Objective. To review retrospectively the combined clinical experience with the surgical treatment of persistently symptomatic gastroesophageal reflux (SGER) in childhood from seven large children's surgery centers in the United States.
Design. During the past 20 years, 7467 children <18 years of age underwent antireflux operations for SGER at the seven participating hospitals. Fifty-six percent were neurologically normal (NN) and 44% were neurologically impaired (NI). The most frequent diagnostic studies were upper gastrointestinal series (68%), esophageal pH monitoring (54%), gastric emptying study (32%), and esophagoscopy (25%). The age at operation was under 12 months in 40% and 1 to 10 years in 48%. The type of fundoplication was Nissen (64%), Thal (34%), and Toupet (1.5%). A gastric emptying procedure was performed on 11.5% of NN patients and 40% of NI patients. Laparoscopic fundoplication was performed on 2.6% of patients.
Results. Good to excellent results were achieved in 95% of NN and 84.6% of NI patients. Major complications occurred in 4.2% of NN and 12.8% of NI patients. The most frequent complications were recurrent reflux attributable to wrap disruption (7.1%), respiratory (4.4%), gas bloat (3.6%), and intestinal obstruction (2.6%). Postoperative death occurred in 0.07% of NN and 0.8% of NI patients. Reoperation was performed in 3.6% of NN and 11.8% of NI patients. The results and complications were similar among the participating hospitals and did not seem related to the type of fundoplication used.
Conclusion. The excellent results (94% cure) and low morbidity with gastroesophageal fundoplication with or without a gastric emptying procedure from a large combined hospital study indicate that operation should be used early for SGER in NN children and to facilitate enteral feedings and care in NI children.
- GER =
- gastroesophageal reflux •
- NN =
- neurologically normal •
- NI =
- neurologically impaired
Gastroesophageal reflux (GER) is one of the most frequent symptomatic clinical disorders affecting the gastrointestinal tract of infants and children as well as adults.1 During the past 2 decades, GER has been recognized more frequently because of increased awareness of the condition and because of the more sophisticated diagnostic techniques that have been developed for both identifying and quantifying the disorder. Gastroesophageal fundoplication is currently one of the three most common major operations performed on infants and children by many pediatric surgeons in the United States.2 Nonetheless, the indications for operation in the era of proton pump inhibitors have undergone reassessment, and the results after antireflux surgery have varied in previous reports.3-4 The present study was therefore undertaken to provide a very large clinical experience in the surgical treatment of GER by combining the clinical data from seven large children's surgery centers from diverse regions in the United States.
PATIENTS AND METHODS
The surgical treatment of GER in children under the age of 18 years during the 20-year period from 1976 through 1996 was retrospectively reviewed by sending a questionnaire to the Chief of Pediatric Surgery from each of the following children's medical centers: Cardinal Glennon Children's Hospital in St Louis, Missouri; Children's Hospital in Oklahoma City, Oklahoma; Children's Mercy Hospital in Kansas City, Missouri; Cook Fort Worth Children's Hospital in Fort Worth, Texas; C. S. Mott Children's Hospital in Ann Arbor, Michigan; J. W. Riley Children's Hospital in Indianapolis, Indiana; and the UCLA Children's Hospital in Los Angeles, California. A combined total of 7467 patients who underwent antireflux operations were included in the study; the number of patients from each hospital is indicated in the Table. Fifty-six percent (46%–76%) (4182 patients) were neurologically normal (NN) and 44% (24%–54%) (3285 patients) were neurologically impaired (NI). The mean period of follow-up was 7.3 years.
The most common symptoms of the operated children were failure-to-thrive in 39.9% (25%–64%) and feeding disorders in 35.9% (25%–59%). Respiratory symptoms believed to be related to reflux were present in 32.6% (19%–47%). Esophagitis was present in 11.9% (8%–18%).
