Search

Background: Injuries to the pancreas from blunt abdominal trauma in children are rare. Most are minor and are best treated conservatively. The mainstay for treatment of major ductal injuries has been prompt surgical resection. Diagnostic imaging modalities are the key to the accurate classification of these injuries and planning appropriate treatment. Computed tomography (CT) scan has been the major imaging modality in blunt abdominal trauma for children, but has shortcomings in the diagnosis of pancreatic ductal injury. Endoscopic retrograde cholangiopancreatography (ERCP) has been shown recently to be superior in diagnostic accuracy. The therapeutic placement of stents in the trauma setting has not been described in children., Methods: Two children sustained major ductal injuries from blunt abdominal trauma that were suspected, but not conclusively noted, on initial CT scan. Both underwent ERCP within hours of injury. In case 1, a stent was threaded through the disruption into the distal duct. In case 2, a similar injury, the stent could only be placed through the ampulla, thereby reducing ductal pressure. In both cases, clinical improvement was rapid with complete resolution of clinical and chemical pancreatitis, resumption of a normal diet, and discharge from the hospital. The stents were removed at 10 and 12 days postinjury, and both children have remained well. Follow-up ERCP and CT scans show complete healing of the ducts and no evidence of pseudocyst formation 1 year post injury., Conclusions: Acute ERCP should be the imaging modality of choice in suspected major pancreatic ductal injury. Successful treatment by placement of an intrapancreatic ductal stent may be possible at the same time. Surgical resection or reconstruction can then be reserved for cases in which stenting is impossible or fails.

Background/purpose: Whether laparoscopic appendectomy (LA) is superior to open appendectomy (OA) for simple (SA) and perforated appendicitis (PA) in children is debatable. The operative experience of 4 senior pediatric surgeons at a single institution was studied over a 6-year period during a transition from OA in all cases to LA in all cases, to answer this question., Methods: All appendectomies from December 1993 to December 1999 were reviewed for operative technique (OA, LA), presence of perforation (SA, PA), operating time (OT), length of stay (LOS), morbidity, and mortality., Results: There were 1,128 appendectomies in children aged 14 months to 19 years, including 955 LA (653 in SA, 302 in PA) and 173 OA (86 in SA, 87 in PA). OT was equal for LA and OA in SA (52 minutes), but has dropped to less than 40 minutes for LA in the past year. OT in PA was slightly longer in LA versus OA (68 v. 58 minutes; P < .001) but recently has dropped in LA to less than 60 minutes. LOS in SA was 2 days for LA and 3 days for OA; in PA, LOS was 7 days in both LA and OA, but has dropped to 5 days for LA recently. Postoperative abscess rates and incidence of bowel obstruction did not differ between LA and OA in either group. There was no mortality., Conclusions: LA is at least as safe and effective as, if not superior to, OA for both simple and perforated appendicitis. Postoperative pain is less, and recovery is faster, thereby reducing LOS and overall cost. The growing demand for this procedure can be satisfied without increase in cost, morbidity, or mortality. Laparoscopic appendectomy is our procedure of choice in children.

Purpose: This study of traumatic liver injuries (LI) in children reviews the authors' management and identifies the indications for surgical intervention., Methods: A total of 11,761 admissions over 13 years to a regional pediatric trauma center were surveyed., Results: LI were identified in 328 children. Mechanisms included 39% pedestrians struck by a vehicle, 34% motor vehicle occupants, 13% falls or discrete blows to the abdomen, 5% bicycle injuries, 5% child abuse, and 4% penetrating injuries. Eighty-seven of patients with LI were treated nonoperatively with a mortality rate of 17%. Six percent of deaths were attributed to the LI and massive hemorrhage, all of which presented with cardiopulmonary resuscitation (CPR) in progress, with the remainder of deaths attributed to massive head, neck, or chest injuries (HNCI, 90%) and sepsis (4%). Surgery was performed in 13% of all LI and included major hepatic vascular injury (MHVI) repairs (34%), hepatorrhaphies (34%), lobectomies (27%), and biliary repairs (4%). Excluding the nonoperative group deaths, the need for blood transfusion of more than 25 mL/kg in the first 2 hours as an indicator of surgical necessity or a MHVI had, respectively, a sensitivity of 34% and 67%, specificity of 98% and 97%, positive predictive value of 79% and 53%, negative predictive value of 89% and 98%, and prediction accuracy of 88% and 95%. In the surgical management group, the mortality rate was 23% with 40% of deaths attributed to MHVI, 30% a combination of MHVI and HNCI, 20% massive HNCI, and 10% sepsis., Conclusions: The majority of pediatric LI are not associated with hemodynamic instability and can be successfully managed nonoperatively (72%). Hemodynamic instability, as defined by the need for blood transfusion in excess of 25 mL/kg within the first 2 hours, was a strong indicator of a MHVI, which was often a lethal injury (70%).

