ;61:533-7 COMPLICATION OF VATS Table 3. Postoperative Complications '~ Complication Patients Percent Prolonged air leak Pleural effusion Neurologic Chest wall involvement Deep vein thrombosis Pneumonia Chylothorax Empyema Septicemia Neoplastic seeding Arrhythmia Pulmonary embolism Respiratory insufficiency Renal failure Total Numbers in parentheses are percentages. diagnosed by a VATS technique required an open procedure for complete resection. In 20 patients, thoracotomy was the decision of choice for economic reasons or because of equipment failure. The latter generally consisted of sticking of the staple line and failure to release tissue, but rupture of the material was observed in 2 patients. In a few patients with giant bullae or parenchymal lesions requiring major wedge resections that would have necessitated numerous hazardous stapling procedures, conversion to thoracotomy was considered optimal. The overall incidence of postoperative complications was 10.9%. The frequency of each complication is reported in Table 3. There were infectious complications in 9 patients. Three empyemas occurred after pleural abrasion for spontaneous pneumothorax and necessitated prolonged irrigation and drainage. Four patients had pneumonia, 2 after wedge resections and 2 after video-assisted lobectomies. These complications necessitated parenteral antibiotic treatment and a prolonged hospital stay (mean stay, 14 days). Two patients had postoperative septicemia, 1 after a video-assisted lobectomy and 1 after talc pleurodesis for a malignant pleural effusion. Five patients had neurologic complications. These complications occurred principally after a mediastinal procedure; 4 (8.2%) of 49 patients having resection or biopsy of a mediastinal mass or tumor experienced a neurologic deficit. There was one left recurrent laryngeal nerve palsy after resection of a left vagus nerve neurinoma and one left phrenic nerve palsy after resection of a neurinoma involving this nerve. Two patients experienced a transient phrenic nerve palsy after resection of a mediastinal lymphoma in 1 and resection of a bronchogenic cyst in the other. These palsies resolved after a few months. One patient sustained a regressive right recurrent laryngeal nerve palsy after a right-sided thoracoscopy for staging of a primary lung cancer. Two patients experienced neoplastic complications. They were operated on for malignant pulmonary nodules and underwent wedge resection of the lung by a videoassisted procedure without the chest being opened. Both had a neoplastic chest wall recurrence at the trocar site a few months after the initial resection. The nodules had been extracted from the chest cavity without the use of an endoscopic bag. Five patients experienced trocar-related or port-related complications. Regressive subcutaneous emphysema developed in 3, and I had a muscular hematoma (pectoralis muscle) around a trocar site. One young patient experienced bilateral submammary chest pain after a bilateral video-assisted thoracic sympathectomy. The pain resolved after a 4-month period. Sixty-three patients had prolonged air leak (longer than 5 days) postoperatively. In 60, it occurred after pleurodesis with or without stapling of apical blebs and in 3, after a pulmonary wedge resection. Forty-seven patients required a 6- to 18-day period of drainage because of incomplete lung reexpansion, 26 patients required a new tube thoracostomy, and 6, a needle aspiration. Postoperative pleural effusion after pleurodesis (pleural abrasion by Vicryl) occurred in 7 patients. Three were serous effusions and four, hemothorax. One patient underwent thoracentesis for evacuation and 3, a new VATS procedure for drainage. Three patients had postoperative chylothorax. In the first, it was secondary to removal of a mediastinal tumor by a right-sided VATS procedure. In the second, it occurred after a procedure in the aortopulmonary window to stage a primary lung cancer. In the last patient, a right-sided chylothorax appeared 3 days after a rightsided pleurodesis performed for spontaneous pneumothorax. Postoperative deep vein thrombosis occurred in 4 patients. All had had operation for malignant pleural effusion and had had talc poudrage of the pleural cavity. One of them died of a massive pulmonary embolism. One patient experienced acute renal failure after chemical pleurodesis (iodine). Comment With the advent of video-assisted techniques and endoscopic equipment that allow the performance of standard thoracic surgical procedures