UPPER GASTROINTESTINAL FIBEROPTIC ENDOSCOPY IN PEDIATRIC PATIENTS

Transcription

1 GASTROENTEROWGY 72:244-24, 977 Copyright 977 by the American Gastroenterological Association Vol. 72, No.6 Printed in U.s.A. UPPER GASTROINTESTINAL FIBEROPTIC ENDOSCOPY IN PEDIATRIC PATIENTS MARVIN E. AMENT, M.D., AND DENNIS L. CHRISTIE, M.D. Departments of Pediatrics and Medicine, Division of Pediatric Gastroenterology, UCLA Center for Health Sciences, Los Angeles, California One hundred and forty-two pediatric patients,between age month and 2 years had 63 endoscopic procedures. Of66 with chronic abdominal pain, 2 had a source identified endoscopically that was seen in only 5 by esophagogram and upper gastrointestinal series. Of 3 with nausea, vomiting, dysphagia, and/or odynophagia and retrosternal pain, endoscopy demonstrated the source in 9 patients and radiographic studies in 4. Of 34 with hematemesis and/or melena, 26 had a bleeding site identified endoscopically, but only 4 of 2 had an identified source by radiographic studies. Duodenal and gastric ulcers and hemorrhagic gastritis were the commonest of upper gastrointestinal bleeding and organically of chronic abdominal pain. Functional abdominal pain was the commonest cause of chronic abdominal pain in those endoscoped. Foreign bodies were removed from the esophagus and stomach of6 patients and dislodged in 2 others. Caustic ingestion was recognized in the esophagus and stomach of 2 patients who did not have mouth burns. The GIF-P 2 -prototype with four-way tip control and ability to retroflex up, 6 down, and right and left was superior to GIF-P and CF-P-prototype for visualization of the entire esophagus, stomach, duodenal bulb, and postbulbar area in patients less than years old. Visualization of the duodenal bulb was possible in 2 of 29 pediatric patients, and of the postbulbar area in 25 of 26 in whom it was attempted. Infants who weighed as little as 3 to 5 kg were successfully examined. Retroflexion was possible in 29 of 3 to see the fundus and cardioesophageal junction. Patients older than years were better examined with the GIF-D because of its increased ability to transmit light. Sedation for the school-age child with.5 to. mg per kg of diazepam and to 2 mg per kg of meperidine given intravenously provides excellent sedation in most instances. General anesthesia is preferable for the preschooler and infant. Minor complications occurred in 2 patients who received less than adequate sedation and in patient with general anesthesia. Upper gastrointestinal fiberoptic endoscopy is increasingly being used as a diagnostic tool for gastrointestinal diseases l - 7 and for removal of foreign bodies from the esophagus and stomach in the pediatric age group.s However, the commercially available pediatric instruments have limitations in their ability to visualize all parts of the esophagus, stomach, duodenum, and postbulbar area. The purposes of this paper are: to document the capabilities and limitations of commercially available and prototype pediatric fiberoptic upper intestinal endoscopes, to report on endoscopic findings and removal of foreign bodies from 42 pediatric patients, to compare endoscopic observations with results of barium contrast studies, and to report on the safety and ade- Received June 2, 976. Accepted December 2, 976. Address requests for reprints to: Marvin E. Ament, M.D., MDCC , UCLA Center for Health Sciences, Los Angeles, California 924. This study was supported in part by Grant AM 732 from the National Institute of Arthritis, Metabolism and Digestive Diseases to CURE (The Center for Ulcer Research and Education). 