Background/purpose: Nonoperative management of low-grade pancreatic injury is widely accepted. Management of major pancreatic parenchymal or ductal injury in children remains controversial. This study will review management strategies as they relate to site and type of pancreatic injury and their outcomes., Methods: A total of 11,794 consecutive admissions to a regional pediatric trauma center from 1984 to 1997 were reviewed to identify children with pancreatic injury as documented by serum amylase; imaging by computed tomographic (CT) scan, ultrasonography (US), and endoscopic retrograde cholangiopancreatography (ERCP); and surgical or postmortem reports., Results: Pancreatic injury was documented in 56 children, age 2 months to 14 years, with blunt mechanisms of injury. Serum amylase level was elevated on admission in 40 (71%), and no correlation was found between admission amylase values and severity of pancreas injury. An admission CT scan was obtained in 44 (79%) patients and was positive for pancreatic injury in 25 (57%). Twelve children, who had diagnoses of other intraabdominal injury by CT scan, had pancreatic injuries discovered on exploratory laparotomy. Seven children with normal admission CT scans, and the 12 children who did not undergo imaging on admission, had pancreatic injuries that were later documented by imaging, exploration, or autopsy. Thirty children were treated with immediate exploration, 17 for nonpancreatic indications and 13 with pancreatic injuries on admission CT scan. Of 19 minor injuries noted in the operating room, only three were treated with external drainage. Seven patients with distal duct injuries underwent distal pancreatectomy with splenic preservation. Twenty-six children were treated nonoperatively, including 19 with minor pancreas injuries, three with proximal pancreatic duct injuries, and four with duct injuries that were missed at admission. There were no advantages or complications of external drainage versus nondrainage of minor pancreatic injuries. There were seven deaths (overall mortality rate, 12.5%), none related to pancreatic injury. Of the seven patients in whom pseudocysts developed (two not imaged at admission, two from known proximal duct injuries treated nonoperatively, and three from injuries missed by initial studies), five were treated by delayed internal drainage and recovered promptly, and two resolved spontaneously. Pancreas-related complications included one persistent fistula and one prolonged hyperamylasemia, both of which resolved., Conclusions: Pancreatic injuries are uncommon in children. Initial serum amylase level does not correlate with the severity of pancreatic injury. The majority of pancreatic injuries can be managed nonoperatively. Initial nonoperative management of injuries of the proximal pancreatic duct allows for the formation and uneventful delayed drainage of a pseudocyst, rather than the risks of early radical interventions. Distal duct injuries are best managed by prompt spleen-sparing distal pancreatectomy. There is no benefit of closed drainage for management of minor pancreatic injuries discovered at laparotomy.

Omphalopagus twin girls were admitted for evaluation of possible separation and repair at age 7 days. Prenatal sonographic diagnosis occurred late in the third trimester and was followed by cesarean section delivery shortly thereafter. Results of extensive evaluation over the next 7 days including x-rays, computed tomography and ultrasound scan of the head and torso, and cardiac catheterization showed: the gastrointestinal tracts were separate and normal, the livers were joined but had separate biliary and vascular systems, and the hearts were separate with vastly different anatomy and function. One twin (twin A) had a normal heart with a small insignificant VSD. Twin B had a single ventricle, an incompetent A-V valve, stenotic pulmonic valve, ASD, PDA, and congenital heart block. Hemodynamic support of twin B was almost entirely from twin A. The vascular communications between the two consisted of a major connection between the internal mammary arteries and large arterial and venous connections traversing the joined livers. Because of continued deterioration of twin B, separation was undertaken at age 15 days. The separation included dividing the liver and the multiple large vascular connections. Two teams then reconstructed each twin separately. Twin B began showing signs of cardiac decompensation shortly after separation in spite of placement of a pacemaker, pulmonary artery banding, and ligation of the PDA. Cardiac function rapidly deteriorated and she died. Tissue from her chest wall was cryopreserved and placed in the tissue bank. Twin A underwent closure of her abdomen, and received a temporary bovine pericardial patch over the chest defect. She subsequently underwent placement of a graft of twin B's rib cage to bridge the bony chest defect and skin flap closure. She is presently taking a normal diet and thriving at home at age 18 months. The use of cryopreserved tissue from a syngeneic source provides a unique method of reconstruction in this situation.

