Journal of Pediatric Surgery 50 (2015) 250–254
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Revisional surgery for recurrent tracheoesophageal fistula and anastomotic complications after repair of esophageal atresia in 258 infants Antti I. Koivusalo ⁎, Mikko P. Pakarinen, Harry G. Lindahl, Risto J. Rintala Children's Hospital, Section of Paediatric Surgery, University of Helsinki, Helsinki, Finland
a r t i c l e
i n f o
Article history: Received 21 October 2014 Accepted 2 November 2014 Key words: Esophageal atresia Tracheoesophageal fistula Complications Surgical series
a b s t r a c t Aim: We assessed the occurrence and outcome of major reoperations following repair of esophageal atresia with or without tracheoesophageal fistula (TOF). Major outcome measures were survival, preservation of native esophagus, and long–term esophageal function. Methods: Hospital charts of 258 consecutive patients treated for esophageal atresia from 1980 to 2013 were reviewed. Results: Forty-two (16%) patients required a total of 57 reoperations after primary repair (n = 37) or esophageal reconstruction (n = 5). The indications were anastomotic leakage (n = 17), anastomotic rupture after endoscopic dilatation (n = 5), recurrent tracheoesophageal fistula (TOF) (n = 12), undiagnosed proximal TOF (n = 3), recalcitrant anastomotic stricture (n = 11, primary anastomosis 9, reconstruction 2), undetected proximal fistula (n = 3), and inadvertently perforated jejunal graft (n = 1). Anastomotic leakage and rupture after dilatation were treated with rethoracotomy and suture and recurrent or undetected TOF by open repair. Strictures not manageable with repeated dilatations were resected and esophageal ends reanastomosed (n = 10) or bridged with jejunum graft (n = 1). Five (12%) patients required further reoperations, two after recurrent TEF (reocclusion n = 1, reconstruction with gastric tube n = 1), two after stricture operations (re-resection n = 1, resuture after leakage n = 1), and one after recurrent dilatation-related rupture. Mortality was 4/42 (10%). Two patients died of recurred leakage or TOF and two of unrelated cause. Of 38 survivors, 35 retained their native or initially reconstructed esophagus, and 3 had secondary reconstruction. After a median follow-up of 23 (range 0.6–32) years, 35 (95%) patients have acceptable esophageal function. Three patients remained dependent on gastrostomy feedings. Conclusion: Anastomotic and TOF complications required a substantial number of reoperations, including esophageal reconstructions. Over 90% of the patients survived with a functioning native or reconstructed esophagus. © 2015 Elsevier Inc. All rights reserved.
The repair of an esophageal atresia (OA) is usually followed by an excellent outcome. Still, the repair may fail or become complicated resulting in morbidity and occasionally mortality. In our previous review from Helsinki [1], leakage of the esophageal anastomosis occurred in 7%, recalcitrant anastomotic stricture occurred in 6% and a recurrent tracheoesophageal fistula (TOF) with or without a primarily undiagnosed proximal TOF (type D OA) occurred in 6%–10% of infants. All complications expose a neonate to infection and problems of nutrition and respiratory dysfunction. Reoperations are often complex and require experience and judicious timing. A variety of reconstructive procedures may be required to regain esophageal continuity [2]. The aim of this study was to describe the outcome after revisional esophageal surgery for the complications following repair of OA in a cohort of 42 children extracted from our series managed in our institution between 1980 and 2013. ⁎ Corresponding author at: Children's Hospital, Stenbackinkatu 11, 00290, PO BOX 281, Helsinki, Finland. Tel.: +358 504272512. E-mail address: antti.koivusalo@hus.fi (A.I. Koivusalo). http://dx.doi.org/10.1016/j.jpedsurg.2014.11.004 0022-3468/© 2015 Elsevier Inc. All rights reserved.
1. Patients and methods 1.1. Study design This was a retrospective case note review and followed institutional ethical committee approval. The case-notes were reviewed of 258 consecutive infants with OA admitted into Children's Hospital of Helsinki University Hospital between 1980 and 2013. We included all infants who underwent reoperative surgery for anastomotic leakage, anastomotic stricture, recurrent TOF or TOF which had remained undiagnosed during the primary repair. We excluded operations for gastroesophageal reflux (GOR), operations for tracheomalacia and other associated disorders, endoscopic procedures and stents. Follow-up included radiological imaging, endoscopy findings and growth outcomes. The initial defect was classified according to Gross classification [3] i.e. Type A (OA without TOF), Type B (proximal TOF only), Type C (OA and distal TOF), Type D (OA and both proximal and distal TOF) and Type E (isolated TOF) depending on the surgical anatomy. Long-gap OA was defined by location of the distal TOF and judgement of the operating surgeon.
A.I. Koivusalo et al. / Journal of Pediatric Surgery 50 (2015) 250–254
Compared with our earlier report [1] this study had a larger initial cohort of patients (258 patients from 1980 to 2013 compared with 130 patients from 1991 to 2012) and concentrates on surgical management of complications of esophageal anastomosis. 1.2. Principles of surgical management Our techniques of repair of OA in our institution have been described in detail elsewhere [1]. Rigid bronchoscopy is routinely performed prior to the repair and the surgical approach in for the majority was a right transpleural thoracotomy. Recently an extrapleural approach was also used. The thoracoscopic approach was used in three infants only. Infants who were not amenable for primary end-to-end repair were managed with feeding gastrostomy and upper pouch suction drainage or cervical esophagostomy. The gap between upper and lower esophageal pouches was measured before definitive repair by double flexible endoscopy under fluoroscopic control. Delayed primary closure or esophageal reconstruction was scheduled at a minimum age of two months. Methods of esophageal reconstruction were reversed gastric tube (RGT) [4] or pedicled jejunal interposition graft (JIG) [5]. 1.3. Treatment of postoperative complications 1.3.1. Anastomotic leakage Postoperative anastomotic leakage was suspected if the patient had recurring pneumothorax and/or leak of air, saliva and gastric contents from the chest drain. Confirmation of the leakage was performed by chest x-ray with oral contrast. Primary treatment of anastomotic leakage included chest drainage, antibiotics and nutrition via nasogastric tube with parenteral nutrition support. If the leakage could not be controlled with these measures and the leakage was recent without signs of severe infection rethoracotomy and resuture was performed. If possible the sutured area was covered with a flap of vascularized tissue (pleura, pericardium, intercostal muscle). After one week healing was confirmed with contrast esophagogram and oral feeding started. Anastomotic dilatations were started 2–3 weeks thereafter. Cervical esophagostomy, suture closure of the distal esophagus and gastrostomy were performed in those with severe pleural contamination or wide anastomotic defect. Reconstruction of esophageal continuity was scheduled after 3 to 12 months. Routine repeated postoperative dilatations were performed until the anastomotic width was 10 mm until 2001. Since then, dilatations have been performed selectively. Major dilatation associated anastomotic ruptures were treated by rethoracotomy and resuture as for anastomotic leakage.