in a less invasive way, it seems to us important to look closely at the complications that result from these procedures. First, we emphasize that all the surgeons had similar indications (except the one who performed video-assisted lobectomies) and similar patients. The endoscopic equipment was similar or equivalent for all surgeons, and all had similar surgical backgrounds and frequent training updates. All procedures performed by VATS techniques were equivalent or similar to those accomplished by open procedures. This review of our experience includes our learning curve, and some of the complications were registered during the initial period of our VATS practice. Now, with the experience of the surgeons and the anesthesiologists,

4 536 JANCOVICI ET AL Ann Thorac Surg COMPLICATION OF VATS 1996;61:533-7 it is easier to determine whether a VATS procedure is appropriate for the patient and technically feasible for the surgeon. Our overall perioperative mortality rate was 0.5% and was mainly due to high-risk patients undergoing talc pleurodesis for malignant pleural effusions. There were no operative deaths, and no death was related to an intraoperative complication. Reported in-hospital mortality rates range between 0% and 5% [1-7], and, as in our experience, the deaths largely involve older patients in poor condition or with malignancy. This prompted us to change our policy toward this group of patients. We now believe that the ideal procedure for them should be performed under local anesthesia using little or no sedation. The high incidence of deep vein thrombosis in this group of patients (4.8%) prompted us to start routine low-molecular-weight heparin therapy (20 to 40 mg/24 h subcutaneously) from admission until discharge. All these patients were routinely mobilized 24 hours after the operation to discharge, and all had intensive nursing. In our experience, 12.4% of patients required conversion to an open procedure. The need of conversion has been reported to be between 4.1% and 24.1%, depending on the procedure [1, 3, 4]. The reasons for converting are usually the need of further parenchymal resection (lobectomy, segmentectomy), inability to find a lesion, too large a lesion, or a lesion in a difficult location for thoracoscopic resection, tight adhesions in the pleural space, equipment failure, and major bleeding. We do not consider conversion a failure. It is necessary when the VATS procedure becomes too dangerous or when we believe the VATS result would be inferior to that achieved by an open procedure. There should be no feeling of defeat when the chest has to be opened; the decision to do so usually requires a wide experience of thoracic surgery. Moreover, our approach to primary lung cancer is to convert to a lobectomy or a pneumonectomy by an open procedure. However, when resections are satisfactory, a wedge resection could be suitable for high-risk patients in poor general or cardiopulmonary condition who could not tolerate thoracotorny [8]. Only one of us (J.A.) performed VATS lobectomies for T1 NO or T2 NO primary lung cancers. Major bleeding was the most serious intraoperative complication in our experience. It was related mainly to video-assisted iobectomies and especially to pulmonary artery wounds. Bleeding occurred during division or stapling of a superior mediastinal artery or a severed artery. In this situation, applying the stapler to the vessel often creates tension along the branch, which can lead to avulsion. Rarely, bleeding occurred because of equipment failure: the endostapler cuts but fails to staple the vessel. In I patient, a massive hemorrhage resulted from a severing of the left pulmonary vein. Temporary hemostasis was obtained by compression using the endostapler, while the assistant surgeon performed an emergency posterolateral thoracotomy. Such a potentially lifethreatening complication has recently been emphasized by Craig and associates [9], who reported a failure rate of 0.82% for endovascular stapling. We have encountered bleeding resulting from intercostal vessel injuries, but at no time did it require conversion to thoracotomy. Two patients undergoing video-assisted thoracic sympathectomy experienced such problems during division of the sympathetic chain as we tried to dissect it free from a superior intercostal vein that crossed it. The hemorrhage was controlled by application of surgical endoclips or by electrocauterization. Tracheobronchial tree wounds were encountered exclusively during video-assisted