Retrosternal ileocolic esophageal replacement in children revisited

Retrosternal ileocolic esophageal replacement in children revisited

Retrosternal ileocolic esophageal replacement in children revisited Antireflux role of the ileocecal valve The risk of postoperative reflux and pulmon...

924KB Sizes 0 Downloads 66 Views

Retrosternal ileocolic esophageal replacement in children revisited Antireflux role of the ileocecal valve The risk of postoperative reflux and pulmonary aspiration with straight colon or gastric tube esophageal replacement in children prompted us to reevaluate the presumed antireflux role of the ileocecal valve with retrosternal ileocolic interposition. This operation was done in eight patients with esophageal atresia (six) and lye stricture (two) from 19 to 50 months of age between 1983 and 1992. There were no operative deaths. The duration of follow-up ranged from 4 to 115 months. Barium swallow obtained in all patients showed unobstructed esophagoileocolic transit without reflux. Two patients with esophageal atresia had localized proximal anastomotic leaks, which healed spontaneously without stricture. In the two patients with lye ingestion ileoesophageal strictures developed that necessitated revision. None of the patients had postoperative respiratory complications or symptomatic gastroesophageal reflux. All eight children have had their gastrostomy tubes removed, are eating a regular diet, and are growing well. In conclusion, the retrosternal ileocolic conduit provides an excellent substitute esophagus in selected pediatric patients, with potential advantages over delayed primary anastomosis or the straight colon or gastric tube interposition because of the antireflux role of the ileocecal valve. (J THORAC CARDIOVASC SURG 1994;107:1067-72)

Robert J. Touloukian, MD, and George Tellides, MD, PhD, New Haven, Conn.

h e optimal operative technique for esophageal replacement in selected infants and children with esophageal atresia or extensive lye stricture remains controversial and complications are frequent. The first use of the colon for total esophageal replacement in a child was described by Lundblad I in 1921. Waterston/ performed his first successful intrathoracic colon interposition in 1954. Javid'' reported the technique of presternal ileocolic interposition for esophageal atresia in 1954 (Fig. 1), but the posterior mediastinal and substernal routes that use From the Section of Pediatric Surgery, Children's Hospital at YaleNew Haven, and the Yale University School of Medicine, New Haven, Conn. Presented at the Seventy-third Annual Meeting of the New England Surgical Society, Balsams Grand Resort, Dixville Notch, N. H., September 25-27, 1992. Received for publication May II, 1993. Accepted for publication July 7, 1993. Address for reprints: Robert J. Touloukian, MD, Yale University, Pediatric Surgery, 333 Cedar St., P.O. Box 3333, New Haven, CT 06510-8062. Copyright

v

1994 by Mosby-Year Book, Inc.

0022-5223/94 $3.00 + 0

12/1/50476

the right or left colon remain the most popular approaches today. We report our experience with the retrosternal ileocolicinterposition operation, which has been preferred to either straight colon or gastric tube interposition for esophageal atresia or lye stricture at our institution during the past 10 years. Materials and methods Clinical series. The retrosternal ileocolic interposition operation was done in six infants with esophageal atresia and two children with lye stricture between 1983 and 1992 (Table I). Two of the infants with esophageal atresia had a proximal tracheoesophageal fistula (TEF) without distal fistula, two had a distal TEF, and two had isolated atresia with no fistulous connection. Wide separation of the upper and lower esophageal pouches in excess of six vertebral bodies precluded primary anastomosis in all six patients despite attempts at elongation with bougies and myotomy. Feeding gastrostomy was initially done, followed by cervical esophagostomy in the infants with a proximal TEF, whereas the two infants with a distal TEF first underwent extrapleural division of the TEF, but attempts at elongation failed, necessitating cervical esophagostomy between 8 to 10 weeks of age. During the period of this study, six additional patients with wide-gap esophageal atresia and a distal TEF eventually had successful staged repair with the native esophagus. Patients who required esophageal replacement were

1067

1 06 8

The Journal of Thoracic and Cardiovascular Surgery April 1994

Touloukian and Tellides

~

"

."r--

~

"--"?