Fig. 1. Origin of patients with major anastomotic complications (n = 42).
251
1.3.2. Recurrent TOF Recurrent TOF or undiagnosed second TOF was suspected in infants with feeding difficulties and recurrent chest infections. A number of methods were used for diagnosing recurrent TOF. These included rigid bronchoscopy with intubation of the pit of the initial TOF with a ureteral catheter or by injection of dilute methylene blue; contrast injected into a nasogastric tube slowly withdrawn during injection; or filling the esophagus with water during gastroscopy during manual continuous positive airway pressure (CPAP) ventilation and watching for a jet of bubbles. If recurrent or undiagnosed TOF allowed intubation with a catheter, the catheter introduced through the fistula from the trachea was retrieved from the mouth with gastroscope and the looped catheter was used as a guide for intraoperative identification. The surgical approach was transpleural through right lateral thoracotomy or right cervical incision for high fistulas. The fistulous openings were closed by suture and an attempt to interpose vascularized pericardium, pleura or a strip of intercostal or neck muscle. Anastomotic resection and reanastomosis was considered in those with concomitant stricture. 1.3.3. Recalcitrant anastomotic stricture Primary treatment of recalcitrant anastomotic stricture was serial endoscopic balloon dilatations, recently with topical mitomycin application [6] coupled with medical and surgical control of GOR. Esophageal resection was considered if a dilatator guide wire could not be passed through the stricture, the stricture failed to yield to dilatations or if, after numerous dilatations, the esophageal lumen repeatedly occluded and caused dysphagia or respiratory problems. Esophageal continuity was reestablished with end-to-end anastomosis or, in long strictures or if substantial length of esophagus was lost, RGT or JIG. Data are quoted as median (range). Statistical comparisons were made with StatView® 512 computer programme (Brain Power, Calabasas CA, USA). Risk ratios for different outcome variables were estimated using univariate logistic regression analysis. P values of b0.05 were considered statistically significant. 2. Results A total of 258 infants with OA ± TOF were treated from 1980 to 2013. Complications included early anastomotic leakage (n = 20, 8%), anastomotic rupture after endoscopic dilatation (n = 5, 2%), recurrent TOF (n = 14, 5%), undiagnosed proximal TOF (n = 3, 1%), recalcitrant anastomotic stricture (n = 11, 4%) and others (n = 1). Revisional surgery was performed in a total of 42 patients and included 35 (14%) of our 251 children with primary repair and 7 children referred from other hospitals. (Fig. 1) their median birth weight was 2.6 (0.9–4.8) kg and gestational age 38 (30–42) weeks. Birth weight was b1500 g in 4 infants. Types of OA were as follows: A (n = 7), C (n = 30), D (n = 3), and E (n = 2). According to the Spitz classification [7], 33 (79%) patients belonged to Group I (no major cardiac anomalies and birth weight ≥ 1500 g) and 9 (21%) to Group II (major cardiac anomaly or birth weight b 1500 g) and none to group III (major cardiac anomaly and birth weight b1500 g). Eighteen (43%) had an associated congenital disease including significant heart disease (n = 7), congenital lobar emphysema (n = 1), anorectal malformation (n = 9), duodenal atresia (n = 4), choanal atresia (n = 2), omphalocele (n = 1), urogenital anomaly (n = 5), anomaly of limbs or vertebrae (n = 5), Down's syndrome (n = 3), Feingold's syndrome (n = 1) and CHARGE association (n = 2). Fifteen (36%) of the 42 patients required antireflux surgery for symptomatic GOR within the first year of life and eight (19%) underwent aortopexy for significant tracheomalacia within the first three months of life. Fifty-seven major revisional operations were performed in 42 children [following primary end-to end repair (n = 37) and esophageal reconstruction (n = 5)]. Two or more operations were performed in 11
252
A.I. Koivusalo et al. / Journal of Pediatric Surgery 50 (2015) 250–254
(26%) children (Table 1). Four children had undergone a total of 9 operations in other hospitals before referral and included leakage (n = 4), failed repairs of recurrent TOF (n = 2), stricture (n = 1), cervical esophagostomy and feeding gastrostomy (n = 2). Median follow-up was 23 (range 0.7–32) years. 2.1. Anastomotic leakage A total of 20 patients had early anastomotic leakage after primary anastomosis (n = 18), or reconstruction (n = 2). We performed rethoracotomy and resuture to 13 with primary anastomosis (n = 12) or reconstruction (n = 1) at 3 (2–13) days after the preceding operation. In addition we managed 3 (reconstruction n = 1) conservatively. Leakage in the remaining 4 children was initially treated in other hospitals before referral to us. Leaks healed in 12 of the 13 children managed surgically in our institution but one extremely premature neonate died of sepsis. Leaks healed also in our 3 conservatively managed children. Ten (66%) then had an uneventful recovery although one child with swallowing disorder related with Feingold's syndrome is permanently dependent on a feeding gastrostomy. Five of our children and four treated in other hospitals (i.e. 47% of the total of 19 survivors) developed late complications including anastomotic stricture (n = 4) and recurrent TOF with (n = 3) or without (n = 2) anastomotic stricture. 2.2. Anastomotic rupture after endoscopic dilatation Five (2%) underwent rethoracotomy and resuture after anastomotic rupture during endoscopic balloon dilatation of primary anastomosis (n = 4) or cervical anastomosis of RGT (n = 1). The rupture was diagnosed immediately in all five patients. During subsequent dilatations one patient had a second anastomotic rupture which was managed surgically. Eventually all five children had primary healing and gained full oral intake with no further complications. One patient died at the age of 13 months from an unrelated cause. 2.3. Recurrent TOF Recurrent TOF occurred in 14 children. Two patients (type C) had an associated anastomotic stricture. Two had a major cardiac disease and three were born premature. Twelve had recurrent TOF after primary repair (type C n = 11, including one with long-gap; type D n = 1). Before referral from other hospitals two patients (type C) had undergone attempted repair of recurrent TOF with (n = 1) or without (n = 1) stricture with the result of re-recurrence of TOF in one and in the other wide defect after resection of stricture resulted in formation of cervical esophagostomy and feeding gastrostomy. In one infant who had had a thoracoscopic primary repair in our institution a metal clip used in the closure of the tracheal end of TOF had cut through the tracheal fistula remnant. A 19 year-old patient with recurrent TOF diagnosed at the age of 13 had two failed endoscopic procedures to close recurrent TOF before she was scheduled to open closure.