244 quacy of the procedure with both sedation and general anesthesia. Patients and Methods One hundred and forty-two patients ranging in age from month to 2 years underwent diagnostic upper gastrointestinal fiberoptic endoscopy (33) or removal offoreign bodies (9) from the esophagus or stomach with such instruments during a 33-month period. Eleven patients were less than 6 months old. Ten were 6 months to 2 years old; 22 were 2-5 years old; 32 were 5 to years old; and 66 were to 2 years old. A total of 63 procedures was performed. Multiple endoscopic procedures were performed in certain patients to determine whether ulcers, esophagitis, or hemorrhagic gastritis recognized endoscopically but not roentgenographically had healed. Patients examined were grouped by symptoms. Observations were tabulated comparing endoscopic to radiographic observations. Group contained 66 patients with abdominal pain with or without associated vomiting. The pain was midepigastric or periumbilical in location and was of 3-month to 2-year duration. Almost all patients had missed an average of month of school. Group 2 consisted of3 patients with one or more of the following symptoms: nausea, vomiting, dysphagia, and/or

2 June 977 UPPER GASTROINTESTINAL FIBEROPTIC ENDOSCOPY IN PEDIATRIC PATIENTS 245 TABLE. Physical characteristics of upper intestinal {iberscopes Instrument Focus Length Diameter Tip control Field of Bending angle Bending angle Biopsy view channel cli- Up/down Right/left ameter em mm mm GIF-D Adjustable way 75 5 / GIF-P Fixed 7.2 (shaft) 7. (bending area) GIF-P 2 Fixed.6 (shaft).2 (bending area) CF-P Fixed.6 F-7 Fixed GIF-J Adjustable 35 (shaft) 2 way 65 5 /5 % 2. 4 way 75 /6 / 2. 4 way 75 5 /5 / 2. 4 way 62 / / 2. 4 way 7 2 /2 2 /2 2. odynophagia, and retrosternal pain (heartburn). Patients with only vomiting as their chief symptom were infants who vomited sufficiently after feeding to impair weight gain. Group 3 contained 34 patients who had hametemesis and/or melena or nasogastric aspirate which was bloody or coffeeground in consistency. Eighteen of 34 developed shock; 4 did not have a significant decrease in hematocrit defined as a fall of greater than 4 %. Group 4 consisted of patients who had accidentally swallowed foreign bodies or were suspected to have ingested caustics. Forty-three patients had sedation and 2 had general anesthesia. Patients were prepared for upper gastrointestinal endoscopy with sedation using () atropine sulfate (. mg per kg), and (2) meperidine (2. mg per kg) to a maximum dose of 5 mg given intramuscularly 45 min before the procedure. The pharynx was anesthetized in two steps: patients first gargled with ounce of.5% Pontocaine and then had their throats sprayed with Cetacaine spray containing 2% tetracaine hydrochloride. A 2-gauge heparin lock was inserted into one of the arm or hand veins. Diazepam (II alium, Roche Laboratories, Nutley, N. J.) was administered for sedation. A dose of.5 to. mg per kg was the usual range required to provide optimal sedation. Teen-agers were particularly difficult to sedate and some required much larger doses of sedation to accomplish the procedures. Patients were prepared for general anesthesia with atropine sulfate (. mg per kg) and hydroxyzine HCl.25 mg per kg (Vistaril) administered intramuscularly 45 min before general anesthesia. The endotracheal tube was taped in the right corner of the mouth and the patient was turned to the left lateral decubitus position. In some infants we found that the endoscope passed more easily in the supine position. The fiberscope was then passed behind the endotracheal tube, the tip was deflected downward, and the instrument was advanced into the esophagus. If light was seen in either corner of the neck the tube was then withdrawn and repositioned. InstrumentS. * The physical characteristics of the fiberscopes used are shown in table. The pediatric fiberscopes: GIF-Pt. CF-P prototype, GIF-P 2 prototype, and the F -7 were primarily used for patients less than years old and for older patients who did not tolerate use of the standard adult instrument, the GIF-D. The GIF-D was primarily used for patients years and older and for removal of foreign bodies. Before the CF-P and GIF-P 2 became available, we were able to use the GIF-D safely in patients as young as 5 years when they were examined under general * The GIF-PJ> GIF-P 2, GIF-D, GIF-J, and CF-P fiberscopes are products of Olympus Corporation of America, New Hyde Park, New York; the F-7 fiberscope is a product of American Cystoscope Makers, Inc., Pelham Manor, New York. anesthesia. However, when the GIF-P 2 became available it was used for removal offoreign bodies. The J scope was used to