procedures for resection of bronchogenic cysts (2/19) and occurred during dissection or resection of tightly adherent cysts. Such wounds are common to surgical intervention for bronchogenic cysts, and surgeons have reported an incidence of up to 10% in patients operated on by an open technique [10, 11]. In difficult cases when the cyst is tightly attached to the tracheobronchial or the esophageal wall, we think that a small part of the cyst wall can be left in place, after electrocauterization of the mucosal lining, to avoid a tracheobronchial wound during resection. If this complication occurs, it requires conversion to an open procedure because of the difficulty of obtaining a safe repair by a "closed" technique. An unsuccessful or incomplete repair performed with a VATS technique may result in such dramatic infectious complications as mediastinitis or empyema. Conversion to thoracotomy allows a direct suture and application of biologic glue to the suture as well as irrigation of the pleural cavity with antiseptic solutions. We believe this procedure is safer and less hazardous for the patient. Early in our experience, we noted two chest wall neoplastic localizations on a trocar site a few months after wedge resection for a malignant pulmonary nodule. Both patients had bronchogenic carcinoma and underwent reoperation for extensive chest wall resection. The incidence of this complication is 1.4% (2/148) in our patients who had operation for resection or biopsy of a pulmonary nodule, and this finding has recently been emphasized by other groups [12, 13]. This dramatic complication seems to be related exclusively to the lack of an endoscopic bag during extraction of the nodule through the chest wall. Although there may be no way to prove it, the evidence implicating the extraction of the tumor through the 2-cm orifice as the source of the chest wall seeding is convincing. Because we now routinely use endoscopic bags for removal of all pulmonary nodules or mediastinal masses, we have not encountered this complication again. Prolonged postoperative chest pain is a well-known complication of open thoracic operations [14]. Many surgeons [3, 15] have reported that in their VATS experience, patients had less postoperative pain, but the authors did not quantify this. Recently, Landreneau and co-workers [16] found that patients undergoing VATS operations experienced less postoperative pain, less shoulder dysfunction, and less pulmonary dysfunction than patients undergoing lateral thoracotomy. We agree with their findings because only 1 of our patients sustained prolonged chest pain after a VATS procedure. This patient had undergone operation for primary hyperhidrosis of the upper limbs and had had a bilateral video-assisted thoracic sympathectomy. She had bilat-

5 Ann Thorac Surg JANCOVICI ET AL ;61:533-7 COMPLICATION OF VATS eral submammary pain for 4 months and required daily analgesic use. Excluding this patient, no substantial longterm pain was encountered after a VATS procedure, and all patients were asymptomatic a few weeks after the operation. The prevalence of postoperative arrhythmia is very, low in our experience. Except for 1 patient in very poor condition who died of a ventricular arrhythmia a few days after talc pleurodesis for malignant pleural effusion, no patient required new antiarrhythmic drugs for perioperative arrhythmias. Arrhythmia is a commonly reported complication after thoracic surgical procedures, especially after pulmonary resections or procedures involving the pericardium [17]. However, the incidence of such complications seems very low after VATS procedures in our experience as well as the experience of other groups [1, 6, 18]. We think that even patients with such known risk factors for arrhythmia as lung cancer, mediastinal lymph node dissection, and age older than 70 years do not require any prophylactic antiarrhythmic treatment when they undergo an operation using a VATS technique. Prolonged air leak was the most common postoperative complication in our experience. It occurred in 6.7 /, of the patients and was related mainly to the great number of procedures performed for pulmonary bullous emphysema or spontaneous pneumothorax (42.5%). Fifteen percent of our patients operated on for this indication had Vicryl abrasion for pleurodesis with or without bullectomy and had postoperative air leak for 6 to 18 days. The incidence of this complication in our experience is higher than that recently reported by others [19, 20]. This is probably