Purpose: The purpose of this study was to determine whether aortic arch anomalies are associated with long gap esophageal atresia and tracheoesophageal fistula (EA-TEF)., Methods: The authors performed a retrospective review of all infants who had EA-TEF from 1980 to 1996 at two pediatric surgery centers. Two hundred three infants who had EA-TEF were identified., Results: Twelve infants were noted to have both long gap EA-TEF defined as a gap length greater than 3 cm and aortic arch anomalies. Of these 12, 7 had aberrant right subclavian arteries originating from the descending aorta. Four of the seven infants who had aberrant right subclavian artery (SCA) had gap lengths greater than 4 cm. All four had their fistulae divided initially through a right thoracotomy with primary repair performed at a later date. The remaining five infants who had long gap EA-TEF had right-sided aortic arch with aberrant left subclavian arteries. All five initially underwent exploration through the right chest. On discovery of the long gap EA and concurrent vascular anomaly, the thoracotomies were closed, and the infants underwent definitive repair of both their EA-TEF and their vascular anomaly through a left thoracotomy., Conclusions: The authors find that aortic arch anomalies are associated with long gap EA-TEF. Patients who have these two anomalies tend to have a long gap. Preoperative diagnosis of these anomalies may alter the timing and technique of surgical intervention. The embryogenesis of these vascular lesions may account for this more severe form of esophageal atresia.

Purpose: Esophageal replacement has been achieved using stomach, gastric tube, small intestine, and colon in various positions and in single or multiple stages. Long-term functional results are of prime importance in children with benign disease. The aim of this study is to present the immediate and long-term results of one-stage esophagectomy and in situ colon interposition esophageal replacement in children., Methods: Seven children have undergone one-stage esophagectomy and in situ colon interposition esophageal replacement for stricture secondary to caustic ingestion (n = 4), battery ingestion (n = 2), and epidermolysis bullosa (n = 1)., Technique: Via thoracoabdominal and cervical incisions, the transverse colon, isolated on the left colic artery, is pulled through the esophageal hiatus and normal esophageal bed into the neck at the time of simultaneous esophagectomy. The stomach is partially wrapped around the colon as an antireflux procedure and a pyloroplasty are performed., Results: There was no immediate postoperative morbidity or mortality. All patients were discharged taking a soft diet by mouth. There were no anastomotic leaks or strictures, and functional swallowing is excellent. Contrast studies show no dilation of the colon in the chest. Growth and development have been normal up to 7 years postoperatively., Conclusion: The one-stage esophagectomy and in situ colon interposition is an excellent technique for esophageal replacement in children. Continued evaluation of this technique will be necessary to confirm these preliminary results.

The protean clinical manifestations and varied complications of abdominal tuberculosis continue to challenge the diagnostic acumen and therapeutic skills of all physicians. Although abdominal tuberculosis in children has not been common in the United States over the past 2 decades, the authors found 26 case reports for the period 1980-1993. Three clinical patterns were evident: intestinal (13) peritoneal (9), and asymptomatic with incidental calcifications apparent on abdominal radiographs (4). The diagnosis was suspected for only 23% of these cases, which emphasizes the nonspecific symptomatology caused by this extrapulmonary manifestation and the need for a high index of suspicion to make a prompt diagnosis. In this study, 24 of the 26 (91%) were of Hispanic origin; the other two were indo-Chinese, another high-risk group. Most patients (88%) had a positive PPD skin test result. Mycobacteria were isolated from 15 of 21 (71.4%) cultures, with M bovis in 80% and M tuberculosis in 20%. Antituberculous chemotherapy is the mainstay of treatment; surgery is reserved for tissue diagnosis in cases of peritoneal tuberculosis and for the management of complications of intestinal tuberculosis. The response to chemotherapy usually is excellent, and long-term sequelae are uncommon. It appears that steroids do not decrease the incidence or degree of fibrosis in intestinal tuberculosis.