cannulate the ampulla of Vater in 2 patients. Whenever it was feasible the entire esophagus, stomach, duodenal bulb, and postbulbar areas were examined. Radiographic studies. An upper gastrointestinal series was performed in 33 patients. Results of visualization with different endoscopic instruments. We were unable to visualize the esophagus and stomach in their entirety in all of the 3 patients in whom the GIF PI was used. Retroflexion of the tip upward to examine the fundus and cardioesophageal junction was all but impossible. We were able to visualize the entire stomach and esophagus in 47 of 49 patients with the CF-P, and in 29 of 3 with the GIF P 2 The duodenal bulb and postbulbar areas were visualized in 43 of 49 patients with the CF-P. With the GIF-P 2, visualization of the duodenal bulb was possible in 2 of 29, and the postbulbar area in 25 of26 in whom it was attempted. The GIF-P 2 has been successfully used to remove two foreign bodies. It was easily passed into the postubular area in 6 infants less than 6 months of age, 2 of whom weighed between 3 and 5 kg. Failure to pass the CF-P and GIF-P 2 into the duodenal bulb and postbulbar areas was attributable to technical inability. The F -7 was used in 3 patients who also had upper intestinal endoscopy with the CF-P. It did not provide an optimal visualization of the esophagus and stomach. The GIF -J was used successfully to cannulate the ampulla of Vater in 2 patients ages 9 and 2 years. The GIF-D was used in 6 patients for esophagoscopy and esophagogastroduodenoscopy and in 6 for removal of foreign bodies. Complete examination of the esophagus and stomach was possible in 5. The duodenal bulb was visualized in 63 of 65, and the postbulbar area in 5 of 63. Failure to complete the examination was secondary to inadequate sedation in all. Morbidity. Morbidity occurred in 6 patients. Two patients, ages and 9 years, while being examined with the GIF-D using sedation, became restless and started struggling; this resulted in gastric mucosal tear. Neither patient had significant detectable bleeding or perforation and were discharged the following day. Two patients who had general anesthesia had fever to 39 C in the immediate 2 hr after the procedure. They were afebrile the following day. Both patients who had cannulation of ampulla had fever, vomiting, and abdominal pain for 24 hr after the procedure and abnormal elevation of urinary amylase and serum amylase. Results Group. A source for abdominal pain could be identified in 2 of 66 patients (table 2). Of 4 gastric ulcers identified by endoscopy only was seen on upper gastrointestinal series. Of duodenal ulcers identified en-

3 246 AMENT AND CHRISTIE Vol. 72, No.6 TABLE 2. Group : diagnoses established by radiology and endoscopy in 66 patients with abdominal pain with or without associated vomiting Diagnosis No abnormalities found Duodenal ulcer 4" Gastric ulcer 4 4 Gastritis 4 4 Pancreatic rest 2 2b 2 Duodenitis 2 2 Crohn's disease " Four other patients had deformed duodenal bulbs recognized in upper gastrointestinal series. b Both lesions identified by barium contrast studies were called ulcers but they proved to be mass lesions. doscopically 4 were correctly identified on upper gastrointestinal series. In 4 other the bulb was called deformed but no ulcer niche was seen. Forty-six patients did not have endoscopic abnormalities; 7 of these were reported to have pylorospasm, an irritable bulb, or a duodenal ulcer in the upper gastrointestinal series. None of 4 patients identified as having friable gastric mucosa confined to either the antrum or body and fundus had a roentgenographic diagnosis of gastritis. The patient with a pancreatic rest in the duodenal bulb had been thought to have a duodenal ulcer. Group 2. Esophagitis was the diagnosis most frequently missed in esophagograms in group 2 (table 3). It was present in all but patient with a stricture. Strictures were identified in of radiologically and in of endoscopically. Achalasia patients were characterized endoscopically by a nonrelaxing lower esophageal sphincter that would yield to gentle but firm foreward thrust of the instrument. Group 3. Table 