Table 1 Complications of esophageal anastomosis and fistula and the number of reoperations in 42 children. Type of complication
Operations (n)
Anastomotic leakage after primary repair or replacement (n = 13) Anastomotic rupture after endoscopic dilatation (n = 5) Recurrent TOF with (n = 2) or without (n = 12) anastomotic stricture Anastomotic stricture after primary repair or replacement (n = 11) Undiagnosed proximal TOF (n = 3) Perforation of jejunal graft by nasogastric tube (n = 1) Total
13 6 18 16 3 1 57
Of the remaining two patients with type E one had earlier undergone open closure in another hospital and thereafter an endoscopic closure in our institution and the other had undergone endoscopic closure in our institution. Recurrent fistula was diagnosed by performing rigid bronchoscopy with tracheal methylene blue injection in all patients. Esophageal contrast fluoroscopy was performed in 3 but recurrent TEF could only be demonstrated in one. Reoperations took place at a median of 4.5 (0.1–230) months after the primary repair. Twelve children underwent open repair of recurrent TOF by thoracotomy (types C and D) and two by cervical incision. One child (type C) with recurrent TOF and stricture underwent open closure of TOF, resection of stricture and end-to-end reanastomosis. The other patient with recurrent TOF and stricture, with cervical esophagostomy and gastrostomy underwent RGT reconstruction. Two children had significant complications. A re-repair for a second recurrence of a distal TOF was required in a child with a type D previously missed proximal TOF. One child (type C) who underwent resection of an anastomotic stricture developed a mediastinal abscess and complete dehiscence of both esophageal and tracheal sutures and required resuturing of trachea, cervical esophagostomy, feeding jejunostomy and closure of the distal esophageal stump. She underwent RGT reconstruction after three months and resumed full oral intake at the age of four years. In one extremely premature patient (type C), the TOF recurred at the eight postoperative day and a cervical esophagostomy, closure of the distal esophagus and feeding gastrostomy were done. The patient, however, died of sepsis one week thereafter. Recurrent TOF eventually healed and full oral intake was resumed in 13 children (native esophagus n = 11, RGT reconstruction n = 2). One patient died. 2.4. Undiagnosed Proximal TOF After primary repair for a supposed type C, a proximal TOF was diagnosed in three patients and the diagnosis was reverted to type D. The bronchoscopy preceding the repair missed the proximal fistula in two and in one it was not performed. Diagnosis was made by repeat bronchoscopy (n = 2) or with prone contrast esophagogram (n = 1). The patients underwent closure via thoracotomy (n = 2) or cervicotomy (n = 1) median of 3 (range 1–4) months after the primary repair. Eventually all healed completely. Two patients have full oral intake, one is gastrostomy fed because of swallowing disorder typical of CHARGE association. 2.5. Anastomotic stricture Eleven patients underwent esophageal resection because of anastomotic stricture (primary anastomosis n = 9, reconstruction n = 2). One patient (type C) was referred to us after primary repair, reoperation for anastomotic leakage and eventually cervical esophagostomy with closure of the distal esophageal stump and feeding gastrostomy after an attempt to resect anastomotic stricture. He subsequently underwent reconstruction with RGT in our institution. The remaining ten children included six with primary anastomosis (type C), two with delayed primary anastomosis (type A), one (type A) with primary JIG with stricture of the proximal esophagus–jejunal graft anastomosis and one (type C, long gap) with primary reconstruction with gastric tube with a stricture in cervical esophagus–gastric tube anastomosis. The ten children had already undergone a median of 11 (5–39) serial endoscopic balloon dilatations, two had had topical mitomycin applications, seven (type A n = 2, type C n = 5) had had antireflux surgery, and all had had medical treatment for GOR. Reoperations occurred at a median 10 (2–38) months after the primary operation. After resection seven patients had end-to-end re-anastomosis, one (type C OA) underwent a JIG anastomosis because of the long gap
A.I. Koivusalo et al. / Journal of Pediatric Surgery 50 (2015) 250–254
after resection. Also the two patients with previous JIG and RGT underwent resection and reanastomosis. To maintain anastomotic patency after revision all children required a median of 3 (2–9) anastomotic dilatations. After resection and end-to-end reanastomosis 3 patients underwent rethoracotomy and suture closure of postoperative leakage (n = 2) or rupture after dilatation (n = 1) and six months later one of patients with leakage required redo resection and reanastomosis for recurrent stricture. One (9%) of the patients died of pneumonia at the age of 13 months, and all ten survivors regained full oral intake by the age of 14 months. In addition one patient with JIG and stricture of distal graft–esophagus anastomosis was treated with endoscopic placement of expandable absorbable stent. After initial success the patient died of pneumonia in another hospital. No postmortem was performed and the relation with death to the stent placement remains unknown. 2.6. Other complications A one-year old child with Down's syndrome underwent primary reconstruction for type A OA with right-sided JIG. On the third postoperative she developed a right-sided pneumothorax and underwent rethoracotomy. The tip of the nasogastric tube had perforated the jejunal graft which otherwise appeared completely vital. After suture of the perforation and reposition of the nasogastric complete healing occurred. She remains gastrostomy fed because of severe heart failure but has the ability to swallow. 2.7. Risk factors for complications Leakage in the primary anastomosis increased the risk of recurrent TOF [RR = 8.0 (95% CI 2.1–31), P = 0.001] and stricture [RR = 25 (95% CI 5.8–111), P b 0.001]. The risk of a stricture was increased by long gap [RR = 6.3 (95% CI 1.6–25), P = 0.009] and GER requiring antireflux surgery [RR = 10 (95% CI 2.6–42), P = 0.001]. Birth weight, gestation age and Spitz classification were not related with occurrence of complications. 2.8. Loss of native esophagus, mortality and oral intake Thirty-eight (90%) of 42 children survived the revisional surgery. Two of the deaths were related to recurrent leakage or TOF following surgery. Thirty-five survivors (91%) retained their own esophagus and three (9%) underwent secondary reconstruction. Thirty–four of 37 patients who had revisional surgery after primary anastomosis or closure of TOF survived, and 30 (91%) retained their own esophagus and three (9%) underwent secondary reconstruction. Loss of native esophagus occurred after revisional surgery for recurrent TOF with (n = 2) or without (n = 1) anastomotic stricture. Of 38 survivors 35 (92%) eventually resumed full oral intake of foods whereas three remained dependent on gastrostomy feeding primarily because of swallowing disorders associated with CHARGE association (n = 2) and Feingold's syndrome (n = 1). 3. Discussion This series was collected over a 34-year period and there have been many improvements in neonatal care resulting in an improvement in outcome of OA throughout the world. The treatment of complications of the esophageal anastomosis and TOF remains, however, an important issue. Patients with the most severe complications following OA repair should be centralized to a national tertiary centre with a surgical team which has experience in alternative surgical approaches and techniques. Our series reflected the complex nature of revisional esophageal surgery in neonates. It was not uncommon that revisional surgery required