L,' ..., )\

,-.

operative antibiotics, and postoperative decompression with a gastrostomy tube . Technique of ileocolic substernal interposition. The patient is placed in the supine position with the head turned to the right and the neck extended. The neck, chest, and abdomen are prepared for operation and draped in a continuous fashion, and the peritoneal cavity is entered through a midline incision extending from the xiphoid to below the umbilicus. The gastrostomy is taken down, the right colon mobilized, and the appendix removed. The marginal artery and vein supplying the distal 10 em of the terminal ileum and the entire right colon distal to the middle colic vessel is isolated by preserving the tertiary arcades of the mesoileum and the terminal branches of the ileocolic vessels (Fig. 2, A and B). Intestinal continuity is restored with a single-layer ileocolic anastomosis. Thereafter, the ileocolic pedicle is brought behind the lesser sac and stomach and through the gastrohepatic ligament. A gastrocolic anastomosis is made close to the antrum near the greater curvature with a single layer of interrupted silk sutures. The diaphragmatic attachments to the xiphoid are divided and a retrosternal tunnel created by blunt finger dissection. The cervical esophagostomy is taken down and the esophagus mobilized to gain adequate length. The retrosternal space is entered from below the sternocleidomastoid muscle and the superior portion of the tunnel enlarged enough to prevent compression of the ileocolic pedicle in the mediastinum or at the thoracic inlet. The smaller caliber of the ileum compared with that of the colon, with generous blunt dissection behind the sternocleidomastoid muscle at the thoracic inlet, avoids the need to resect or use a rongeur on the upper sternum or proximal portion of the clavicle. The esophagoileal anastomosis is constructed with a single layer of interrupted Vicryl sutures. The neck and anterior mediastinum are drained with a Jackson-Pratt drain (Zimmer Walker Associates, Hartford, Conn.) through a separate stab incision. The sternocleidomastoid and platysma muscles are reapproximated over the anastomosis and the skin closed. The gastrostomy tube is replaced proximal to the cologastrostomy and brought out through the original stab wound site in the abdominal wall.

Results

Fig. 1. Technique of staged presternal ileocolic interposition as described by Javid ' in 1954. Note preservation of both midcolic and right colic artery. Q., Artery; sup., superior; Rt, right; Esoph., esophagus.

operated on between 19 and 45 months of age (mean of 26 months) with weights varying from 19 to 30 pounds (mean of 24 pounds). The two patients with lye stricture had failed attempted dilation and had gastrostomy tubes placed to facilitate retrograde dilation and for nutrition. Their ages were 34 and 50 months and their weights were 26 and 31 pounds at the time of the operation. All patients received preoperative bowel preparation, peri-

Clinical. Patient follow-up ranged from 4 to 115 months (mean of 49 months). There were no operative deaths. The postoperative course ranged from 10 to 21 days with a period of 24 to 72 hours in the intensive care unit for endotracheal intubation and pulmonary toilet. Minor postoperative complications included atelectasis with fever in the first few postoperative days in four patients, a pneumohydrothorax on postoperative day 3 that necessitated chest tube drainage, and a cervical wound infection that necessitated incision and drainage 4 weeks after operation. Two patients (25%) with esophageal atresia had an ileoesophageal anastomotic leak at 10 and 14 days after operation, respectively. The fistulae were treated with antibiotics and topical wound care and healed spontaneously over 2- and 4-week periods. In one of the patients with an anastomotic leak a subsequent stricture developed

The Journal of Thoracic and Cardiovascular Surgery Volume 107, Number 4

I 069

Touloukian and Tellides

Fig. 2. A, Vascular supply to ileocolic conduit showing preservation of right branch of midcolic artery, marginal artery, and tertiary arcades to terminal ileum. B, Variations in ileal and colic branches of ileocolic artery are shown. Most common finding is communicating arcade (left) that allows surgical division of ileal and colic branches, whereas its absence necessitates preservation of main trunk (right) to avoid ischemia of terminal ileum. A, Artery.

