- Email: [email protected]
Thoracoscopic repair of esophageal atresia andtracheo-esophageal fistula
Seminars in Pediatric Surgery (2005) 14, 2-7
Thoracoscopic repair of esophageal atresia and tracheo-esophageal fistula Steven S. Rothenberg, MD From the Mother and Child Hospital at Presbetyrian/St. Lukes, Denver, Colorado. KEYWORDS Tracheo-esophageal fistula; Esophageal atresia; Thoracoscopy
Advancements in minimally invasive surgical techniques and instruments for neonates have allowed even the most complex neonatal procedures to be approached endoscopically. In 1998 the first successful thoracoscopic repair of an esophageal atresia was performed in a 2-month-old infant. One year later the first totally thoracoscopic repair of an atresia with distal fistula was realized. Over the ensuing 5 years these techniques have become more refined and widespread so that this technique is now being performed all over the world with excellent results, while avoiding the significant short- and long-term morbidity associated with thoracotomy in neonates. © 2005 Elsevier Inc. All rights reserved.
Esophageal atresia (EA) with or without a tracheoesophageal (TEF) fistula is one of the rarer congenital anomalies occurring in 1/5000 births. Traditionally these patients have presented shortly after birth because of an inability to pass an oro-gastric tube, respiratory distress, or an inability to tolerate feeds. The condition maybe associated with other major congenital anomalies (VATER Syndrome), or maybe an isolated defect. Improvements in maternal–fetal ultrasound have resulted in prenatal diagnosis in a number of cases. This allows the surgeon to plan for delivery and eventual surgery. Patients with a TEF require relatively emergent surgical intervention to prevent aspiration of gastric acid and over distension of the intestines. Those with pure atresia can be dealt with in a more leisurely fashion as long the infant’s oral secretions are controlled by continuous or intermittent suction. Recent advancements in technique and instrumentation in pediatric endoscopic surgery have allowed significantly more complex and delicate procedures to be performed, even in small neonates. Over the last 10 years the number and breadth of minimally invasive surgical (MIS) proceAddress reprint requests and correspondence: Steven S. Rothenberg, 1601 E. 19th Ave., Suite 5500, Denver, Colorado 80218. E-mail address: [email protected].
1055-8586/$ -see front matter © 2005 Elsevier Inc. All rights reserved. doi:10.1053/j.sempedsurg.2004.10.020
dures performed in infants has increased dramatically.1-4 However, one procedure, successful ligation of a TEF with repair of the esophageal atresia, has remained relatively elusive. In 1999 a stepping-stone was laid when a successful thoracoscopic repair of a pure esophageal atresia was completed in a 2-month-old male.5 In 2000 we reported on the first successful repair of an esophageal atresia with TEF in a newborn using a completely thoracoscopic approach6 and 2 years later reported on the first significant series.7 These milestones allowed for a more widespread adoption of these techniques so that numerous pediatric surgical units around the world are now performing minimally invasive TEF repair. This paper will review the technique as well as the author’s personal experience.
Indications All patients with esophageal atresia and/or TEF requiring repair were considered for thoracoscopic repair. The decision to approach these patients endoscopically depends on a number of different factors, the most important being the stability of the infant and the experience of the surgeon and anesthesiologists with thoracoscopy in neonates. The infant
Rothenberg
Thoracoscopic Repair
first needs to undergo a thorough evaluation to determine if there are any other congenital anomalies. The most relevant is an echocardiogram to determine the anatomy and function of the heart as well as the location of the aortic arch. There are very few absolute contra-indications to a thoracoscopic approach. Currently these would include severe hemodynamic instability requiring significant pressor support and significant prematurity (⬍1500 g). Relative contraindications include significant congenital cardiac defects, smaller size (1500-2000 g), or significant abdominal distension. Congenital heart defects are not an absolute contraindication to a thoracoscopic approach and repairs have been successfully completed in patients with ASD, VSD, and Tetralogy of Falot. However, patients with severe cardiac anomalies may not be able to tolerate even the short periods of single lung ventilation which are required to ligate the fistula. It is also important to determine the side of the aortic arch as this will determine the surgical approach. A right-sided arch will necessitate a left-sided approach which can be done with minimal added difficulty. Other anomalies which may change the surgical approach include intestinal atresias. Imperforate anus or cloacal anomalies are not a contra-indication and a thoracoscopic TEF repair can be combined with a laparoscopic evaluation/repair of these defects. However, a proximal bowel atresia may result in significant early gastric distension which may cause immediate problems. The surgeon must evaluate each situation separately and determine the most appropriate approach. Much of the decision will depend on the surgeon’s and the anesthesiologist’s experience with thoracoscopy in neonates.
Advantages and disadvantages of a thoracoscopic approach There are two main advantages to a thoracoscopic approach. The first is the avoidance of a postero-lateral thoracotomy incision and all of its inherent morbidity, including pain and respiratory compromise in the postoperative period, and scoliosis, shoulder girdle weakness, and chest wall asymmetry in the long term. The second is the improved visualization afforded by the endoscopic approach. The scope provides a magnified view superior to that obtained with surgical loops. It also gives the surgeon a superior view of the fistula, upper and lower esophageal segments, the plane between the upper pouch and membranous trachea, the vagus nerve, and other vital structures. The primary disadvantages are the small working space and the fine manipulations, especially suturing the anastomosis, required in this confined area. Also, currently the sutures can only be placed one at a time, creating a great deal of tension on the initial sutures (see “Technique”) increasing the risk of the sutures tearing out. It is also necessary to achieve some degree of single lung ventilation for the procedure, which can create both anesthetic and technical problems.