253
still further corrective operations. Only two of the four deaths in the series were directly related to revisional surgery. The overall mortality (10%) in the present series appears acceptable and rates of loss of native esophagus (12%) and gastrostomy feeding (8%) were relatively low. After revisional surgery the risk of mortality was highest in premature neonates. A large percentage of anastomotic complications consisted of leakages shortly after primary repair or anastomotic ruptures after dilatation. The large number of operations for early leaks reflects the fact that in the majority of cases the surgical approach was transpleural after which containment of an anastomotic leakage is less likely than with an extrapleural approach. Reoperations for leakage were of emergent nature without as much planning as in more complex operations for stricture. In order to save the native esophagus repair of significant anastomotic leakage should be attempted without delay before local infection and necrosis renders the esophageal wall unsuitable for resuturing [8]. Anastomotic leakage was a significant risk factor for recurrent TOF and stricture. In some cases leakage also eventually leads to subsequent loss of the native esophagus. In accordance with our results Bruch et al. [9] found a rate as high as 64% of anastomotic leakage in those who subsequently developed recurrent TOF. Children with feeding and respiratory problems and a history of anastomotic leakage should be investigated, if necessary repeatedly, by contrast radiograms and endoscopy with methylene blue test [10,11]. Revision operations for patients with recurrent TOF without anastomotic stricture had a good success rate. The repair requires careful separation of esophagus and trachea and interposition of vascularized tissue. However, operations for recurrent TOF and stricture require esophageal resection and a new anastomosis and are more likely to develop complications. Both our patients with recurrent TOF and stricture eventually required esophageal reconstruction. Thoracoscopic repair has had high rate of complications [12] including recurrent TOF and for this reason our preference is still open rather than a thoracoscopic approach. Repeated dilatations with bougie or pneumatic balloon are an effective treatment in the majority of anastomotic strictures. Pearson et al. [13] reported a 3% rate of resection in 114 children with reflux stricture. In our previous series (1991–2012), 29% required N five anastomotic dilatations [1]. In the present series we had 11 children (5% of the population of 237 patients) in whom resection of the stricture was required. After resection of a stricture the new anastomosis may require several dilatations before constant functional luminal diameter is reached and even then occasional need for dilatation may persist for years. At least some paediatric esophageal anastomotic strictures may be treated with expandable absorbable stents. Manfredi et al. [14] had promising results in the use of removable expandable stents in strictures in patients with OA. It is, however, not yet known whether safety issues related with recurrent structuring after stents or, possible risk of recurrent TOF because of stent pressure, preclude the use of these in OA. Our experience of with an expandable absorbable stent is limited in one patient (type A OA, primary reconstruction with pedicled jejunal graft, not included in the present series) with a stented stricture in distal jejunal graft–esophageal anastomosis. After esophageal reconstructions reoperations for graft necrosis, anastomotic leakage, fistulas, strictures or excess and redundancy of the graft are not uncommon [15,16]. Revisional operations for pedicled grafts have the risk of damaging the vascular supply during surgical dissection. In our series no graft loss occurred in revisional operations for complications in primary RGT and JIG or, when RGT and JIG were used for secondary reconstruction. The limitations of the study are the retrospective nature and reliance on reports written by at least ten operating surgeons. Follow-up of some patients from the early part of the series has not been done in our institution and although their present situation is known, they might have suffered from events unknown to us during the decades after revisional surgery.
254
A.I. Koivusalo et al. / Journal of Pediatric Surgery 50 (2015) 250–254
Anastomotic complications or recurrent TOF following primary repair or primary reconstruction for esophageal atresia with or without TOF required a substantial number of reoperations including esophageal reconstructions. Nonetheless, over 90% of the patients survived with a functioning native or reconstructed esophagus.
References [1] Koivusalo AI, Pakarinen MP, Rintala RJ. Modern outcomes of oesophageal atresia: single centre experience over the last twenty years. J Pediatr Surg 2013;48:297–303. [2] Lillehei GW, Shamberger RC. Reoperative esophageal surgery. Semin Pediatr Surg 2003;12:100–6. [3] Gross RE. The surgery of infancy and childhood. Philadelphia: WB Saunders; 1953. [4] Heimlich HJ. Elective replacement of the oesophagus. Br J Surg 1966;53:913–6. [5] Bax NM, van der Zee DC. Jejunal pedicle grafts for reconstruction of the esophagus in children. J Pediatr Surg 2007;42:363–9. [6] Rosseneu S, Afzal N, Yerushalmi B, et al. Topical application of mitomycin-C in oesophageal strictures. J Pediatr Gastroenterol Nutr 2007;44:336–41. [7] Spitz L, Kiely EM, Morecroft JA. Oesophageal atresia: at-risk groups for the 1990s. J Pediatr Surg 1994;29:723–5.
[8] Chavin K, Field G, Chandler J, et al. Save the child's esophagus: management of major disruption after repair of esophageal atresia. J Pediatr Surg 1996;31: 48–51. [9] Bruch SW, Hirschl RB, Coran AG. The diagnosis and management of recurrent tracheoesophageal fistulas. J Pediatr Surg 2010;45:337–40. [10] Coran AG. Redo esophageal surgery: the diagnosis and management of recurrent tracheooesophageal fistula. Pediatr Surg Int 2013;29:995–9. [11] Guo W, Li Y, Jiao A, et al. Tracheoesophageal fistula after primary repair of type C esophageal atresia in the neonatal period: recurrent or missed second congenital fistula. J Pediatr Surg 2010;45:2351–5. [12] Szavay PO, Zundel S, Blumenstock G, et al. Perioperative outcome of patients with esophageal atresia and tracheo-esophageal fistula undergoing open versus thoracoscopic surgery. J Laparoendosc Adv Surg Tech A 2011;21:439–43. [13] Pearson EG, Downey EC, Barnhart DC, et al. Reflux esophageal stricture—a review of 30 years' experience in children. J Pediatr Surg 2010;45:2356–60. [14] Manfredi MA, Jennings RW, Anjum MW, et al. Externally removable stents in the treatment of benign recalcitrant strictures and esophageal perforations in pediatric patients with esophageal atresia. Gastrointest Endosc 2014;80: 246–52. [15] Dunn JC, Fonkalsrud EW, Applebaum H, et al. Reoperation after esophageal replacement in childhood. J Pediatr Surg 1999;34:1630–2. [16] Arul GS, Parikh D. Oesophageal replacement in children. Ann R Coll Surg Engl 2008; 90:7–12.