An esophagram with upper gastrointestinal series was the most frequently performed diagnostic study with 67.9% (56%–94%) having the study (5070 patients). This procedure was abnormal in 68.2% (44%–86%) of the patients on whom it was performed. An esophageal pH monitoring study was performed on 54.2% (14%–88%) (3300 patients). The study was abnormal in 94.5% (83%–98%) of those patients who had the procedure. Esophagoscopy was performed on 25.6% (12%–74%) (1912 patients) of patients. The procedure demonstrated esophagitis in 70.1% (16%–96%) of the patients on whom the study was performed. An isotope gastric emptying study was performed on 32.0% (5%–72%) (2389 patients). Abnormal gastric emptying with >50% retention of the isotope meal at 90 minutes was demonstrated in 50.8% (30%–65%) of those who had the study.
Of the 7467 children undergoing fundoplication, 40.4% (18%–56%) (3017 patients) were under 12 months of age at the time of operation. Only 11.9% (6%–30%) (889 patients) were age 10 years or older at operation.
The most frequently performed antireflux operation was the Nissen fundoplication that was used on 64.4% (2%–98%) (4809 patients). The Thal fundoplication was performed on 33.7% (2%–98%) (2516 patients) and the Toupet fundoplication was performed on 1.5% (0%–10%) (112 patients). Other types of antireflux operations were performed on 0.4% (0%–2%) (30 patients). Laparoscopic fundoplication was performed on 2.6% of the children.
A gastric emptying procedure (antroplasty or pyloroplasty) was performed on 481, or 11.5% (1%–26%), of NN children at the time of fundoplication and on 1311, or 39.9% (2%–76%), of NI children. A simultaneous tube gastrostomy was performed on 29.9% (3%–85%) of NN children; 89.4% (51%–99%) of NI children had a gastrostomy.
The mean hospital stay for NN children was 4.2 days (2.5 to 6 days); for NI children the mean hospital stay was 7.3 days (6 to 11.6 days).
Good to excellent results after fundoplication as noted by significant clinical improvement in preoperative symptoms were recorded in 94% (88%–97%) of NN children and in 84.6% (75%–93%) of NI children. No symptomatic improvement was recorded on 2.6% (0.4%–6.0%) of NN children or on 8.5% (4%–16%) of NI children.
Major postoperative complications occurred in 4.2% (2%–6%) of NN children and in 12.8% (4%–16%) of NI children. Postoperative death within 1 month after operation occurred in 0.07% of NN patients and in 0.8% (0.2%–1.0%) of NI children. Intestinal obstruction occurred in 2.6% (0.5%–6.2%) of patients. Respiratory complications occurred in 4.4% (3.0%–5.3%) of children. Postoperative gas bloat was recorded in 3.6% (2%–10%) of patients, and was higher in those hospitals in which a gastric emptying procedure was rarely used. Postoperative transient dumping was recorded in 0.9% (0%–5%) of patients and had no relation to whether a gastric emptying procedure had been performed. Esophageal obstruction attributable to a tight fundoplication was reported in 2.4% (1.0%–4.2%) of children. Postoperative bleeding requiring reoperation occurred in 3 patients (0.0004%). Splenic injury requiring reconstruction or resection occurred in 5 patients (0.0007%). Wound infection was recorded in 127 children (1.7%). Disruption of the fundoplication was reported in 7.1% (1.3%–18.2%) of children.
Reoperation was performed on 3.6% (2%–10%) of NN patients and on 11.8% (6%–24%) of NI children for reconstruction of the fundoplication or for adhesions.