Table I. Patients with retrosternal ileocolic esophageal replacement Case No. I

2 3 4 5 6 7

8

Age*

Weight*

Follow-up

Sex

Primary disease

(rna)

(lbs}

(rna)

M F F F F M F M

EA and proximal TEF EA and distal TEF; VATER EA EA; VATER EA and proximal TEF; VATER EA and distal TEF Lye stricture Lye stricture

22 24 24 19 23 45 34 50

25 24 23 19 24 30 26 31

4 7

Complication Leak Leak

15 35 65 72

80 115

Stricture Stricture

Diet]

Weight perceniilei

Regular Regular Regular Regular Regular Regular Regular Regular

20% 5% 5% 10% 10% 10% 5% 25%

EA, Esophageal atresia; TEF, tracheoesophageal fistula; VATER, vertebral defects, imperforate anus; tracheoesophageal fistula, and radial and renal dysplasia .

• Age and weight at time of operation. tDiet and weight percentile at last clinic visit.

I 070

The Journal of Thoracic and Cardiovascular Surgery April 1994

Touloukian and Tellides

Fig. 3. A, Completed retrosternal ileocolic interposition for esophageal atresia in 18-month-old infant. Barium swallow shows esophagoileal anastomosis (top arrow), ileocecal valve in middle mediastinum (center arrow), and cologastric anastomosis (lower arrow) in antrum. Barium flow is unobstructed without kink, twisting, or aspiration. B, Spot film during swallowing shows distinctive mucosal pattern of terminal ileum beyond proximal anastomosis (arrow) .

that necessitated one esophageal dilation II weeks after operation. In the two patients with lye strictures ileoesophageal strictures developed 4 and 5 weeks after operation. The strictures were resected 6 and 54 weeks postoperatively with complete healing. One patient with esophageal atresia (no anastomotic leak or stricture) and one patient with lye stricture required endoscopic removal of a swallowed coin obstructing the proximal anastomosis I and 3 years after operation, respectively. In one patient (12.5%) with esophageal atresia smallbowel obstruction developed, necessitating lysis of adhesions 7 months after operation. A nasogastric tube was passed through the ileocolic interposition under fluoroscopic guidance for preoperative gastric decompression. None of the patients had aspiration pneumonia or symptomatic gastroesophageal reflux. Gastrostomy tube feeding was resumed between the fourth and seventh postoperative days. Oral feeding was started on the ninth postoperative day after barium swallow showed the anas tomosis to be healed. All eight children have had the gastrostomy tubes removed and are tolerating regular diets. Weights range from the 5th to the 25th percentile (mean of l Oth percentile). No significant change in weight occurred after the gastrostomy tubes were removed. Radiographic evaluation. All children were studied between 10 and 14 days after operation. Swallowed bar-

ium usually emptied so quickly from the proximal esophagus that the anastomsosis was not clearly delineated on fluoroscopy, but it could be seen on the spot film (Fig. 3, A). Occasionally, some slowing was seen at the thoracic inlet and the ileocecal valve before contrast entered the colon, but dysphagia, choking, or spitting did not occur. The colon was relatively straight with a slight concavity to the right hemithorax (Fig. 3, B). Barium passage slowed once the barium reached the colon and mixed with retained secretions, but it never refluxed into the esophagus via the ileocecal valve during a 30-second to I-minute period of fluoroscopic observation. During this time, the patient was turned from the upright into the recumbent position and rotated from side to side in an attempt to cause barium to reflux from the colon into the terminal ileum. Total clearance of barium from the colon often took more than 2 minutes, but gastrocolic reflux did not occur. Discussion Colonic interposition remains an option for primary repair in the infant with esophageal atresia in which the gap exceeds six vertebral bodies. The trend in recent years, however, has been delayed primary anastomosis in conjunction with esophageal myotomy after a period of growth and elongation of the upper esophagus, a plan that