3 Some question whether or not the fact that the thoracoscopic approach is trans-pleural rather then retro-pleural is a disadvantage. This would assume that the risk of infectious complications (mediastinitis/empyema) is less with the retro-pleural approach. Although this theoretically is true, the use of appropriately placed chest drains and antibiotics should negate any perceived advantage.
Preoperative preparation Preparation does not differ from that for the open procedure. A suction catheter should be placed in the upper blind pouch and the patient should be placed in a head-up position to prevent reflux of gastric contents thru the fistula. Broad spectrum antibiotics should be given and the workup for other congenital anomalies completed. Mechanical ventilation should be avoided until induction of anesthesia if at all possible. This will minimize the amount of abdominal distension caused by air passing thru the fistula. Once the baby has been stabilized, surgery should be performed in a prompt time frame to minimize the risk of aspiration. In the case of pure atresia surgery can be performed in a more elective manner. Often it will be necessary to place a gastrostomy tube first. The decision to proceed with the esophageal repair will depend on the gap between the two segments. If this is not clear, thoracoscopy may offer a less invasive way of evaluating the two ends and the gap between them.
Instrumentation Instruments consist of standard laparoscopic/thoracoscopic instruments for neonates (short 18 to 20 cm in length, 2.5 to 3.0 mm in diameter). Three to four 3.0- and/or 5.0-mm ports are necessary and size used depends on the size of instruments and scope used. In general at least one 5-mm port is necessary for introduction of the endoscopic clip applier and suture. A 3- to 5-mm 30°-wide angle telescope is used and should be approximately 20 cm in length. Some prefer a more angled scope and this is strictly dependent on surgeon preference. However, a 0° scope should be avoided as the limited space of the thorax will create significant instrument and scope “dueling.” Specific instrumentation is also surgeon-dependent but in general a basic set is all that is needed. This should include a curved dissector (Maryland), curved dissecting scissors (insulated), a needle driver, a 5-mm endoscopic clip applier, 4-0 or 5-0 PDS suture on a TF needle. Other instruments can include a hook cautery, bipolar forceps or Ligasure 5-mm dissector, and a knot pusher.
Technique General endotracheal anesthesia is administered but low peak pressures should be used until the fistula is ligated to prevent over distension of the abdomen. Local anesthetic
4
Seminars in Pediatric Surgery, Vol 14, No 1, February 2005
Figure 2 Figure 1 The patient is positioned in a modified prone position with the left side elevated approximately 30°. This gives access to the area between the anterior and posterior axillary lines for port placement while exposing the posterior mediastinum.
Standard room set-up.
selves so that they are in the most ergonomic and comfortable position.
Port placement (0.25% marcaine) is inserted at the trocar sites. An attempt should be made to obtain a left mainstem intubation by blind manipulation of the endotracheal tube. However, if this cannot be achieved easily the endotracheal tube should be left in the trachea just above the carina. Extra time should not be wasted trying to manipulate the tube down the left side as excellent right lung collapse can be achieved with CO2 insufflation alone. Wasting even minutes trying to place a bronchial blocker or other manipulations can compromise the eventual success of a thoracoscopic approach. The decision to perform preoperative bronchoscopy is controversial and the surgeon can follow the same guidelines he does for open TOF repair. A urinary catheter is optional.
Positioning Once the endotracheal tube is secure the patient is placed in a modified prone position with the right side elevated approximately 30-45° (Figure 1). If there is a right-sided arch then the left side is approached. This positioning gives the surgeon access to the area between the anterior and posterior axillary line for trocar placement while allowing gravity to retract the lung away from the posterior mediastinum. This arrangement gives excellent exposure of the fistula and esophageal segments without the need of an extra trocar for a lung retractor. Generally small rolls are sufficient to provide stabilization for positioning or a small bean bag can be used. The surgeon and the assistant stand at the patient’s front and the monitor is placed at the patient’s back (Figure 2). This allows the surgeon and the assistant to work in line with the camera toward the point of dissection. The assistant should not be placed on the opposite side of the table as this will place him at a complete paradox with the telescope. The scrub nurse can be on either side of the patient, depending on the room layout. Because of the fine manipulation necessary the surgeon and the assistant should position them-
Port placement is extremely important because of the small chest cavity and the intricate nature of the dissection and reconstruction. The procedure can be performed with three ports but occasionally a fourth port is necessary to retract the lung (Figure 3). The initial port (3-5 mm) is placed in the fifth intercostals space at approximately the posterior axillary line. This is the camera port and gives the surgeon excellent visualization of the posterior mediastinum in the area of the fistula and eventual anastomosis. As mentioned, a 30° lens is used to allow the surgeon to “look down” on his instruments and avoid “dueling.” The two instrument ports are placed in the mid-axillary line one to two interspaces above and below the camera port. The upper port is 5 mm to allow for the clip applier and suture. The lower port is 3 mm in size. Ideally, these ports are placed so that the instrument tips will approximate a right angle (90°) at the level of the
Figure 3 Port placement. Three-millimeter ports are used for instruments and the telescope and a single 5-mm port is used for introduction of the endoclip and suture.
Rothenberg
Thoracoscopic Repair
5
Figure 5 Figure 4
The completed anastomosis.
Exposure of the fistula for ligation.
fistula. This positioning will facilitate suturing the anastomosis. A fourth port can be placed either higher or lower in the thoracic cavity to help retract the lung but this has not been necessary in the majority of cases. The operation then follows the same pattern of the open procedure.