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Revisional surgery for recurrent tracheoesophageal fistula and anastomotic complications after repair of esophageal atresia in 258 infants Antti I. Koivusalo ⁎, Mikko P. Pakarinen, Harry G. Lindahl, Risto J. Rintala Children's Hospital, Section of Paediatric Surgery, University of Helsinki, Helsinki, Finland
a r t i c l e
i n f o
Article history: Received 21 October 2014 Accepted 2 November 2014 Key words: Esophageal atresia Tracheoesophageal fistula Complications Surgical series
a b s t r a c t Aim: We assessed the occurrence and outcome of major reoperations following repair of esophageal atresia with or without tracheoesophageal fistula (TOF). Major outcome measures were survival, preservation of native esophagus, and long–term esophageal function. Methods: Hospital charts of 258 consecutive patients treated for esophageal atresia from 1980 to 2013 were reviewed. Results: Forty-two (16%) patients required a total of 57 reoperations after primary repair (n = 37) or esophageal reconstruction (n = 5). The indications were anastomotic leakage (n = 17), anastomotic rupture after endoscopic dilatation (n = 5), recurrent tracheoesophageal fistula (TOF) (n = 12), undiagnosed proximal TOF (n = 3), recalcitrant anastomotic stricture (n = 11, primary anastomosis 9, reconstruction 2), undetected proximal fistula (n = 3), and inadvertently perforated jejunal graft (n = 1). Anastomotic leakage and rupture after dilatation were treated with rethoracotomy and suture and recurrent or undetected TOF by open repair. Strictures not manageable with repeated dilatations were resected and esophageal ends reanastomosed (n = 10) or bridged with jejunum graft (n = 1). Five (12%) patients required further reoperations, two after recurrent TEF (reocclusion n = 1, reconstruction with gastric tube n = 1), two after stricture operations (re-resection n = 1, resuture after leakage n = 1), and one after recurrent dilatation-related rupture. Mortality was 4/42 (10%). Two patients died of recurred leakage or TOF and two of unrelated cause. Of 38 survivors, 35 retained their native or initially reconstructed esophagus, and 3 had secondary reconstruction. After a median follow-up of 23 (range 0.6–32) years, 35 (95%) patients have acceptable esophageal function. Three patients remained dependent on gastrostomy feedings. Conclusion: Anastomotic and TOF complications required a substantial number of reoperations, including esophageal reconstructions. Over 90% of the patients survived with a functioning native or reconstructed esophagus. © 2015 Elsevier Inc. All rights reserved.
The repair of an esophageal atresia (OA) is usually followed by an excellent outcome. Still, the repair may fail or become complicated resulting in morbidity and occasionally mortality. In our previous review from Helsinki [1], leakage of the esophageal anastomosis occurred in 7%, recalcitrant anastomotic stricture occurred in 6% and a recurrent tracheoesophageal fistula (TOF) with or without a primarily undiagnosed proximal TOF (type D OA) occurred in 6%–10% of infants. All complications expose a neonate to infection and problems of nutrition and respiratory dysfunction. Reoperations are often complex and require experience and judicious timing. A variety of reconstructive procedures may be required to regain esophageal continuity [2]. The aim of this study was to describe the outcome after revisional esophageal surgery for the complications following repair of OA in a cohort of 42 children extracted from our series managed in our institution between 1980 and 2013. ⁎ Corresponding author at: Children's Hospital, Stenbackinkatu 11, 00290, PO BOX 281, Helsinki, Finland. Tel.: +358 504272512. E-mail address: antti.koivusalo@hus.fi (A.I. Koivusalo). http://dx.doi.org/10.1016/j.jpedsurg.2014.11.004 0022-3468/© 2015 Elsevier Inc. All rights reserved.
1. Patients and methods 1.1. Study design This was a retrospective case note review and followed institutional ethical committee approval. The case-notes were reviewed of 258 consecutive infants with OA admitted into Children's Hospital of Helsinki University Hospital between 1980 and 2013. We included all infants who underwent reoperative surgery for anastomotic leakage, anastomotic stricture, recurrent TOF or TOF which had remained undiagnosed during the primary repair. We excluded operations for gastroesophageal reflux (GOR), operations for tracheomalacia and other associated disorders, endoscopic procedures and stents. Follow-up included radiological imaging, endoscopy findings and growth outcomes. The initial defect was classified according to Gross classification [3] i.e. Type A (OA without TOF), Type B (proximal TOF only), Type C (OA and distal TOF), Type D (OA and both proximal and distal TOF) and Type E (isolated TOF) depending on the surgical anatomy. Long-gap OA was defined by location of the distal TOF and judgement of the operating surgeon.
A.I. Koivusalo et al. / Journal of Pediatric Surgery 50 (2015) 250–254
Compared with our earlier report [1] this study had a larger initial cohort of patients (258 patients from 1980 to 2013 compared with 130 patients from 1991 to 2012) and concentrates on surgical management of complications of esophageal anastomosis. 1.2. Principles of surgical management Our techniques of repair of OA in our institution have been described in detail elsewhere [1]. Rigid bronchoscopy is routinely performed prior to the repair and the surgical approach in for the majority was a right transpleural thoracotomy. Recently an extrapleural approach was also used. The thoracoscopic approach was used in three infants only. Infants who were not amenable for primary end-to-end repair were managed with feeding gastrostomy and upper pouch suction drainage or cervical esophagostomy. The gap between upper and lower esophageal pouches was measured before definitive repair by double flexible endoscopy under fluoroscopic control. Delayed primary closure or esophageal reconstruction was scheduled at a minimum age of two months. Methods of esophageal reconstruction were reversed gastric tube (RGT) [4] or pedicled jejunal interposition graft (JIG) [5]. 1.3. Treatment of postoperative complications 1.3.1. Anastomotic leakage Postoperative anastomotic leakage was suspected if the patient had recurring pneumothorax and/or leak of air, saliva and gastric contents from the chest drain. Confirmation of the leakage was performed by chest x-ray with oral contrast. Primary treatment of anastomotic leakage included chest drainage, antibiotics and nutrition via nasogastric tube with parenteral nutrition support. If the leakage could not be controlled with these measures and the leakage was recent without signs of severe infection rethoracotomy and resuture was performed. If possible the sutured area was covered with a flap of vascularized tissue (pleura, pericardium, intercostal muscle). After one week healing was confirmed with contrast esophagogram and oral feeding started. Anastomotic dilatations were started 2–3 weeks thereafter. Cervical esophagostomy, suture closure of the distal esophagus and gastrostomy were performed in those with severe pleural contamination or wide anastomotic defect. Reconstruction of esophageal continuity was scheduled after 3 to 12 months. Routine repeated postoperative dilatations were performed until the anastomotic width was 10 mm until 2001. Since then, dilatations have been performed selectively. Major dilatation associated anastomotic ruptures were treated by rethoracotomy and resuture as for anastomotic leakage.