The Journal of Thoracic and Cardiovascular Surgery Volume 107, Number 4

may require preoperative hospitalization in the neonatal intensive care unit for many weeks, incur great expense, and dislocate the nuclear family. The risk of stricture and severe gastroesophageal reflux with delayed primary repair is as high as 30%,4which warrants reconsideration of esophageal replacement in selected cases of pure esophageal atresia or extreme wide-gap atresia with fistula. Advantages of cervical esophagostomy and gastrostomy include shorter-term initial hospitalization and the benefits of adequate nutrition, growth and development, and minimal, if any, respiratory complications from aspiration during infancy. Long-term follow-up of right- and left-colon substitutes that use the posterior mediastinal and retrosternal routes have shown that the colon functions principally as a conduit with ineffective peristalsis, increasing the risk of reflux and aspiration. For these reasons, we have favored the terminal ileum and right colon as first described by Javid ' in 1954, because of the potential advantages that preserving the ileocecal valve may have in preventing coloesophageal reflux and aspiration. This technique was largely abandoned by pediatric and thoracic surgeons in favor of procedures with either the retrosternal gastric tube or proximal or distal colon, which are viewed as simpler to perform and have a lower risk of ischemic leak or stricture at the upper anastomosis. The Waterston procedure.' which uses a distal coloesophageal anastomosis via a transdiaphragmatic incision, also remains popular in some centers. The benefits of certain anatomic advantages of the retrosternal ileocolic esophageal replacement procedure over the straight colon procedure include the similar caliber of the terminal ileum and cervical esophagus, which facilitates the neck anastomosis; avoidance of a thoracotomy; and preservation of the ileocecal valve to prevent reflux esophagitis from retained secretions and food in the colonic segment. Renewed enthusiasm for this operation was reported by Mao-TangS as esophageal replacement in patients with extensive lye stricture, but has not recently been advocated for infants with esophageal atresia. Previous studies in patients with the straight colon conduit have shown that peristalsis is generally ineffective and food empties largely by gravity." 7 Gastroesophageal reflux results in nocturnal coughing, choking spells, and may even cause aspiration pneumonia and colonic ulceration with bleeding.": 8 In a recent report, 60% of children with substernal colon interposition had significant reflux." Some have suggested that pyloroplasty will enhance gastric emptying and reduce the risk of reflux, 10-12 but alkaline gastritis has been reported as a complication.P For this reason, pyloroplasty was not initially done in our

Touloukian and Tellides

1 07 1

patients nor subsequently needed for either gastric retention or reflux into the colon conduit. Since an operative mortality rate of higher than 2% was reported in an early series of colon interposition.!" the morbidity has declined with modern anesthesia and postoperative care. Furthermore, the retrosternal route avoids a thoracotomy with division and repair of the diaphragm, which carries its own risk of pulmonary complications. Proximal anastomotic leak or stricture ranging from 9% in one series 13 to as high as 70% in another!" is not totally preventable. In most series the incidence of leak is approximately 30%.9 Ischemia from a redundant or angulated pedicle and tension at the anastomosis can be avoided, but perfusion of the ileal segment is variable. We retain all the tertiary arcades of the mesoileum to minimize ischemia at the distal end of the marginal artery blood supply. Unfortunately, the vascular anatomy of the ileal and colic branches of the ileocolic artery is variable enough to make surgical division of the ileocolic artery itself necessary if the terminal communicating arcade is absent. Compression of the mesoileum at the thoracic inlet must be avoided by generous blunt dissection behind the sternocleidomastoid muscle. With this method, resection of the clavicular head has been unnecessary. Staging the neck anastomosis has also been advocated to reduce the prevalence ofleak or stricture.l" but has not appeared necessary in light of our results to date. An excellent functional result was achieved in our series with all eight children eating a regular diet and growing well without choking, eructation, or the need for gastrostomy feedings. Growth percentiles remain between the 5th and 25th percentile in our series, whereas others report a tendency for the weight to fall below the 5th percentile with advancing age after straight colon interposition.P possibly because of food intolerance or eating difficulties. Two children in our group initially had some difficulty in learning to eat solid foods. For that reason, we begin "sham feedings" with the assistance of a therapist to prepare the infant to chew and swallow properly after the reconstruction. None of the infants with esophageal atresia required revision of the anastomosis, a result we attribute, in part, to our care in preserving terminal blood flow and minimizing the effects of acid reflux at the anastomosis. This has been accomplished by bringing the cologastric anastomosis close to the greater curvature of the stomach, thereby creating some angulation, which inhibits gastrocolic reflux. We found that the ileocecal valve itself also has an additional antireflux action, which prevents retained, pooled colon secretions and food from being aspirated. The barium swallow obtained routinely in these