Ligating the fistula Once the chest has been insufflated and the lung collapsed the surgeon must identify the fistula. In most cases the fistula is attached to the membranous portion of the trachea just above the carina. This level is usually demarcated by the Azygus vein. After the azygus is identified it should be mobilized for a short segment using a curved dissector or scissors. The vein is then cauterized and divided. It is often easiest to do this with a small hook cautery although bipolar cautery or other sealing devices can be used. Ties or clips are generally not necessary and could interfere with the dissection of the fistula. With the vein divided, the lower esophageal segment is identified and followed proximally to the fistula. Because of the magnification afforded by the thoracoscopic approach it is easy to visualize exactly where the fistula enters the back wall of the trachea. A 5-mm endo clip can then be applied safely (Figure 4). Care should be taken to avoid the vagus nerve. A single clip is usually sufficient. The fistula can then be safely divided with scissors. The distal segment may retract making it difficult to visualize; therefore, it may be preferable to wait until the upper pouch is dissected out before completely dividing the fistula. The fistula may also be suture ligated but this requires delicate suturing at a time when an air leak from
the divided fistula may cause increased respiratory compromise.
Mobilize the upper pouch Attention is now turned to the thoracic inlet. The anesthesiologist places pressure on the NG tube to help identify the upper pouch. The pleura overlying the pouch is incised sharply and the pouch is mobilized with blunt and sharp dissection. In some cases it is helpful to place a stay suture in the tip of the pouch to aid in applying traction. The plane between the esophagus and trachea can be seen well and the two should be separated by sharp dissection. Mobilization of the upper pouch is carried on up into the thoracic inlet. Once adequate mobilization is achieved the distal tip of the pouch is resected. This should be an adequate section so that there is a sufficient opening to prevent later stricture formation.
Figure 6
Skin incisions 1 year post repair.
6
Seminars in Pediatric Surgery, Vol 14, No 1, February 2005
The anastomosis
Results
With the two ends mobilized the anastomosis is performed using a 4-0 or 5-0 mono-filament absorbable suture on a small taper needle. The sutures are placed one at a time in an interrupted fashion. The back wall is placed first (3 to 5 sutures) with the knots intraluminal. The Ng tube is then passed under direct vision into the lower pouch and on into the stomach. The anterior wall is the completed with the Ng tube acting as a guide to prevent incorporation of the posterior wall and ensuring patency of the anastomosis (Figure 5). Adequate bites should be placed to prevent the sutures from tearing out and just as in the open procedure it is important to get mucosa with all bites incorporated. The anastomosis generally requires only seven to eight sutures. Once the anastomosis is completed a chest tube is then placed through the lower trocar site and the tip is placed near the anastomosis (under direct vision with the endoscope). The other ports are removed and the sites are closed with absorbable sutures.
Twenty-seven of 28 procedures were completed successfully thoracoscopically. The average operative time was 95 minutes (range 55-120 minutes). All but 3 patients were extubated by the second postoperative day. One patient required re-intubation 4 hours postextubation for increasing respiratory distress (patient 1). He was re-extubated on pod #3. The fifth patient had an extremely long gap (4 vertebral bodies) with significant tension on the anastomosis. He was kept sedated and intubated for 4 days. Esophageal contrast studies were obtained on pod #5 in 27 patients and the anastomosis was patent with no evidence of a leak in 26 of 27. The first patient repaired with this technique had clinical evidence of a leak—saliva in the chest tube— on pod #4. He was kept NPO and drainage stopped after 24 hours. He was studied on day 8 with no evidence of a leak. One patient with a long gap pure atresia had a small leak which had sealed on a repeat study on day 9. The one conversion was performed in a patient with an unrecognized distal congenital esophageal stenosis. The anastomosis was successfully completed thoracoscopically but because the NG tube could not be passed into the stomach the case was converted to ensure there was not a false passage. Eight patients have required esophageal dilations and all are now on full oral feeds.
Author’s experience From March 2000 to July 2004, 26 consecutive patients with esophageal atresia and a distal TEF and 2 with pure esophageal atresia were referred to the author for repair. Twelve had been diagnosed prenatally and were delivered at the high-risk, perinatal/neonatal center. Sixteen others were diagnosed postnatally and transferred after birth. Gestational age of the patients ranged from 31 weeks to 40 weeks at the time of delivery. Three other infants with TEF were also referred to the author during this period but were excluded because of size and associated anomalies. These three weighed 800 g, 1100 g with a Tetralogy of Falot, and 1800 g with an omphalocele. In the thoracoscopic group patients ranged in weight from 1.8 to 3.8 kg. Preoperative evaluations revealed congenital heart disease in 10, including 1 patient with a Tetralogy of Falot as well as a right-sided aortic arch. Two patients had a high imperforate anus and one had a cloacca. Four patients required intubation before surgery for increasing respiratory distress. The gap length was estimated preoperatively based on the position of the tip of the NG tube and the apparent bifurcation of the trachea as seen on the CXR. This ranged from 2 to 3.5 vertebral bodies. At the time of surgery the longest gap in a patient with TEF was closer to 4 vertebral bodies as this patient had a trifurcation type fistula. The longest gap in a patient with pure atresia was 5 vertebral bodies. Seven other patients with TEF were referred to our practice during the study period and were repaired via a standard thoracotomy by surgeons not currently performing the procedure thoracoscopically. However, all four surgeons in our group are now performing the procedure thoracoscopically.