Surgical construction of an antireflux procedure has rapidly evolved as the definitive therapy for persistently symptomatic gastroesophageal reflux disease in infants and children refractory to medical treatment. Some controversy has developed during recent years, however, regarding the timing and possible need for antireflux surgery with the availability of proton pump inhibitors (omeprazole) and prokinetic agents (cisapride), which allow optimal medical therapy.3 5 The high frequency of postoperative complications, as well as fundoplication breakdown, have been cited as indications for more extensive and lengthy use of medical therapy despite the fact that volume reflux with pulmonary aspiration as well as inadequate nutrition cannot be corrected by cytoprotective medications or by reducing gastric acid production.4 6 7Furthermore, certain medications (eg, omeprazole) are difficult to administer to infants, have side effects when given for prolonged periods, and are expensive.8
There are numerous published reports from individual hospitals that document the good results and the relatively few major complications of antireflux operations in children.2 9 10 There is, however, some difference of opinion among surgeons regarding which preoperative diagnostic procedures are most helpful and which operative technique is most efficacious, because many decisions have been made on the basis of individual or institutional experience and bias. The present report combining the clinical experience from seven large pediatric surgical centers from diverse regions in the United States is an attempt to assess the long-term effectiveness and the type and frequency of complications after antireflux surgery in children.
Many large children's surgery centers currently perform more than 100 antireflux operations each year, as in all but two of the hospitals in the present study. NI children who require a feeding gastrostomy have undergone a concomitant fundoplication with increasing frequency, and now constitute 44% of the total number of children undergoing antireflux surgery, although the frequency of complications in this high risk group of patients was four times higher than in NN patients.11-13 In each of the seven participating hospitals, extensive medical therapy administered by the gastroenterology staff was given before considering antireflux surgery.
The most frequent indications for fundoplication in the present study were failure-to-thrive, feeding disorders, and respiratory symptoms, which are similar to the indications reported by individual hospitals previously.2 Esophagitis was considered an indication for fundoplication in only 11.9% of children.
An esophagram with upper gastrointestinal series was the most frequently used diagnostic study in six of the hospitals and was performed on 67.9% of patients. This study was used as the primary diagnostic procedure for determining the presence of GER in two hospitals, although in all seven hospitals it was considered helpful in defining the anatomy of the gastroesophageal junction and ruling out congenital or acquired obstruction of the proximal small intestine.
An esophageal pH monitoring study was performed on 54.2% of patients and had the greatest diagnostic specificity of all the studies evaluated; this study was considered the definitive diagnostic procedure to determine the presence of GER in five of the hospitals.14 Esophagoscopy was performed on one-fourth of the children in the study, but was used with wide variance among the seven hospitals, depending on the recommendation of the gastroenterologist. When performed, esophagoscopy demonstrated esophagitis in 70% of patients.
An isotope gastric emptying study to determine the presence and severity of delayed gastric emptying was used with wide variance and frequency among the hospitals, particularly in NN children (2%–76%).2 Furthermore, the manner in which the study was performed differed substantially between some of the participating hospitals. When performed, the study was abnormal in approximately 50% of patients.
The criteria for selection of children for fundoplication seems to vary considerably among the participating hospitals. Those hospitals performing the largest number of fundoplications on NN children were the least likely to perform preoperative esophageal pH monitoring or esophagoscopy and to rely on history and esophagram to establish the diagnosis of reflux. NI children who required a feeding gastrostomy were less likely to undergo esophageal pH monitoring or esophagoscopy before undergoing fundoplication than were NN patients. In one hospital all NI children requiring a feeding gastrostomy received a percutaneous endoscopic gastrostomy unless severe GER was demonstrated, thus accounting for a smaller number of fundoplication in NI children than in the other hospitals.
With little variation among hospitals, the majority of antireflux operations in NN children were performed during the first 2 years of life. The average age at operation for NI children was somewhat older in each of the hospitals.
The type of antireflux operation performed varied greatly with only 2% receiving a Nissen fundoplication in one hospital, whereas >98% of patients from two other hospitals had this procedure. Conversely, the Thal fundoplication was used in 98% of patients in one hospital and in <2% of children in two other hospitals.15 The Thal procedure was used in almost all hospitals for children who had previously undergone repair of esophageal atresia with tracheoesophageal fistula who had symptomatic GER. The Toupet, Dor, Boix-Ochoa, or other antireflux procedures were used in less 2% of all patients.