4 describes patients in group 3. All patients who had sufficient hemorrhage to decrease the hematocrit greater than 4% had their source of bleeding identified. Endoscopy identified the source of bleeding in 2 of36. Of the others, 2 with Meckel's diverticulum were identified at exploratory laparotomy and 2 with epistaxis were identified by nasopharyngeal examination when they re-bled. A bleeding source ultimately was not identified in 5.9% of. All gastric and duodenal ulcers were identified endoscopically, and in patient with pyloric stenosis and gastrointestinal hemorrhage bright red blood was seen to reflux back into the stomach from the duodenum. Only of 3 duodenal and of 5 gastric ulcers were identified by upper gastrointestinal series. The esophageal mucosal tear occurred at the gastroesophageal junction and was secondary to use of a vinyl plastic nasogastric tube which had stiffened and had torn the mucosa. Gastritis was diffuse in 2 patients and was confined to the body and fundus in the remainder. It was not identified by barium contrast studies. One patient with recurrent aphthous ulcers developed superficial mucosal erosions in the postbulbar area. Esophagitis was recognized in of 3 patients on esophagogram that had been identified endoscopically. Barrett's esophagus was diagnosed in of these 3. Group 4. Data are shown in table 5. Foreign bodies were effectively removed from 6 of 9 patients. In the 2 patients in whom they could not be removed, their adhesions were broken and they were pushed into the duodenal bulb. Evidence of caustic ingestion was found in the esophagus and stomach of two children who did not have any burns in their mouths. Foreign bodies could not be removed with the GIF-P t because of its lack of maneuverability; the GIF-D, CF-P, and GIF-P 2 were equally effective in removing objects. Discussion Upper gastrointestinal fiberoptic endoscopy is a more useful tool for diagnosis of upper gastrointestinal disease than barium contrast studies in almost all conditions,, and provides a means to remove foreign bodies from areas of the stomach but reached by rigid endoscopic instruments. It represents one of the major diagnostic advances in pediatric gastroenterology in this decade. TABLE 3. Group 2: diagnosed established by endoscopy and radiology in 39 patients with nausea, vomiting, dysphagia, odynophagia, and retrosternal pain a Diagnosis Esophagitis 2 2 Esophageal stricture No abnormalities found 9 4" 9 Esophageal polyp Esophageal ulcer Achalasia Duodenal ulcer " These 4 patients were reported to have pylorospasm; antral spasm, or gastroesophageal reflux. TABLE 4. Group 3: diagnoses established by endoscopy and radiology in 3 patients with hematemesis and/or melena or nasogastric aspirate which was bloody or coffee-ground in consistency Diangosis Duodenal ulcer Gastric ulcer Hemorrhagic gastritis No diagnoses found Esophagitis Esophageal varices Esophageal mucosal tear Epistaxis" Meckel's diverticulum b Esophageal ulcer Postbulbar ulcers Submucosal hemorrhage caused by retching " nasopharyngeal examination. b Identified at exploratory laparotomy of3 lof5 6 3

4 June 977 UPPER GASTROINTESTINAL FIBEROPTIC ENDOSCOPY IN PEDIATRIC PATIENTS 247 TABLE 5. Group 4: foreign bodies Case Foreign body or ingestion Removed en- Instrument no. doscopically used Pin in gastric antrum Yes GIF-D 2 Penny in esophagus Yes GIF-D 3 Multiple foreign bodies in No GIF-D stomach and duodenum 4 Overcoat button in stomach No GIF-D 5 Nickel in stomach Yes GIF-D 6 Quarter in stomach No GIF-D 7 Penny in stomach Yes CF-P Quarter in stomach Yes GIF-P. 9 Cherry pit in esophageal stric- Yes GIF-P. ture Drano No CF-P Drano No GIF-P. Adhesions broken, foreign bodies pushed into duodenal bulb or antrum-came out below. Diagnostic capability depends on the instruments used. In our experience the earliest pediatric fiberscope, the GIF-Pt> was unsatisfactory for examination of the smallest infants and children and for removal of foreign bodies because of its stiffness and the limited ability to manipulate the tip. The CF-P gave us the four-way tip control a flexible endoscope must have to be effective. However, it