Discussion Recent advances in minimally invasive surgery (MIS) in infants and children have allowed for a wide expansion of applications over the last few years. Procedures that were previously thought impossible in children, let alone neonates, have become commonplace at a number of pediatric centers. We first reported on our experience in infants less than 5 kg in 1998 and recently reported on our first decade’s experience.4,5 These studies clearly show that complex procedures such as Nissen fundoplications, repair intestinal atresias, bowel resections, and thoracic procedures, such as lung biopsy and PDA ligation, were not only possible but were associated with less morbidity than standard open techniques. Others have reported similar positive findings. Fujimoto and coworkers reported that there is a decreased stress response as measured by interleukin-2 and other stress mediators when MIS techniques are employed in infants, and this correlated with improved clinical outcomes. Concerns that neonates could not tolerate, or would be adversely affected by, abdominal insufflation or single lung ventilation have not been borne out. With these encouraging results the impetus to develop the tools and techniques to perform even the most complicated procedures in neonates became a driving force. Many recent studies have documented the long-term outcome of patients with EA and TEF and the overall morbidity is significant, some related to the initial surgical
Rothenberg
Thoracoscopic Repair
approach and technique.8,9 The benefits of performing a TEF using minimally invasive techniques are obvious but the technical hurdles are many. The greatest advantage is avoiding a postero-lateral thoracotomy in a neonate. This has been shown to be associated with a high degree of scoliosis and shoulder girdle weakness later in development.10 Another is the improved cosmetic result (Figure 6). Bianchi and others have advocated muscle sparing and/or lateral thoracotomies, with various skin incisions to improve on these problems but these incisions can be difficult to develop, offer more limited access, and still require spreading of the rib innerspace.11 While an improvement over a standard thoracotomy, there are still significant issues associated with these techniques and they may not eliminate the risk of scoliosis. An unanticipated benefit of the thoracoscopic approach was the superior visualization of the anatomy and especially the fistula. Since the fistula is visualized perpendicular to its insertion to the membranous trachea, the exact site for ligation could be identified easily, minimizing the residual pouch attached to the trachea. The use of the 5-mm titanium clips has proven to be simple and effective with no evidence of tracheal leak or recurrent fistula. In fact, we have adopted this method of fistula ligation in the few open procedures we have performed in the last year. A recognized advantage after the first case was performing the dissection and anastomosis in situ. Because the separation of the fistula and the upper pouch from the trachea was performed under direct magnified vision from a lateral approach there was little manipulation or force applied to the trachea itself. This may help diminish the degree of tracheo-malacia that these children have postoperatively. Also the plane between the upper pouch and trachea was more obvious, making injury to the membranous wall of the trachea less likely. Another advantage of performing the anastomosis in situ may be less tension on the esophageal ends allowing longer gaps to be brought together without tearing. This appeared to be the case in one patient with a trifurcation fistula and long gap, although he did develop an anastomotic stricture that required dilation. The major technical hurdle in this operation is the suturing of the anastomosis. The placement of the sutures and knot tying are technically demanding and relatively imprecise. Also, as opposed to the open technique where the entire posterior row of sutures can be placed and then brought together to disperse the tension along multiple points during knot tying, this method places all the tension
7 on one suture at a time. So far this has not been a significant problem but it could prove to be. For this procedure to become more widely accepted it maybe necessary to develop a mechanical anastomotic device or self-knotting suture. The rate of anastomotic narrowing requiring at least one dilation in this series is 30% but was initially almost 50%, a figure much higher then our open experience. This may have been secondary to inadequate approximation of the mucosal ends or an insufficient opening being made in the upper pouch. We have modified the technique to eliminate both of these problems as well as switched to a monofilament absorbable suture. These changes seem to have resolved the problem with only 2 of the last 16 patients requiring dilation. Clearly the technical and physiologic hurdles to accomplish this type of repair are many and it will require continued advances before this surgery becomes commonplace. However, the ability to perform this complex reconstruction without a thoracotomy lays further groundwork in minimizing surgical morbidity in even the smallest pediatric patients.
References 1. Rothenberg SS, Chang HT, Bealer JF: Experience with minimally invasive surgery in infants. Am J Surg 176:654-658, 1998 2. Fujimoto T, Segawa O, Lane GJ, et al: Laparoscopic surgery for newborn infants. Surg Endosc 13:773-777, 1999 3. Rothenberg SS, Chang JHT, Toews WH, et al: Thoracscopic closure of patent ductus arteriosus: A less traumatic and more cost-effective technique. J Pediatr Surg 30:1057-1060, 1995 4. Rothenberg SS, Chang JHT, Bealer J: First decades experience with minimally invasive surgery in neonates. Pediatr Endosurg Innovative Tech 8:89-94, 2004 5. Lobe TE, Rothenberg SS, Waldschmidt J, Stroeder L: Thoracoscopic repair of esophageal atresia in an infant: A surgical first. Pediatr Endosurg Innovative Tech 3:141-148, 1999 6. Rothenberg SS: Thoracoscopic repair of a tracheoesophageal in a neonate. Pediatr Endosurg Innovative Tech 4:150-156, 2000 7. Rothenberg SS: Thoracoscopic Repair of tracheo-esophageal fistula and esophageal atresia in newborns. J Pediatr Surg 37:869-872, 2002 8. Little DC, Rescorla FJ, Grosfeld JL, et al: Long-term analysis of children with esophageal atresia and tracheoesophageal fistula. J Pediatr Surg 38:852-856, 2003 9. Konkin DE, O’Hali WA, Webber EM, et al: Outcomes in esophageal atresia and tracheoesophageal fistula. J Pediatr Surg 38:1726-1729, 2003 10. Vaiquez JJ, Murcia J, DiezPardo JA: Morbid musculoskeletal sequelae of thoracotomy for tracheoesophageal fistula. J Pediatr Surg 20:511514, 1995 11. Bianchi A, Sowende O, Aliza NK, et al: Aesthetics and lateral thoracotomy in the neonate. J Pediatr Surg 12:1798-1800, 1998
Thoracoscopic repair of esophageal atresia and tracheo-esophageal fistula Steven S. Rothenberg, MD From the Mother and Child Hospital at Presbetyrian/St. Lukes, Denver, Colorado. KEYWORDS Tracheo-esophageal fistula; Esophageal atresia; Thoracoscopy
Advancements in minimally invasive surgical techniques and instruments for neonates have allowed even the most complex neonatal procedures to be approached endoscopically. In 1998 the first successful thoracoscopic repair of an esophageal atresia was performed in a 2-month-old infant. One year later the first totally thoracoscopic repair of an atresia with distal fistula was realized. Over the ensuing 5 years these techniques have become more refined and widespread so that this technique is now being performed all over the world with excellent results, while avoiding the significant short- and long-term morbidity associated with thoracotomy in neonates. © 2005 Elsevier Inc. All rights reserved.