Despite the wide variation in surgical technique used for the antireflux procedure among hospitals, the percentage of patients experiencing major complications, as well as the need for reoperation, was low and remarkably similar between each of the hospitals. The selection of technique seems to have been based on the previous operative experience and the bias of the surgeon. This study is thus unable to recommend one operative technique more than another based on the reported good and bad results or the complications from the participating hospitals.
A gastrostomy was used much more frequently in NI children (89.4%) compared with NN patients (29.9%) because of the need for assisted feedings. Although a concomitant gastric emptying procedure was used more frequently in NI children (39.9%) compared with NN patients (11.5%), there was wide variation in the use of this procedure among the participating hospitals in the study.16 17 Although there was a slight decrease in the number of complications in hospitals that used pyloroplasty more often, there was no statistically significant difference in results between hospitals that used a gastric emptying procedure more frequently than those that rarely used the procedure.
It is apparent from the present compiled large series that regardless of the type of fundoplication performed, the number of NN children who experienced a good result as determined by each of the contributing authors was very high (mean, 94%). The mortality was very low and comparable to that for inguinal herniorrhaphy (mean, 0.07%). Major postoperative complications occurred in only 4.2% of the NN patients. The mean hospital stay for the 4182 NN children was 4.2 days. Thus, there is little evidence to support the prolonged use of nonoperative therapy for infants and children with persistently symptomatic GER from both the standpoints of efficacy of care, as well as cost.
Reoperation for disrupted fundoplication in children with persistent symptomatic GER was recommended by the surgeons from each of the participating hospitals, although recurrent breakdown of the wrap was more than twice as common as after the initial repair.18
Similarly, for NI children, regardless of the type of fundoplication performed, the majority (mean, 84.6%) achieved a good result after operation. The complication rate in this high-risk group of children was four times higher than in the NN children, although the postoperative mortality was low (mean, 0.8%). Several reviews have evaluated the advantages of the type of fundoplication in NI children without a clear consensus, however, some authors who use the Nissen as well as the Thal procedure have recorded a higher recurrence rate in the NI children after the Thal repair.19
It is the consensus of the surgeons in the present study, particularly in the high-risk group of NI patients, that the initial fundoplication should be performed with meticulous attention to details, including construction of a loose wrap and suture approximation of the crura posterior to the esophagus. The majority favor routine division of the short gastric vessels, particularly in the NI children, although this technical detail is often omitted with the Thal repair. Particular attention was directed to correcting or improving factors that favor breakdown of the fundoplication, including poor nutrition, gastric distention, delayed gastric emptying, mechanical small bowel obstruction, chronic lung disease, and seizures with retching. Sedation and antiseizure medications during the postoperative period were helpful in reducing seizure activity and the commonly observed retching and posturing. Because many of the NI children were aerophagic, the gastrostomy was commonly used for aspiration of air as well as for feeding purposes.
Persistently symptomatic GER refractory to medical therapy is one of the two most frequent disorders for which abdominal surgery is performed in infants and children. In a seven hospital combined clinical experience in 7467 children, good to excellent long-term results were achieved with both the Nissen and Thal fundoplication in 94% of NN patients and 84.6% of NI patients. Major complications occurred in 4.2% of NN patients and were four times more frequent in NI patients. The results and complications were similar among the participating hospitals despite variations in operative technique. This study does not support the prolonged use of nonoperative therapy for children with persistently symptomatic GER from both the standpoints of efficacy of care as well as cost.
- Received July 21, 1997.
- Accepted September 9, 1997.
Reprint requests to (E.W.F.) Department of Surgery, UCLA School of Medicine, 10833 Le Conte, Los Angeles, CA 90095.
- Copyright © 1998 American Academy of Pediatrics