The recognition and treatment of childhood gastroesophageal reflux has increased during the past two decades. At the Children's Hospital of Oklahoma, no patient had an operation for gastroesophageal reflux until 1977; in the five year period from 1984 to 1988 approximately 80 patients a year underwent operative correction. Although surgeons are gratified by patients' satisfaction with operative treatment, they must assume that such patients represent a small segment of those evaluated for and found to have gastroesophageal reflux.
The symptoms of gastroesophageal reflux leading to operation are ubiquitous (Table 1), ranging from the trivial to the life-threatening. Therefore, the following questions must be answered: With what frequency is operation indicated in the concurrent care of patients with gastroesophageal reflux ? What is the natural history of existing reflux when no urgent operation is indicated or performed? Neither question can be answered comprehensively, but fragments of relevant data have been published. Among 500 infants referred for possible operation for gastroesophageal reflux, Randolph identified 83 (14%) as needing operations. l Presumably even these referred infants represent but a portion of those evaluated by their own physicians for possible gastroesophageal reflux. In the absence of lifethreatening disease, the identification of gastroesophageal reflux as a clinical problem raises the issue: What is the expected natural history of any particular child's reflux? Again no conclusive data exist. However, both Carre,2 in a retrospective analysis of patients without treatment, and Jolley, 3 in a prospective study of treated patients, suggest a 60% to 65% rate of clinical recovery from childhood gastroesophageal reflux, typically as the child assumes the erect position. Co-relative to clinical resolution, remission of gastroesophageal reflux on both barium meal and esophageal pH examination has been described. Long term follow-up of children with persistent gastroesophageal reflux is scant, but as might be expected indicates that both esophageal stricture and marked pulmonary function abnormalities are possible.2·4 The identification and indications for surgical treatment of children with significant, persisting, health impairing gastroesophageal reflux is the subject of this article.
Presenting Problems of Gastroesophageal Reflux
The diagnosis of clinically significant gastroesophageal reflux most often is occasioned by identification of a recognized consequence of vomiting or aspiration. The most common presenting problems of gastroesophageal reflux are listed in Table 1.
Whatever the preferred local diagnostic testing technique used, a child must be assessed as to the severity of illness and the likelihood of symptom remission without operation. For example, a large fixed sliding esophageal hiatal hernia is unlikely to reduce spontaneously. Determining the severity of illness and predicting the probability of symptom remission may not be easy, even with the help of parents or guardians. Nevertheless, these considerations are essential for proper patient selection and family counseling. Even without categoric advice, families typically agree with a reasoned medical/surgical approach, particularly when such discussions follow obvious adequate consideration of the child, the child's symptoms, the symptoms' severity, and the danger of the illness to the child. When specifics of medical versus surgical treatment, the length of prospective treatment, and the expected results are considered, families may be more anxious for prompt surgical treatment than are the physicians.
Summary of Diagnostic Studies for Gastroesophageal Reflux
In the absence of other obvious disease, vomiting with or without failure to thrive, respiratory symptoms including airway disease, and recurrent pneumonia or apnea are the most common symptoms suggestive of gastroesophageal reflux.5·6
Evaluation for gastroesophageal reflux ideally should provide the following information: (1) presence or absence of gastroesophageal reflux; (2) identification of reflux as the cause for the child's symptoms or complaint; (3) analysis of severity of gastroesophageal reflux as compared with normals; (4) likelihood of persistence or remission of gastroesophageal reflux; and (5) evidence that, except for gastroesophageal reflux, the remainder of the upper gastrointestinal function is normal and will remain so. Obtaining this information is complex, and no single test or study provides all the answers. However, thoughtful use of the available study modes will most often answer enough questions for a reasoned and sensible treatment approach (Table 2).7
Esophageal pH monitoring. Absolutes in identifying the presence of gastroesophageal reflux and, in turn, reflux as the cause of symptomatology do not exist. Nevertheless, 24-hour esophageal pH monitoring can be used as the gold standard, as this test measures esophageal acid over a prolonged period. Additionally, Meyers7 has shown that of any single test, esophageal pH monitoring most closely parallels the identification of patients with gastroesophageal reflux as determined by clinical evaluation and multiple testing methods. A variety of pH study methods are popular including those of DeMeester8 in adults and those of Euler,9 Sondhemier, ,0 and Jolley and Johnson3 in children. The latter three methods have data bases skewed to infants and small children and none has ideal nonoperative follow-up. Each method has its advantages and disadvantages but the method of Jolley and Johnson provides the greatest amount of useful data regarding the needs of children.3 Their technique identifies variance from the normal amount of reflux (pH score); likelihood for childhood remission of reflux (reflux type); and probability that respiratory symptoms are related to gastroesophageal reflux (mean duration of sleep reflux). Although not perfect, this technique gives reasonable results permitting a careful evaluation of infants and small children.