Esophageal atresia (EA) with or without a tracheoesophageal (TEF) fistula is one of the rarer congenital anomalies occurring in 1/5000 births. Traditionally these patients have presented shortly after birth because of an inability to pass an oro-gastric tube, respiratory distress, or an inability to tolerate feeds. The condition maybe associated with other major congenital anomalies (VATER Syndrome), or maybe an isolated defect. Improvements in maternal–fetal ultrasound have resulted in prenatal diagnosis in a number of cases. This allows the surgeon to plan for delivery and eventual surgery. Patients with a TEF require relatively emergent surgical intervention to prevent aspiration of gastric acid and over distension of the intestines. Those with pure atresia can be dealt with in a more leisurely fashion as long the infant’s oral secretions are controlled by continuous or intermittent suction. Recent advancements in technique and instrumentation in pediatric endoscopic surgery have allowed significantly more complex and delicate procedures to be performed, even in small neonates. Over the last 10 years the number and breadth of minimally invasive surgical (MIS) proceAddress reprint requests and correspondence: Steven S. Rothenberg, 1601 E. 19th Ave., Suite 5500, Denver, Colorado 80218. E-mail address: [email protected].
1055-8586/$ -see front matter © 2005 Elsevier Inc. All rights reserved. doi:10.1053/j.sempedsurg.2004.10.020
dures performed in infants has increased dramatically.1-4 However, one procedure, successful ligation of a TEF with repair of the esophageal atresia, has remained relatively elusive. In 1999 a stepping-stone was laid when a successful thoracoscopic repair of a pure esophageal atresia was completed in a 2-month-old male.5 In 2000 we reported on the first successful repair of an esophageal atresia with TEF in a newborn using a completely thoracoscopic approach6 and 2 years later reported on the first significant series.7 These milestones allowed for a more widespread adoption of these techniques so that numerous pediatric surgical units around the world are now performing minimally invasive TEF repair. This paper will review the technique as well as the author’s personal experience.
Indications All patients with esophageal atresia and/or TEF requiring repair were considered for thoracoscopic repair. The decision to approach these patients endoscopically depends on a number of different factors, the most important being the stability of the infant and the experience of the surgeon and anesthesiologists with thoracoscopy in neonates. The infant
Rothenberg
Thoracoscopic Repair
first needs to undergo a thorough evaluation to determine if there are any other congenital anomalies. The most relevant is an echocardiogram to determine the anatomy and function of the heart as well as the location of the aortic arch. There are very few absolute contra-indications to a thoracoscopic approach. Currently these would include severe hemodynamic instability requiring significant pressor support and significant prematurity (⬍1500 g). Relative contraindications include significant congenital cardiac defects, smaller size (1500-2000 g), or significant abdominal distension. Congenital heart defects are not an absolute contraindication to a thoracoscopic approach and repairs have been successfully completed in patients with ASD, VSD, and Tetralogy of Falot. However, patients with severe cardiac anomalies may not be able to tolerate even the short periods of single lung ventilation which are required to ligate the fistula. It is also important to determine the side of the aortic arch as this will determine the surgical approach. A right-sided arch will necessitate a left-sided approach which can be done with minimal added difficulty. Other anomalies which may change the surgical approach include intestinal atresias. Imperforate anus or cloacal anomalies are not a contra-indication and a thoracoscopic TEF repair can be combined with a laparoscopic evaluation/repair of these defects. However, a proximal bowel atresia may result in significant early gastric distension which may cause immediate problems. The surgeon must evaluate each situation separately and determine the most appropriate approach. Much of the decision will depend on the surgeon’s and the anesthesiologist’s experience with thoracoscopy in neonates.
Advantages and disadvantages of a thoracoscopic approach There are two main advantages to a thoracoscopic approach. The first is the avoidance of a postero-lateral thoracotomy incision and all of its inherent morbidity, including pain and respiratory compromise in the postoperative period, and scoliosis, shoulder girdle weakness, and chest wall asymmetry in the long term. The second is the improved visualization afforded by the endoscopic approach. The scope provides a magnified view superior to that obtained with surgical loops. It also gives the surgeon a superior view of the fistula, upper and lower esophageal segments, the plane between the upper pouch and membranous trachea, the vagus nerve, and other vital structures. The primary disadvantages are the small working space and the fine manipulations, especially suturing the anastomosis, required in this confined area. Also, currently the sutures can only be placed one at a time, creating a great deal of tension on the initial sutures (see “Technique”) increasing the risk of the sutures tearing out. It is also necessary to achieve some degree of single lung ventilation for the procedure, which can create both anesthetic and technical problems.