Fig. 1. Origin of patients with major anastomotic complications (n = 42).
251
1.3.2. Recurrent TOF Recurrent TOF or undiagnosed second TOF was suspected in infants with feeding difficulties and recurrent chest infections. A number of methods were used for diagnosing recurrent TOF. These included rigid bronchoscopy with intubation of the pit of the initial TOF with a ureteral catheter or by injection of dilute methylene blue; contrast injected into a nasogastric tube slowly withdrawn during injection; or filling the esophagus with water during gastroscopy during manual continuous positive airway pressure (CPAP) ventilation and watching for a jet of bubbles. If recurrent or undiagnosed TOF allowed intubation with a catheter, the catheter introduced through the fistula from the trachea was retrieved from the mouth with gastroscope and the looped catheter was used as a guide for intraoperative identification. The surgical approach was transpleural through right lateral thoracotomy or right cervical incision for high fistulas. The fistulous openings were closed by suture and an attempt to interpose vascularized pericardium, pleura or a strip of intercostal or neck muscle. Anastomotic resection and reanastomosis was considered in those with concomitant stricture. 1.3.3. Recalcitrant anastomotic stricture Primary treatment of recalcitrant anastomotic stricture was serial endoscopic balloon dilatations, recently with topical mitomycin application [6] coupled with medical and surgical control of GOR. Esophageal resection was considered if a dilatator guide wire could not be passed through the stricture, the stricture failed to yield to dilatations or if, after numerous dilatations, the esophageal lumen repeatedly occluded and caused dysphagia or respiratory problems. Esophageal continuity was reestablished with end-to-end anastomosis or, in long strictures or if substantial length of esophagus was lost, RGT or JIG. Data are quoted as median (range). Statistical comparisons were made with StatView® 512 computer programme (Brain Power, Calabasas CA, USA). Risk ratios for different outcome variables were estimated using univariate logistic regression analysis. P values of b0.05 were considered statistically significant. 2. Results A total of 258 infants with OA ± TOF were treated from 1980 to 2013. Complications included early anastomotic leakage (n = 20, 8%), anastomotic rupture after endoscopic dilatation (n = 5, 2%), recurrent TOF (n = 14, 5%), undiagnosed proximal TOF (n = 3, 1%), recalcitrant anastomotic stricture (n = 11, 4%) and others (n = 1). Revisional surgery was performed in a total of 42 patients and included 35 (14%) of our 251 children with primary repair and 7 children referred from other hospitals. (Fig. 1) their median birth weight was 2.6 (0.9–4.8) kg and gestational age 38 (30–42) weeks. Birth weight was b1500 g in 4 infants. Types of OA were as follows: A (n = 7), C (n = 30), D (n = 3), and E (n = 2). According to the Spitz classification [7], 33 (79%) patients belonged to Group I (no major cardiac anomalies and birth weight ≥ 1500 g) and 9 (21%) to Group II (major cardiac anomaly or birth weight b 1500 g) and none to group III (major cardiac anomaly and birth weight b1500 g). Eighteen (43%) had an associated congenital disease including significant heart disease (n = 7), congenital lobar emphysema (n = 1), anorectal malformation (n = 9), duodenal atresia (n = 4), choanal atresia (n = 2), omphalocele (n = 1), urogenital anomaly (n = 5), anomaly of limbs or vertebrae (n = 5), Down's syndrome (n = 3), Feingold's syndrome (n = 1) and CHARGE association (n = 2). Fifteen (36%) of the 42 patients required antireflux surgery for symptomatic GOR within the first year of life and eight (19%) underwent aortopexy for significant tracheomalacia within the first three months of life. Fifty-seven major revisional operations were performed in 42 children [following primary end-to end repair (n = 37) and esophageal reconstruction (n = 5)]. Two or more operations were performed in 11
252
A.I. Koivusalo et al. / Journal of Pediatric Surgery 50 (2015) 250–254
(26%) children (Table 1). Four children had undergone a total of 9 operations in other hospitals before referral and included leakage (n = 4), failed repairs of recurrent TOF (n = 2), stricture (n = 1), cervical esophagostomy and feeding gastrostomy (n = 2). Median follow-up was 23 (range 0.7–32) years. 2.1. Anastomotic leakage A total of 20 patients had early anastomotic leakage after primary anastomosis (n = 18), or reconstruction (n = 2). We performed rethoracotomy and resuture to 13 with primary anastomosis (n = 12) or reconstruction (n = 1) at 3 (2–13) days after the preceding operation. In addition we managed 3 (reconstruction n = 1) conservatively. Leakage in the remaining 4 children was initially treated in other hospitals before referral to us. Leaks healed in 12 of the 13 children managed surgically in our institution but one extremely premature neonate died of sepsis. Leaks healed also in our 3 conservatively managed children. Ten (66%) then had an uneventful recovery although one child with swallowing disorder related with Feingold's syndrome is permanently dependent on a feeding gastrostomy. Five of our children and four treated in other hospitals (i.e. 47% of the total of 19 survivors) developed late complications including anastomotic stricture (n = 4) and recurrent TOF with (n = 3) or without (n = 2) anastomotic stricture. 2.2. Anastomotic rupture after endoscopic dilatation Five (2%) underwent rethoracotomy and resuture after anastomotic rupture during endoscopic balloon dilatation of primary anastomosis (n = 4) or cervical anastomosis of RGT (n = 1). The rupture was diagnosed immediately in all five patients. During subsequent dilatations one patient had a second anastomotic rupture which was managed surgically. Eventually all five children had primary healing and gained full oral intake with no further complications. One patient died at the age of 13 months from an unrelated cause. 2.3. Recurrent TOF Recurrent TOF occurred in 14 children. Two patients (type C) had an associated anastomotic stricture. Two had a major cardiac disease and three were born premature. Twelve had recurrent TOF after primary repair (type C n = 11, including one with long-gap; type D n = 1). Before referral from other hospitals two patients (type C) had undergone attempted repair of recurrent TOF with (n = 1) or without (n = 1) stricture with the result of re-recurrence of TOF in one and in the other wide defect after resection of stricture resulted in formation of cervical esophagostomy and feeding gastrostomy. In one infant who had had a thoracoscopic primary repair in our institution a metal clip used in the closure of the tracheal end of TOF had cut through the tracheal fistula remnant. A 19 year-old patient with recurrent TOF diagnosed at the age of 13 had two failed endoscopic procedures to close recurrent TOF before she was scheduled to open closure.