still was not optimal because its tip did not retroflex upward and the shaft was too inflexible. The GIF-P 2 with its ability to retroflex upward, its four-way tip control, its wider field of view, and the shaft's more optimal flexibility gave us the ideal instrument to use. The F-7, another pediatric endoscope with four-way tip control, may have promise. However, our experience with its use was limited and insufficient for critical evaluation. Two of 42 patients who received sedation developed the complications of gastric mucosal tear because they became hyperactive during the procedure. Four of the procedures were prematurely terminated because of paradoxical hyperactivity secondary to diazepam. These patients had initially received sedation at doses recommended in two other reports. 5 7 However, in all other patients larger doses of sedation were routinely and safely administered. No patient developed respiratory depression. The dsoe of diazepam needed to sedate and endoscope children safely is greater on a milligram per kilogram basis than that required for adults. This experience is not unique to us but has been observed by others (J. Vanderhoof and W. Liebman). Problems with sedation have been reported 3 7. in similar patients but not to the extent we encountered. Further studies will have to be done in preschool children to determine how safely the pediatric endoscopes may be passed using sedation. General anesthesia was safer, allowed for an easier examination, and was associated with minimal postendoscopy morbidity. Its disadvantages are that it adds to the cost and time of the procedure. However, it may still be the safest way to endoscope infants and preschool children. Functional abdominal pain was the commonest cause of chronic abdominal pain in those who were endoscoped. An endoscopic abnormality was not found in 45; 7 had been diagnosed as having pylorospasm or duodenal ulcer by barium contrast studies. These false-positive reports resulted in patients being treated for peptic ulcer disease when it was not necessary. Finding a normal endoscopic examination enabled us to discontinue medications for ulcer disease and pylorospasm. Duodenal and gastric ulcers were the commonest cause of organic disease in patients with chronic abdominal pain who were endoscoped. Gastric ulcers were diagnosed on barium contrast studies only 25 % of the time and duodenal ulcers in 5% with barium contrast studies. A deformed bulb was recognized roentgenographically in the others diagnosed endoscopically. These observations are comparable to those reported by others. 5 We were surprised to find hemorrhagic gastritis in 4 patients because we thought they had functional abdominal pain or irritable bowel syndrome as the cause of their symptoms. Two of them did have pain characteristic of acid peptic disease. We are uncertain of the relationship of the lesions to the abdominal pain in the remainder of the patients. Although esophageal strictures were easily identified by esophagograms in most patients, patient diagnosed by contrast study as having a dysautonomic upper esophagus at endoscopy was found to have a stricture in the upper one-third of the esophagus and reflux esophagitis, and another patient said to have a normal esophagogram had diffuse narrowing and esophagitis. Esophagitis was consistently missed in all patients with strictures shown by contrast studies and was only recognized in a single patient with monilial esophagitis. Prior to fiberoptic endoscopy, esophageal varices were the most frequently reported source of upper gastrointestinal bleeding in' children. In nearly one-third of reported a source was never identified, whereas in our series it was 6%..2 Our experience revealed that gastric ulcers, duodenal ulcers, and hemorrhagic gastritis are the commonest causes, and that bleeding esophageal varices are rather uncommon. The patients with hemorrhagic gastrititis surprised us because other diagnoses had been anticipated in all of these. In those with esophagitis, esophageal ulcer, and esophageal mucosal tear, endoscopy alone established the diagnosis. Diagnoses were not established in 4 small infants who did not have a significant fall in hematocrit. The results of our findings and those of others reporting endoscopically identified sources of gastrointestinal. 