3 Some question whether or not the fact that the thoracoscopic approach is trans-pleural rather then retro-pleural is a disadvantage. This would assume that the risk of infectious complications (mediastinitis/empyema) is less with the retro-pleural approach. Although this theoretically is true, the use of appropriately placed chest drains and antibiotics should negate any perceived advantage.
Preoperative preparation Preparation does not differ from that for the open procedure. A suction catheter should be placed in the upper blind pouch and the patient should be placed in a head-up position to prevent reflux of gastric contents thru the fistula. Broad spectrum antibiotics should be given and the workup for other congenital anomalies completed. Mechanical ventilation should be avoided until induction of anesthesia if at all possible. This will minimize the amount of abdominal distension caused by air passing thru the fistula. Once the baby has been stabilized, surgery should be performed in a prompt time frame to minimize the risk of aspiration. In the case of pure atresia surgery can be performed in a more elective manner. Often it will be necessary to place a gastrostomy tube first. The decision to proceed with the esophageal repair will depend on the gap between the two segments. If this is not clear, thoracoscopy may offer a less invasive way of evaluating the two ends and the gap between them.
Instrumentation Instruments consist of standard laparoscopic/thoracoscopic instruments for neonates (short 18 to 20 cm in length, 2.5 to 3.0 mm in diameter). Three to four 3.0- and/or 5.0-mm ports are necessary and size used depends on the size of instruments and scope used. In general at least one 5-mm port is necessary for introduction of the endoscopic clip applier and suture. A 3- to 5-mm 30°-wide angle telescope is used and should be approximately 20 cm in length. Some prefer a more angled scope and this is strictly dependent on surgeon preference. However, a 0° scope should be avoided as the limited space of the thorax will create significant instrument and scope “dueling.” Specific instrumentation is also surgeon-dependent but in general a basic set is all that is needed. This should include a curved dissector (Maryland), curved dissecting scissors (insulated), a needle driver, a 5-mm endoscopic clip applier, 4-0 or 5-0 PDS suture on a TF needle. Other instruments can include a hook cautery, bipolar forceps or Ligasure 5-mm dissector, and a knot pusher.
Technique General endotracheal anesthesia is administered but low peak pressures should be used until the fistula is ligated to prevent over distension of the abdomen. Local anesthetic
4
Seminars in Pediatric Surgery, Vol 14, No 1, February 2005
Figure 2 Figure 1 The patient is positioned in a modified prone position with the left side elevated approximately 30°. This gives access to the area between the anterior and posterior axillary lines for port placement while exposing the posterior mediastinum.
Standard room set-up.
selves so that they are in the most ergonomic and comfortable position.
Port placement (0.25% marcaine) is inserted at the trocar sites. An attempt should be made to obtain a left mainstem intubation by blind manipulation of the endotracheal tube. However, if this cannot be achieved easily the endotracheal tube should be left in the trachea just above the carina. Extra time should not be wasted trying to manipulate the tube down the left side as excellent right lung collapse can be achieved with CO2 insufflation alone. Wasting even minutes trying to place a bronchial blocker or other manipulations can compromise the eventual success of a thoracoscopic approach. The decision to perform preoperative bronchoscopy is controversial and the surgeon can follow the same guidelines he does for open TOF repair. A urinary catheter is optional.
Positioning Once the endotracheal tube is secure the patient is placed in a modified prone position with the right side elevated approximately 30-45° (Figure 1). If there is a right-sided arch then the left side is approached. This positioning gives the surgeon access to the area between the anterior and posterior axillary line for trocar placement while allowing gravity to retract the lung away from the posterior mediastinum. This arrangement gives excellent exposure of the fistula and esophageal segments without the need of an extra trocar for a lung retractor. Generally small rolls are sufficient to provide stabilization for positioning or a small bean bag can be used. The surgeon and the assistant stand at the patient’s front and the monitor is placed at the patient’s back (Figure 2). This allows the surgeon and the assistant to work in line with the camera toward the point of dissection. The assistant should not be placed on the opposite side of the table as this will place him at a complete paradox with the telescope. The scrub nurse can be on either side of the patient, depending on the room layout. Because of the fine manipulation necessary the surgeon and the assistant should position them-
Port placement is extremely important because of the small chest cavity and the intricate nature of the dissection and reconstruction. The procedure can be performed with three ports but occasionally a fourth port is necessary to retract the lung (Figure 3). The initial port (3-5 mm) is placed in the fifth intercostals space at approximately the posterior axillary line. This is the camera port and gives the surgeon excellent visualization of the posterior mediastinum in the area of the fistula and eventual anastomosis. As mentioned, a 30° lens is used to allow the surgeon to “look down” on his instruments and avoid “dueling.” The two instrument ports are placed in the mid-axillary line one to two interspaces above and below the camera port. The upper port is 5 mm to allow for the clip applier and suture. The lower port is 3 mm in size. Ideally, these ports are placed so that the instrument tips will approximate a right angle (90°) at the level of the
Figure 3 Port placement. Three-millimeter ports are used for instruments and the telescope and a single 5-mm port is used for introduction of the endoclip and suture.
Rothenberg
Thoracoscopic Repair
5
Figure 5 Figure 4
The completed anastomosis.
Exposure of the fistula for ligation.
fistula. This positioning will facilitate suturing the anastomosis. A fourth port can be placed either higher or lower in the thoracic cavity to help retract the lung but this has not been necessary in the majority of cases. The operation then follows the same pattern of the open procedure.