related to the great number of bleb electrocauterizations that we performed during the early part of our VATS practice. This technique has proved to be less efficient than the endostapler, and since we abandoned it, we have lessened significantly the incidence of prolonged air leaks. Despite a high incidence of such problems, our reoperation rate for prolonged air leak is 0%, which is not dissimilar to that reported by others [19]. The incidence of infectious complications was 1% in our experience. This is not dissimilar from the incidences previously reported by other groups [1, 6, 7, 20]. Such complications seem to be more frequent after videoassisted lobectomy (7.3% in our experience) than after standard VATS procedures, and we think they can be prevented by adequate positioning of chest tubes, early respiratory physical therapy, and perioperative prophylactic antibiotic therapy. In conclusion, on the basis of our experience with 937 VATS procedures performed consecutively over a 5-year period, we conclude that the incidence of complications is acceptable. Excepting prolonged postoperative air leak, the incidence of minor complications decreased compared with that reported during and after analogous open procedures. We expect that the incidence of complications will continue to decrease as we gain further experience in the choice of indications and specific endoscopic techniques. Dramatic perioperative accidents are rare, but massive bleeding may require immediate conversion to an open procedure to achieve hemostasis. For this reason, we believe only trained thoracic surgeons, who have the skill, expertise, and judgment to know when and how to open the chest, should perform VATS procedures. References 1. Hazelrigg SR, Nunchuck SK, LoCicero J Ili, the Video Assisted Thoracic Surgery Study Group. Video Assisted Thoracic Surge D' Study Group data. Ann Thorac Surg 1993; 56: Lewis RJ, Caccavale RJ, Sisler GE, Mackenzie JW. One hundred consecutive patients undergoing video-assisted thoracic operations. Ann Thorac Surg 1992;54: Kaiser LR, Bavaria JE. Complications of thoracoscopy. Ann Thorac Surg 1993;56: Solaini L, Bagioni P, Grandi U. Role of videoendoscopy in pulmonary surged': present experience. Eur J Cardio-thorac Surg 1995;9: Miller DL, Allen MS, Trastek VF, Deschamps C, Pairolero PC. Videothoracoscopic wedge excision of the lung. Ann Thorac Surg 1992;54: Dowling RD, Keenan RJ, Ferson PF, Landreneau RJ. Videoassisted thoracoscopic resection of pulmonary metastases. Ann Thorac Surg 1993;56: Ohri SK, Os~'al SK, Townsend ER, Fountain SW. Early and late outcome after diagnostic thoracoscopy and talc pleurodesis. Ann Thorac Surg 1992;53: Ginsberg RH. Limited resection in the treatment of stage l non-small cell lung cancer: an overview. Chest 1989; 96(Suppl):50S-1S. 9. Craig SR, Walker WS. Potential complications of vascular stapling in thoracoscopic pulmonary resection. Ann Thorac Surg 1995;59: Ribet ME, Copin MC, Gosselin B. Bronchogenic cysts of the mediastinum. J Thorac Cardiovasc Surg 1995;109: Suen H-C, Mathisen DJ, Grillo HC, et al. Surgical management and radiological characteristics of bronchogenic cysts. Ann Thorac Surg 1993;55: Fry WA, Siddiqui A, Pensler JM, Mostafavi H. Thoracoscopic implantation of cancer with a fatal outcome. Ann Thorac Surg 1995;59: Walsh GL, Nesbitt JC. Tumor implants after thoracoscopic resection of a metastatic sarcoma. Ann Thorac Surg 1995;59: Dajczman E, Gordon A, Kreisman H, Wolkove N. Long-term postthoracotomy pain. Chest 1991;99: Landreneau RJ, Hazelrigg SR, Ferson PF, et al. Thoracoscopic resection of 85 pulmonary, lesions. Ann Thorac Surg 1992;54: Landreneau RJ, Hazelrigg SR, Mack MJ, et al. Postoperative pain-related morbidity: video-assisted thoracic surgery versus thoracotomy. Ann Thorac Surg 1993;56: Asamura H, Naruke T, Tsuchiya R, Goya T, Kondo H, Suemasu K. What are the risk factors for arrhythmias after thoracic operations? A retrospective multivariate analysis of 267 consecutive thoracic operations. J Thorac Cardiovasc Surg 1993;106: McKenna R Jr. Lobectomy by video-assisted thoracic surgery with mediastinal node sampling for lung cancer. J Thorac Cardiovasc Surg 1994;107: Naunheim KS, Mack MJ, Hazelrigg SR, et al. Safety and efficacy of video-assisted thoracic surgical techniques for the treatment of spontaneous pneumothorax. J Thorac Cardiovasc Surg 1995;109: Waller DA, Forty J, Morritt GN. Video-assisted thoracoscopic surgery versus thoracotomy for spontaneous pneumothorax. Ann Thorac Surg 1994;58:372-7.