Table 1 Complications of esophageal anastomosis and fistula and the number of reoperations in 42 children. Type of complication
Operations (n)
Anastomotic leakage after primary repair or replacement (n = 13) Anastomotic rupture after endoscopic dilatation (n = 5) Recurrent TOF with (n = 2) or without (n = 12) anastomotic stricture Anastomotic stricture after primary repair or replacement (n = 11) Undiagnosed proximal TOF (n = 3) Perforation of jejunal graft by nasogastric tube (n = 1) Total
13 6 18 16 3 1 57
Of the remaining two patients with type E one had earlier undergone open closure in another hospital and thereafter an endoscopic closure in our institution and the other had undergone endoscopic closure in our institution. Recurrent fistula was diagnosed by performing rigid bronchoscopy with tracheal methylene blue injection in all patients. Esophageal contrast fluoroscopy was performed in 3 but recurrent TEF could only be demonstrated in one. Reoperations took place at a median of 4.5 (0.1–230) months after the primary repair. Twelve children underwent open repair of recurrent TOF by thoracotomy (types C and D) and two by cervical incision. One child (type C) with recurrent TOF and stricture underwent open closure of TOF, resection of stricture and end-to-end reanastomosis. The other patient with recurrent TOF and stricture, with cervical esophagostomy and gastrostomy underwent RGT reconstruction. Two children had significant complications. A re-repair for a second recurrence of a distal TOF was required in a child with a type D previously missed proximal TOF. One child (type C) who underwent resection of an anastomotic stricture developed a mediastinal abscess and complete dehiscence of both esophageal and tracheal sutures and required resuturing of trachea, cervical esophagostomy, feeding jejunostomy and closure of the distal esophageal stump. She underwent RGT reconstruction after three months and resumed full oral intake at the age of four years. In one extremely premature patient (type C), the TOF recurred at the eight postoperative day and a cervical esophagostomy, closure of the distal esophagus and feeding gastrostomy were done. The patient, however, died of sepsis one week thereafter. Recurrent TOF eventually healed and full oral intake was resumed in 13 children (native esophagus n = 11, RGT reconstruction n = 2). One patient died. 2.4. Undiagnosed Proximal TOF After primary repair for a supposed type C, a proximal TOF was diagnosed in three patients and the diagnosis was reverted to type D. The bronchoscopy preceding the repair missed the proximal fistula in two and in one it was not performed. Diagnosis was made by repeat bronchoscopy (n = 2) or with prone contrast esophagogram (n = 1). The patients underwent closure via thoracotomy (n = 2) or cervicotomy (n = 1) median of 3 (range 1–4) months after the primary repair. Eventually all healed completely. Two patients have full oral intake, one is gastrostomy fed because of swallowing disorder typical of CHARGE association. 2.5. Anastomotic stricture Eleven patients underwent esophageal resection because of anastomotic stricture (primary anastomosis n = 9, reconstruction n = 2). One patient (type C) was referred to us after primary repair, reoperation for anastomotic leakage and eventually cervical esophagostomy with closure of the distal esophageal stump and feeding gastrostomy after an attempt to resect anastomotic stricture. He subsequently underwent reconstruction with RGT in our institution. The remaining ten children included six with primary anastomosis (type C), two with delayed primary anastomosis (type A), one (type A) with primary JIG with stricture of the proximal esophagus–jejunal graft anastomosis and one (type C, long gap) with primary reconstruction with gastric tube with a stricture in cervical esophagus–gastric tube anastomosis. The ten children had already undergone a median of 11 (5–39) serial endoscopic balloon dilatations, two had had topical mitomycin applications, seven (type A n = 2, type C n = 5) had had antireflux surgery, and all had had medical treatment for GOR. Reoperations occurred at a median 10 (2–38) months after the primary operation. After resection seven patients had end-to-end re-anastomosis, one (type C OA) underwent a JIG anastomosis because of the long gap
A.I. Koivusalo et al. / Journal of Pediatric Surgery 50 (2015) 250–254
after resection. Also the two patients with previous JIG and RGT underwent resection and reanastomosis. To maintain anastomotic patency after revision all children required a median of 3 (2–9) anastomotic dilatations. After resection and end-to-end reanastomosis 3 patients underwent rethoracotomy and suture closure of postoperative leakage (n = 2) or rupture after dilatation (n = 1) and six months later one of patients with leakage required redo resection and reanastomosis for recurrent stricture. One (9%) of the patients died of pneumonia at the age of 13 months, and all ten survivors regained full oral intake by the age of 14 months. In addition one patient with JIG and stricture of distal graft–esophagus anastomosis was treated with endoscopic placement of expandable absorbable stent. After initial success the patient died of pneumonia in another hospital. No postmortem was performed and the relation with death to the stent placement remains unknown. 2.6. Other complications A one-year old child with Down's syndrome underwent primary reconstruction for type A OA with right-sided JIG. On the third postoperative she developed a right-sided pneumothorax and underwent rethoracotomy. The tip of the nasogastric tube had perforated the jejunal graft which otherwise appeared completely vital. After suture of the perforation and reposition of the nasogastric complete healing occurred. She remains gastrostomy fed because of severe heart failure but has the ability to swallow. 2.7. Risk factors for complications Leakage in the primary anastomosis increased the risk of recurrent TOF [RR = 8.0 (95% CI 2.1–31), P = 0.001] and stricture [RR = 25 (95% CI 5.8–111), P b 0.001]. The risk of a stricture was increased by long gap [RR = 6.3 (95% CI 1.6–25), P = 0.009] and GER requiring antireflux surgery [RR = 10 (95% CI 2.6–42), P = 0.001]. Birth weight, gestation age and Spitz classification were not related with occurrence of complications. 2.8. Loss of native esophagus, mortality and oral intake Thirty-eight (90%) of 42 children survived the revisional surgery. Two of the deaths were related to recurrent leakage or TOF following surgery. Thirty-five survivors (91%) retained their own esophagus and three (9%) underwent secondary reconstruction. Thirty–four of 37 patients who had revisional surgery after primary anastomosis or closure of TOF survived, and 30 (91%) retained their own esophagus and three (9%) underwent secondary reconstruction. Loss of native esophagus occurred after revisional surgery for recurrent TOF with (n = 2) or without (n = 1) anastomotic stricture. Of 38 survivors 35 (92%) eventually resumed full oral intake of foods whereas three remained dependent on gastrostomy feeding primarily because of swallowing disorders associated with CHARGE association (n = 2) and Feingold's syndrome (n = 1). 3. Discussion This series was collected over a 34-year period and there have been many improvements in neonatal care resulting in an improvement in outcome of OA throughout the world. The treatment of complications of the esophageal anastomosis and TOF remains, however, an important issue. Patients with the most severe complications following OA repair should be centralized to a national tertiary centre with a surgical team which has experience in alternative surgical approaches and techniques. Our series reflected the complex nature of revisional esophageal surgery in neonates. It was not uncommon that revisional surgery required