To identify the physiological and anatomic factors that characterize the need for operative management of blunt pediatric liver injuries, the case records of 106 pediatric trauma victims with liver injuries over a 6-year period were reviewed. Sixty-nine patients were managed without operation (nonoperative) and 37 underwent operation, 7 with penetrating and 30 with blunt liver injuries. Of these 30 patients, 21 underwent laparotomy due to blunt liver injuries (operative); the remaining 9 patients required operation due to associated intraabdominal injuries. Nine (45%) of the 21 operative patients had major hepatic vein or retrohepatic vena caval injuries, 7 of whom died. Overall mortality was 9.4% (10/106). When nonoperative and operative groups were compared, those who underwent laparotomy due to blunt liver injuries: (1) had significantly lower Champion and Pediatric Trauma Scores due to multisystem injury; (2) had 25% or greater lobar disruption with pelvic blood collections on computed tomography scan; (3) underwent early transfusion within 2 hours of admission (18/21); and (4) were frequently found to have a major hepatic vein or retrohepatic vena caval injury at the time of operation. Only one patient successfully managed without operation received greater than 30 mL/kg of blood products within 24 hours of admission. As selective nonoperative management of pediatric liver injuries gains widespread acceptance, the identification of factors that predict the need for operative intervention will limit the potential risks of delay in treatment.

Two patients, aged 2 and 6 months, underwent extended right hepatectomy for hepatoblastoma. Tumor resection was complete in both and postoperative chemotherapy was begun. One patient became markedly jaundiced 4 months postoperatively. Imaging evaluation showed marked nodular liver enlargement and intrahepatic ductal dilatation. At laparotomy an obstructed distal common duct was draped over a huge regenerating liver nodule. A high Roux-En-Y choledochojejunostomy was followed by a decrease in bilirubin to normal within 4 days. Chemotherapy was restarted and the patient remains anicteric and healthy. No evidence of tumor recurrence was found. A second patient became jaundiced 2 weeks after resection. Imaging evaluation suggested tumor recurrence. At laparotomy biopsy failed to confirm recurrent tumor but the area of the porta was not explored. Jaundice persisted and the patient died of liver failure thought secondary to tumor recurrence 4 months postoperation. Postmortem examination showed a completely obstructed common duct draped over a liver nodule and no evidence of tumor. These two patients illustrate the heretofore unrecorded entity of biliary obstruction from regenerating liver following resection for tumor. The sudden appearance of jaundice following a successful liver resection for tumor should raise the suspicion of benign repairable biliary obstruction rather than unresectable tumor recurrence. The definitive evaluation is reexploration rather than various imaging modalities, which may not only be misleading, but may delay life-saving intervention.

Nine infants underwent operations for perforation of a Meckel's diverticulum over a 21-yr period. All of the infants were acutely ill when initially seen by the surgical service. Lethargy, irritability, anorexia, fever, abdominal tenderness, and passage of blood in the stools were common clinical manifestations. The preoperative diagnosis of perforated Meckel's diverticulum was not made in any of the cases. Skilled preoperative and postoperative management is important for the survival of these infants. There was no mortality and minimal morbidity. In retrospect, the peptic etiology of perforation in this age group is a key to early recognition. The past history of most of the patients includes previous episodes of blood in the stools and episodic abdominal complaints.

Previous attempts to produce sustained portal hypertension in experimental animals by portal venous obstruction have not been successful. Experiments were designed to study the effects of gradual portal venous occlusion plus hepatic lymphatic ligation in the rat. Adult male Lewis rats were divided into three groups: (A) nonoperated or sham-operated; (B) operated, with hepatic lymphatic ligation only; and (C) operated, with placement of a portal vein ameroid constrictor and hepatic lymphatic ligation. There were no differences noted in portal venous pressures (control 9.6 +/- 0.4 cm H2O) and portovenograms between groups A and B throughout the study. In group C there was greater than doubling of portal venous pressure by 1 wk which remained for 2 mo. Portal pressure then fell but remained elevated at greater than 30% above control values for 1 yr. Three systems of collateral circulation developed rapidly in response to this extrahepatic portal venous occlusion: (A) veins directly bridging across the occluding ameroid reestablishing hepatopetal flow which steadily increased; (B) spontaneous portosystemic shunts at the splenorenal area that occur early and then regress; and (C) retroperitoneal, paraesophageal and submucosal esophageal veins that develop late and remain stable. Quantitative assessment of the lumen size of the submucosal esophageal veins revealed a twofold increase in these veins in group B when compared to controls by 4 wk and a fivefold increase in these veins in group C by 8 wk that persisted for 1 yr. This model appears useful for studying the course and effects of extrahepatic portal hypertension and its gradual alteration by the spontaneous development of portosystemic collaterals in the rat. This model may have enough similarities to the extrahepatic portal hypertension seen in children that its use for future studies may be fruitful.