Ligating the fistula Once the chest has been insufflated and the lung collapsed the surgeon must identify the fistula. In most cases the fistula is attached to the membranous portion of the trachea just above the carina. This level is usually demarcated by the Azygus vein. After the azygus is identified it should be mobilized for a short segment using a curved dissector or scissors. The vein is then cauterized and divided. It is often easiest to do this with a small hook cautery although bipolar cautery or other sealing devices can be used. Ties or clips are generally not necessary and could interfere with the dissection of the fistula. With the vein divided, the lower esophageal segment is identified and followed proximally to the fistula. Because of the magnification afforded by the thoracoscopic approach it is easy to visualize exactly where the fistula enters the back wall of the trachea. A 5-mm endo clip can then be applied safely (Figure 4). Care should be taken to avoid the vagus nerve. A single clip is usually sufficient. The fistula can then be safely divided with scissors. The distal segment may retract making it difficult to visualize; therefore, it may be preferable to wait until the upper pouch is dissected out before completely dividing the fistula. The fistula may also be suture ligated but this requires delicate suturing at a time when an air leak from
the divided fistula may cause increased respiratory compromise.
Mobilize the upper pouch Attention is now turned to the thoracic inlet. The anesthesiologist places pressure on the NG tube to help identify the upper pouch. The pleura overlying the pouch is incised sharply and the pouch is mobilized with blunt and sharp dissection. In some cases it is helpful to place a stay suture in the tip of the pouch to aid in applying traction. The plane between the esophagus and trachea can be seen well and the two should be separated by sharp dissection. Mobilization of the upper pouch is carried on up into the thoracic inlet. Once adequate mobilization is achieved the distal tip of the pouch is resected. This should be an adequate section so that there is a sufficient opening to prevent later stricture formation.
Figure 6
Skin incisions 1 year post repair.
6
Seminars in Pediatric Surgery, Vol 14, No 1, February 2005
The anastomosis
Results
With the two ends mobilized the anastomosis is performed using a 4-0 or 5-0 mono-filament absorbable suture on a small taper needle. The sutures are placed one at a time in an interrupted fashion. The back wall is placed first (3 to 5 sutures) with the knots intraluminal. The Ng tube is then passed under direct vision into the lower pouch and on into the stomach. The anterior wall is the completed with the Ng tube acting as a guide to prevent incorporation of the posterior wall and ensuring patency of the anastomosis (Figure 5). Adequate bites should be placed to prevent the sutures from tearing out and just as in the open procedure it is important to get mucosa with all bites incorporated. The anastomosis generally requires only seven to eight sutures. Once the anastomosis is completed a chest tube is then placed through the lower trocar site and the tip is placed near the anastomosis (under direct vision with the endoscope). The other ports are removed and the sites are closed with absorbable sutures.
Twenty-seven of 28 procedures were completed successfully thoracoscopically. The average operative time was 95 minutes (range 55-120 minutes). All but 3 patients were extubated by the second postoperative day. One patient required re-intubation 4 hours postextubation for increasing respiratory distress (patient 1). He was re-extubated on pod #3. The fifth patient had an extremely long gap (4 vertebral bodies) with significant tension on the anastomosis. He was kept sedated and intubated for 4 days. Esophageal contrast studies were obtained on pod #5 in 27 patients and the anastomosis was patent with no evidence of a leak in 26 of 27. The first patient repaired with this technique had clinical evidence of a leak—saliva in the chest tube— on pod #4. He was kept NPO and drainage stopped after 24 hours. He was studied on day 8 with no evidence of a leak. One patient with a long gap pure atresia had a small leak which had sealed on a repeat study on day 9. The one conversion was performed in a patient with an unrecognized distal congenital esophageal stenosis. The anastomosis was successfully completed thoracoscopically but because the NG tube could not be passed into the stomach the case was converted to ensure there was not a false passage. Eight patients have required esophageal dilations and all are now on full oral feeds.
Author’s experience From March 2000 to July 2004, 26 consecutive patients with esophageal atresia and a distal TEF and 2 with pure esophageal atresia were referred to the author for repair. Twelve had been diagnosed prenatally and were delivered at the high-risk, perinatal/neonatal center. Sixteen others were diagnosed postnatally and transferred after birth. Gestational age of the patients ranged from 31 weeks to 40 weeks at the time of delivery. Three other infants with TEF were also referred to the author during this period but were excluded because of size and associated anomalies. These three weighed 800 g, 1100 g with a Tetralogy of Falot, and 1800 g with an omphalocele. In the thoracoscopic group patients ranged in weight from 1.8 to 3.8 kg. Preoperative evaluations revealed congenital heart disease in 10, including 1 patient with a Tetralogy of Falot as well as a right-sided aortic arch. Two patients had a high imperforate anus and one had a cloacca. Four patients required intubation before surgery for increasing respiratory distress. The gap length was estimated preoperatively based on the position of the tip of the NG tube and the apparent bifurcation of the trachea as seen on the CXR. This ranged from 2 to 3.5 vertebral bodies. At the time of surgery the longest gap in a patient with TEF was closer to 4 vertebral bodies as this patient had a trifurcation type fistula. The longest gap in a patient with pure atresia was 5 vertebral bodies. Seven other patients with TEF were referred to our practice during the study period and were repaired via a standard thoracotomy by surgeons not currently performing the procedure thoracoscopically. However, all four surgeons in our group are now performing the procedure thoracoscopically.