253
still further corrective operations. Only two of the four deaths in the series were directly related to revisional surgery. The overall mortality (10%) in the present series appears acceptable and rates of loss of native esophagus (12%) and gastrostomy feeding (8%) were relatively low. After revisional surgery the risk of mortality was highest in premature neonates. A large percentage of anastomotic complications consisted of leakages shortly after primary repair or anastomotic ruptures after dilatation. The large number of operations for early leaks reflects the fact that in the majority of cases the surgical approach was transpleural after which containment of an anastomotic leakage is less likely than with an extrapleural approach. Reoperations for leakage were of emergent nature without as much planning as in more complex operations for stricture. In order to save the native esophagus repair of significant anastomotic leakage should be attempted without delay before local infection and necrosis renders the esophageal wall unsuitable for resuturing [8]. Anastomotic leakage was a significant risk factor for recurrent TOF and stricture. In some cases leakage also eventually leads to subsequent loss of the native esophagus. In accordance with our results Bruch et al. [9] found a rate as high as 64% of anastomotic leakage in those who subsequently developed recurrent TOF. Children with feeding and respiratory problems and a history of anastomotic leakage should be investigated, if necessary repeatedly, by contrast radiograms and endoscopy with methylene blue test [10,11]. Revision operations for patients with recurrent TOF without anastomotic stricture had a good success rate. The repair requires careful separation of esophagus and trachea and interposition of vascularized tissue. However, operations for recurrent TOF and stricture require esophageal resection and a new anastomosis and are more likely to develop complications. Both our patients with recurrent TOF and stricture eventually required esophageal reconstruction. Thoracoscopic repair has had high rate of complications [12] including recurrent TOF and for this reason our preference is still open rather than a thoracoscopic approach. Repeated dilatations with bougie or pneumatic balloon are an effective treatment in the majority of anastomotic strictures. Pearson et al. [13] reported a 3% rate of resection in 114 children with reflux stricture. In our previous series (1991–2012), 29% required N five anastomotic dilatations [1]. In the present series we had 11 children (5% of the population of 237 patients) in whom resection of the stricture was required. After resection of a stricture the new anastomosis may require several dilatations before constant functional luminal diameter is reached and even then occasional need for dilatation may persist for years. At least some paediatric esophageal anastomotic strictures may be treated with expandable absorbable stents. Manfredi et al. [14] had promising results in the use of removable expandable stents in strictures in patients with OA. It is, however, not yet known whether safety issues related with recurrent structuring after stents or, possible risk of recurrent TOF because of stent pressure, preclude the use of these in OA. Our experience of with an expandable absorbable stent is limited in one patient (type A OA, primary reconstruction with pedicled jejunal graft, not included in the present series) with a stented stricture in distal jejunal graft–esophageal anastomosis. After esophageal reconstructions reoperations for graft necrosis, anastomotic leakage, fistulas, strictures or excess and redundancy of the graft are not uncommon [15,16]. Revisional operations for pedicled grafts have the risk of damaging the vascular supply during surgical dissection. In our series no graft loss occurred in revisional operations for complications in primary RGT and JIG or, when RGT and JIG were used for secondary reconstruction. The limitations of the study are the retrospective nature and reliance on reports written by at least ten operating surgeons. Follow-up of some patients from the early part of the series has not been done in our institution and although their present situation is known, they might have suffered from events unknown to us during the decades after revisional surgery.
254
A.I. Koivusalo et al. / Journal of Pediatric Surgery 50 (2015) 250–254
Anastomotic complications or recurrent TOF following primary repair or primary reconstruction for esophageal atresia with or without TOF required a substantial number of reoperations including esophageal reconstructions. Nonetheless, over 90% of the patients survived with a functioning native or reconstructed esophagus.
References [1] Koivusalo AI, Pakarinen MP, Rintala RJ. Modern outcomes of oesophageal atresia: single centre experience over the last twenty years. J Pediatr Surg 2013;48:297–303. [2] Lillehei GW, Shamberger RC. Reoperative esophageal surgery. Semin Pediatr Surg 2003;12:100–6. [3] Gross RE. The surgery of infancy and childhood. Philadelphia: WB Saunders; 1953. [4] Heimlich HJ. Elective replacement of the oesophagus. Br J Surg 1966;53:913–6. [5] Bax NM, van der Zee DC. Jejunal pedicle grafts for reconstruction of the esophagus in children. J Pediatr Surg 2007;42:363–9. [6] Rosseneu S, Afzal N, Yerushalmi B, et al. Topical application of mitomycin-C in oesophageal strictures. J Pediatr Gastroenterol Nutr 2007;44:336–41. [7] Spitz L, Kiely EM, Morecroft JA. Oesophageal atresia: at-risk groups for the 1990s. J Pediatr Surg 1994;29:723–5.
[8] Chavin K, Field G, Chandler J, et al. Save the child's esophagus: management of major disruption after repair of esophageal atresia. J Pediatr Surg 1996;31: 48–51. [9] Bruch SW, Hirschl RB, Coran AG. The diagnosis and management of recurrent tracheoesophageal fistulas. J Pediatr Surg 2010;45:337–40. [10] Coran AG. Redo esophageal surgery: the diagnosis and management of recurrent tracheooesophageal fistula. Pediatr Surg Int 2013;29:995–9. [11] Guo W, Li Y, Jiao A, et al. Tracheoesophageal fistula after primary repair of type C esophageal atresia in the neonatal period: recurrent or missed second congenital fistula. J Pediatr Surg 2010;45:2351–5. [12] Szavay PO, Zundel S, Blumenstock G, et al. Perioperative outcome of patients with esophageal atresia and tracheo-esophageal fistula undergoing open versus thoracoscopic surgery. J Laparoendosc Adv Surg Tech A 2011;21:439–43. [13] Pearson EG, Downey EC, Barnhart DC, et al. Reflux esophageal stricture—a review of 30 years' experience in children. J Pediatr Surg 2010;45:2356–60. [14] Manfredi MA, Jennings RW, Anjum MW, et al. Externally removable stents in the treatment of benign recalcitrant strictures and esophageal perforations in pediatric patients with esophageal atresia. Gastrointest Endosc 2014;80: 246–52. [15] Dunn JC, Fonkalsrud EW, Applebaum H, et al. Reoperation after esophageal replacement in childhood. J Pediatr Surg 1999;34:1630–2. [16] Arul GS, Parikh D. Oesophageal replacement in children. Ann R Coll Surg Engl 2008; 90:7–12.