I 072

Touloukian and Tellides

children showed no evidence of either reflux or aspiration. Late-appearing redundancy of the colon may contribute to reflux, but can be largely prevented by making the conduit as straight as possible at the first operation. Children with lye ingestion have a higher risk of the development of anastomotic strictures because of residual chronic inflammation and submucosal fibrosis in the proximal esophagus, an observation also made by Mao-TangS in his patients with esophageal stricture. We conclude that the classic retrosternal ileocolic conduit has proved to be an effective operation for esophageal replacement and may offer advantages to either the straight colon or gastric tube because ofthe antireflux role of the ileocecal valve. Certain advantages of esophageal replacement over delayed primary anastomosis with extreme wide-gap esophageal atresia have been identified. REFERENCES 1. Lunblad O. Uber antethorakale osophagoplastik. Acta

Chir Scand 1921;53:535-54. 2. Waterston DJ. Esophageal atresia. In: Gairdner D, ed. Recent advances in pediatrics. London: Churchill, 1954: 151-7. 3. Javid H. Esophageal reconstruction using colon and terminal ileum. Surgery 1954;36:132-4. 4. Puri P, Ninan GK, Blake NS. Delayed primary anastomosis for esophageal atresia. 18 months to II years' followup. J Pediatr Surg 1992;27:1127-30.

The Journal of Thoracic and Cardiovascular Surgery April 1994

5. Mao-Tang H. Ileocolic replacement of esophagus in children with esophageal stricture. J Pediatr Surg 1991;26: 755-7. 6. Othersen HB, Clatworthy HW. Functional evaluation of esophageal replacement in children. J THORAC CARDIO. VASC SURG 1967;53:55-63. 7. Schiller M, Frye TR, Boles ET. Evaluation of colonic replacement of the esophagus in children. J Pediatr Surg 1971;6:753-60. 8. Jones EL, Booth DJ, Cameron JL. Functional evaluation of esophageal reconstructions. Ann Thorac Surg 1971;12: 331-46. 9. Mitchell 1M, Goh DW, Roberts KD. Colon interposition in children. Br J Surg 1989;76:681-6. 10. Gross RE, Firestone FN. Colonic reconstruction of the esophagus in infants and children. Surgery 1967;61:955-64. II. Azar H, Chrispin AR, Waterston DJ. Esophageal replacement with transverse colon in infants and children. J Pediatr Surg 1971;6:3-9. 12. German JC, Waterston DJ. Colon interposition for replacement of the esophagus in children. J Pediatr Surg 1976;II: 227-34. 13. Hendren WH, Hendren WG. Colon interposition for esophagus in children. J Pediatr Surg 1985;20:829-39. 14. Louhimo I, Pasila M, Visakorpi JK. Late gastrointestinal complications in patients with colonic replacement of the esophagus. J Pediatr Surg 1969;4:663-73. 15. Anderson KD, Noblett H, Belsey R. Long-term followup of children with colon and gastric tube interposition for esophageal atresia. Surgery 1992;III: 131-6.