Selective hepatic dearterialization for exsanguinating liver injury has been performed in two young children without untoward sequelae. It appears to be a safe, simple, life-saving maneuver, as has been shown in adults. Simultaneous cholecystectomy and drainage of the extrahepatic biliary tree does not seem necessary. Postoperative studies in these two patients showed normal liver functions, normal liver scintiscans, and in one case, early rearterialization of the devascularized lobe by translobar collaterals. We would urge the use of this relatively simple procedure in place of major hepatic resection for selected severe hepatic injuries.

In a series of 49 consecutive cases of Wilms' tumor from a single institution, there was a 20% incidence of nonvisualization on the initial intravenous pyelogram. Even though nonvisualization was secondary to either gross or microscopic invasion of the renal vein or renal pelvis, this did not forecast a grim progosis. Seven of the nine patients with nonvisualization are presently alive and tumor-free at least 5 yr after resection. The only two deaths in this particular group were unrelated to the Wilms' tumor.

The optimal surgical management of neonates with gastroschisis and omphalocele remains controversial. Early attempts at primary closure were limited by respiratory and vena cava compromise and were gradually replaced by skin closure only, staged reduction of implanted prostheses, and escharification of the intact omphalocele. Advances in surgical technique, neonatal intensive care, and mechanical ventilation have made primary closure a viable alternative once again. We wish to report a consecutive series of 73 infants, including 54 with gastroschisis and 19 with omphalocele, over a 16-year period. Fifty gastroschisis patients were closed primarily, with four deaths (8%), and four with staged prostheses, with one death (25%). Sixteen omphaloceles were closed primarily, with three deaths (19%); two closed with a staged prosthesis, and one patient, closed with skin flaps only, died (100%). Overall survival for primary closure was 59/66 (89%) and for staged reduction, 5/7 (43%). In the past 5 years all patients have been closed primarily with no deaths (100% survival). All survivors were alive and well 1 to 15 years postoperatively. Our current operative technique involves vigorous manual stretching of the abdominal cavity, manual emptying of the entire Gl tract, and primary fascial closure, with no gastrostomy. Key points in postoperative management include total paralysis and mechanical ventilation, as the initially tense abdominal wall softens over 1 to 2 days, followed by weaning from the ventilator, and vigorous fluid support. Gl function returns more rapidly after primary closure (average of 3 days) and hospital stay is not prolonged. Other complications (intestinal obstruction, two patients; fistulae, one patient, etc) are much less frequent.(ABSTRACT TRUNCATED AT 250 WORDS)

Severe acquired subglottic stenosis occurs most commonly in infants and children who require long-term ventilatory support with indwelling endotracheal tubes for underlying respiratory disease. A variety of operative and endoscopic procedures have been advocated for this problem including endoscopic fulguration, cryotherapy, direct incision or excision, and several types of extensive direct laryngotracheoplasties. The failure rate with these procedures is high. We have treated 22 patients including 12 neonates and 10 older children with acquired airway injury and obstruction with a simple cricoid split. In 15 patients the airway obstruction was completely relieved and these youngsters were extubated without difficulty. An additional 3 patients failed initial extubation which was subsequently successful, however, after repeat intubation for a short period of time. Endoscopic follow-up shows complete healing of the incised area. In 2 neonates the procedure was unsuccessful, and tracheotomy was needed. One of these had additional severe airway injury in the distal tracheobronchial tree secondary to selective bronchial intubation in the newborn period. Two older children also required tracheotomy because of severe airway scarring unrelieved by this procedure. The cricoid split is a simple and successful way of dealing with subglottic stenosis, especially in the newborn premature infant. A significant advantage is that it disturbs the anatomy very little, allowing for more extensive laryngotracheoplasty in the future should it fail.

This is a report of a case in which a significant quantity of sulfuric acid was ingested, causing life threatening respiratory distress, secondary to glottic injury and edema, and cardiovascular collapse. Vigorous resuscitive efforts allowed survival through the first 48 hours and two major operations. Persistent and complex surgical management eventually restored oropharyngeal competence and gastrointestinal integrity. Survival of a child from such an overwhelming injury has not been previously reported.