Discussion Recent advances in minimally invasive surgery (MIS) in infants and children have allowed for a wide expansion of applications over the last few years. Procedures that were previously thought impossible in children, let alone neonates, have become commonplace at a number of pediatric centers. We first reported on our experience in infants less than 5 kg in 1998 and recently reported on our first decade’s experience.4,5 These studies clearly show that complex procedures such as Nissen fundoplications, repair intestinal atresias, bowel resections, and thoracic procedures, such as lung biopsy and PDA ligation, were not only possible but were associated with less morbidity than standard open techniques. Others have reported similar positive findings. Fujimoto and coworkers reported that there is a decreased stress response as measured by interleukin-2 and other stress mediators when MIS techniques are employed in infants, and this correlated with improved clinical outcomes. Concerns that neonates could not tolerate, or would be adversely affected by, abdominal insufflation or single lung ventilation have not been borne out. With these encouraging results the impetus to develop the tools and techniques to perform even the most complicated procedures in neonates became a driving force. Many recent studies have documented the long-term outcome of patients with EA and TEF and the overall morbidity is significant, some related to the initial surgical
Rothenberg
Thoracoscopic Repair
approach and technique.8,9 The benefits of performing a TEF using minimally invasive techniques are obvious but the technical hurdles are many. The greatest advantage is avoiding a postero-lateral thoracotomy in a neonate. This has been shown to be associated with a high degree of scoliosis and shoulder girdle weakness later in development.10 Another is the improved cosmetic result (Figure 6). Bianchi and others have advocated muscle sparing and/or lateral thoracotomies, with various skin incisions to improve on these problems but these incisions can be difficult to develop, offer more limited access, and still require spreading of the rib innerspace.11 While an improvement over a standard thoracotomy, there are still significant issues associated with these techniques and they may not eliminate the risk of scoliosis. An unanticipated benefit of the thoracoscopic approach was the superior visualization of the anatomy and especially the fistula. Since the fistula is visualized perpendicular to its insertion to the membranous trachea, the exact site for ligation could be identified easily, minimizing the residual pouch attached to the trachea. The use of the 5-mm titanium clips has proven to be simple and effective with no evidence of tracheal leak or recurrent fistula. In fact, we have adopted this method of fistula ligation in the few open procedures we have performed in the last year. A recognized advantage after the first case was performing the dissection and anastomosis in situ. Because the separation of the fistula and the upper pouch from the trachea was performed under direct magnified vision from a lateral approach there was little manipulation or force applied to the trachea itself. This may help diminish the degree of tracheo-malacia that these children have postoperatively. Also the plane between the upper pouch and trachea was more obvious, making injury to the membranous wall of the trachea less likely. Another advantage of performing the anastomosis in situ may be less tension on the esophageal ends allowing longer gaps to be brought together without tearing. This appeared to be the case in one patient with a trifurcation fistula and long gap, although he did develop an anastomotic stricture that required dilation. The major technical hurdle in this operation is the suturing of the anastomosis. The placement of the sutures and knot tying are technically demanding and relatively imprecise. Also, as opposed to the open technique where the entire posterior row of sutures can be placed and then brought together to disperse the tension along multiple points during knot tying, this method places all the tension
7 on one suture at a time. So far this has not been a significant problem but it could prove to be. For this procedure to become more widely accepted it maybe necessary to develop a mechanical anastomotic device or self-knotting suture. The rate of anastomotic narrowing requiring at least one dilation in this series is 30% but was initially almost 50%, a figure much higher then our open experience. This may have been secondary to inadequate approximation of the mucosal ends or an insufficient opening being made in the upper pouch. We have modified the technique to eliminate both of these problems as well as switched to a monofilament absorbable suture. These changes seem to have resolved the problem with only 2 of the last 16 patients requiring dilation. Clearly the technical and physiologic hurdles to accomplish this type of repair are many and it will require continued advances before this surgery becomes commonplace. However, the ability to perform this complex reconstruction without a thoracotomy lays further groundwork in minimizing surgical morbidity in even the smallest pediatric patients.
References 1. Rothenberg SS, Chang HT, Bealer JF: Experience with minimally invasive surgery in infants. Am J Surg 176:654-658, 1998 2. Fujimoto T, Segawa O, Lane GJ, et al: Laparoscopic surgery for newborn infants. Surg Endosc 13:773-777, 1999 3. Rothenberg SS, Chang JHT, Toews WH, et al: Thoracscopic closure of patent ductus arteriosus: A less traumatic and more cost-effective technique. J Pediatr Surg 30:1057-1060, 1995 4. Rothenberg SS, Chang JHT, Bealer J: First decades experience with minimally invasive surgery in neonates. Pediatr Endosurg Innovative Tech 8:89-94, 2004 5. Lobe TE, Rothenberg SS, Waldschmidt J, Stroeder L: Thoracoscopic repair of esophageal atresia in an infant: A surgical first. Pediatr Endosurg Innovative Tech 3:141-148, 1999 6. Rothenberg SS: Thoracoscopic repair of a tracheoesophageal in a neonate. Pediatr Endosurg Innovative Tech 4:150-156, 2000 7. Rothenberg SS: Thoracoscopic Repair of tracheo-esophageal fistula and esophageal atresia in newborns. J Pediatr Surg 37:869-872, 2002 8. Little DC, Rescorla FJ, Grosfeld JL, et al: Long-term analysis of children with esophageal atresia and tracheoesophageal fistula. J Pediatr Surg 38:852-856, 2003 9. Konkin DE, O’Hali WA, Webber EM, et al: Outcomes in esophageal atresia and tracheoesophageal fistula. J Pediatr Surg 38:1726-1729, 2003 10. Vaiquez JJ, Murcia J, DiezPardo JA: Morbid musculoskeletal sequelae of thoracotomy for tracheoesophageal fistula. J Pediatr Surg 20:511514, 1995 11. Bianchi A, Sowende O, Aliza NK, et al: Aesthetics and lateral thoracotomy in the neonate. J Pediatr Surg 12:1798-1800, 1998