5 bleeding materially alter previously reported series on upper gastrointestinal bleeding in children.. 2 Most foreign bodies will spontaneously pass through the gastrointestinal tract. However, sharp objects should be removed if they do not promptly pass out of the stomach to prevent perforation of the gastric mucosa. Objects which cause symptomatic obstruction to the esophagus and stomach should be removed as soon as possible. Several conclusions may be reached from our study: the GIF-P 2 is the superior pediatric upper intestinal

5 24 AMENT AND CHRISTIE Vol. 72, No.6 fiberoptic endoscope we used for visualizing the entire esophagus, stomach, duodenal bulb, and postbulbar areas in patients less than years old because of its four-way tip control, the ability to retroflex the tip upward, and the greater flexibility of the shaft. In patients older than years, standard fiberoptic endoscopes are preferable because of their increased capacity to transmit an image. Upper gastrointestinal fiberoptic endoscopy should be the first diagnostic tool used for the diagnosis of upper gastrointestinal bleeding because it is the most sensitive means to establish the diagnosis. Ulcers and gastritis are the commonest causes of upper gastrointestinal bleeding in children and teen-agers. An esophagogram should be done first in patients with retrosternal burning, odynophagia, and vomiting associated with failure to gain weight and grow. Upper gastrointestinal and small bowel series should be done first in the patient with chronic abdominal pain. Endoscopy should be considered for patients with negative or equivocal radiographic studies in whom the symptoms are severe and a functional diagnosis is not thought probable. The technique should be used for removal of sharp foreign bodies which fail to leave the stomach rapidly, and for others which cause obstructive symptoms. Asymptomatic foreign bodies should be removed from the stomach endoscopically if they do not leave it within a month. Sedation with meperidine and diazepam may be safe for some patients but it has not been established as safe for the youngest infants. Doses of intravenous diazepam greater than to 5 mg are usually necessary in the preadolescent and teen-age groups to accomplish a satisfactory examination and prevent complications. General anesthesia, we believe, may still be the safest anesthesia for the infant and preschool child. REFERENCES. Ament ME, Gans SL, Christie DL: Experience with esophagogastroduodenoscopy in diagnosis of 79 pediatric patients with hematemesis, melena or chronic abdominal pain (abstr). Gastroenterology 6:5, Ament ME, Christie DL, Liebman WM: Fiberoptic endoscopy of the upper gastrointestinal tract in pediatrics (abstr). Clin Res 23:44A, Goldstein PD, Tedesco F J, Gleason W A: Gastrointestinal fiberendoscopy as a diagnostic tool in infants and children (abstr). Gastroenterology 6:997, Gleason W A, Tedesco F J, Keating JP, et al: Fiberoptic gastrointestinal endoscopy in infants and children. J Pediatr 5:-3, Tedesco FJ, Goldstein PD, Gleason WA, et al: Upper gastrointestinal endoscopy in the pediatric patients. Gastroenterology 7: , Gans S, Ament ME, Christie DL, et al: Pediatric endoscopy with flexible fiberscopes. J Pediatr Surg :375-3, Cremer M, Peeters JP, Emonts P, et al: Fiberendoscopy of the gastrointestinal tract in children: experience with newly designed fiberscopes. Endoscopy 6:6-9, 974. Christie DL, Ament ME: Removal offoreign bodies from esophagus and stomach with flexible fiberoptic panendoscopes. Pediatrics 57:93-934, Cotton PB, Rosenberg MT, Waldrom RPL, et al: Early endoscopy of the esophagus, stomach and duodenal bulb in patients with hematemesis and melena. Br. Med J 2:55-59, 973. Cotton PB: Fiberendoscopy and the barium meal-results and implications. Br Med J 2:6-65, 973. Spencer R: Gastrointestinal hemorrhage in infancy and childhood: 476. Surgery 55:7-734, Collins REC: Some problems of gastrointestinal bleeding in children. Arch Dis Child 46:-2, 97