Total abdominal colectomy and endorectal ileal pull-through have been widely used in children as a definitive continence-preserving procedure for ulcerative colitis (UC) and familial polyposis (FP). Controversy exists in regards to the results achieved with the various techniques utilized to accomplish this procedure, including straight ileal pull-through, pull-through with balloon dilatation of the lower ileal segment, and the construction of a variety of ileal reservoirs: S-shaped, J-shaped, or lateral. We have utilized the two-stage lateral reservoir approach advocated by Fonkalsrud et al, on a total of 12 patients, aged 1 to 17 years, including ten patients with UC and two with FP. Twelve patients have undergone stage 1 and seven of these stage 2. Numerous complications of stage 1 occurred early in the series, including breakdown of the ileoanal anastomosis, 3 patients; cuff abscess, 2 patients; postoperative intestinal obstruction, 2 patients; and intraabdominal abscess, 1 patient. All have fully recovered. The last six stage 1 procedures have been without complications. Five of the seven stage 2 patients are now well, continent, and having four to six controlled bowel movements a day 1 to 2 years postoperatively. Two patients are using small amounts of Immodium and Metamucil. One patient has developed marked dilatation of the pouch at 8 months postoperatively with episodic "pouchitis" requiring operative revision and diversion. One stage 2 patient suffered numerous complications including spontaneous perforation of the reservoir 2 months postoperatively, eventually leading to re-creation of a permanent ileostomy and removal of the pouch. Five patients await further treatment. The lateral reservoir technique is capable of producing a satisfied and continent patient; however, the potential technical problems and early complications are significant. The occurrence of delayed "pouchitis" is worrisome. Although the initial results with the lateral reservoir technique appear encouraging, further longterm evaluation is necessary.

Ten surgical neonates with postoperative intractable diarrhea and secondary weight loss were treated with combination cholestyramine and paregoric therapy. Within 3-5 days all infants except two showed significant clinical improvement with a decreasing number of stools, an increase in the consistency of the stool, and gradual weight gain. The exact mechanism of action of cholestyramine is not clear. It may act by binding with bile salts and/or endotoxins in the bowel lumen or decreasing the motility of the bowel. Used in combination with paregoric, a known bowel motility depressant, the doses of each medication can be kept quite low thus avoiding undesirable side effects. Medium chain triglyceride formula is helpful in some of these infants to improve fat absorption further. Medication in all of these infants has been discontinued without any adverse effects.

The treatment of biliary atresia by variations of the original Kasai hepatoportoenterostomy has shown early success with good bile flow and the elimination of jaundice in 50% to 70% of cases in many series. Long-term follow-up in many of these patients shows continued problems with ascending cholangitis and progressive liver disease leading to death. Our recent experience with a modified Sawaguchi hepatoportoenterostomy is encouraging. Twelve patients were operated on before two months of age. All but one became jaundice free within 2 to 4 months and had biliary intestinal continuity reestablished within 3 to 6 months. These 11 patients have remained jaundice free with normal growth and development 1 to 8 years postoperatively. Two patients had one and two episodes of cholangitis, respectively. All have continued mild elevations of hepatocellular enzymes but no patient has obvious signs of liver failure. Serial liver biopsies have shown clearing of bile stasis and continued periportal fibrosis. Size and number of ductules in the excised biliary remnant did not correlate with clinical outcome. One patient remained jaundiced after hepatoportoenterostomy and reoperation, and eventually expired. In contrast, two patients operated at 4 and 9 months of age never drained bile and eventually died of bleeding varices and hepatic failure, respectively. The atypical success and relative lack of cholangitis in this series is not readily explained, but may be related to specific technical modifications of the original Sawaguchi procedure.

The original Duhamel procedure has undergone several important modifications over the years by Martin, Ravitch, Soper, Talbert, and others. Combining the best features of these various modifications, a modified Duhamel procedure has been developed and utilized in 41 consecutive infants and children with all types of Hirschsprung's disease, including three with long segment involvement. The procedure includes the use of a formal upper and lower anastomosis and division of the spur by the GIA stapling device. There were no deaths and no early complications. At 2-7 yr all children are toilet-trained, continent, and growing well. With appropriate modifications, the Duhamel operation is a very satisfactory definitive procedure for Hirschsprung's disease in infancy and childhood.