Upper gastrointestinal series (barium meal). Historically, upper gastrointestinal series was the preferred means for diagnosing gastroesophageal reflux.11 The presence of a hiatal hernia or gastroesophageal reflux could be documented radiographically by means of an upper gastrointestinal series (barium meal examination). Additionally, an inexact quantification of risk from existent gastroesophageal reflux could be made from the density and height of the refluxed esophageal barium column. Authors of seminal articles in the understanding of hiatal hernia and gastroesophageal reflux depended on this method.12·13 Although supplanted by esophageal pH monitoring as a single indicator of reflux and its risk, barium meal is often chosen as a screening examination. When so used, it should be extended to assess the anatomy of the pylorus and duodenum as well as the upper gastrointestinal tract to exclude such obstructive lesions as pyloric or duodenal obstruction and malrotation. As a screening examination, upper gastrointestinal series can also provide a gross estimate of gastric emptying. Typically, a barium meal examination will be used to identify esophageal stricture, nasopharyngeal dysfunction with aspiration, and other motility problems leading to aspiration with swallowing.
Radio nuclide scan. Radio nuclide scanning of the esophagus, stomach, and small bowel has been used to diagnose gastroesophageal reflux. The advantages of this method over a barium meal include: less radiation, a longer period of observation, an estimate of pulmonary aspiration, and a measurement of gastric emptying. Nuclide scanning is particularly useful in assessment of gastric emptying but has been replaced in most applications by pH monitoring for diagnosis. Gastroesophageal reflux often exists with apparent or real delay in gastric emptying. Using the method of Jolley an accurate sense of the adequacy of gastric emptying can be established preoperatively, enabling a rational consideration of the role of pyloroplasty in patients requiring antireflux surgery. M
Esophagoscopy and biopsy. In children endoscopy plays little role in the diagnosis of gastroesophageal reflux because few symptomatic children have visible signs of esophagi t is on study. However, Barrett's esophagus, uncommon at present, is recognized increasingly in chronic gastroesophageal reflux.15 Therefore, esophagoscopy with biopsy will play a concomitantly greater role in determining this structural abnormality preoperatively and in following it postoperatively. Follow-up is crucial because the importance of childhood Barrett's is uncertain.
Manometry. Except to rule out other disease processes such as esophageal dysfunction and achalasia, for which it is essential, manometry has seldom been used in childhood gastroesophageal reflux. The information manometry provides can be gathered from other studies.
The principle for treatment of gastroesophageal reflux in all age groups is reflux control and symptom prevention. Application of this principle differs for children as compared with adults. For children, nonoperative treatment must be based on a reasonable expectation of symptom ablation and resolution of gastroesophageal reflux with growth and development. Medical treatment should not be instituted when existing symptoms or signs, such as esophageal stricture or near death from gastroesophageal reflux-related aspiration, indicate that operative correction is required. Medical treatment should be time-limited by either success or failure in eliminating gastroesophageal reflux defined by symptoms or signs. For example, in a child with vomiting and failure to thrive but no esophagitis or pulmonary disease, elimination of the symptom (vomiting) and resumption of growth indicate success. On the other hand, in a child with gastroesophageal reflux, esophagitis, and gastrointestinal bleeding but no other symptomatology, documented healing of the esophagus is essential to define success.16 With these caveats, initial treatment of gastroesophageal reflux in infants and children commonly can be nonoperative, with operation reserved for those unresponsive to medical management.
Current medical management is nonspecific. Its success relies on a comprehensive workup for gastroesophageal reflux to quantify reflux and rule out medically untreatable conditions such as esophageal stricture or anatomic abnormality. Because most pediatric patients with gastroesophageal reflux are infants, positional treatment with thickened feedings, although controversial and experimentally nonverifiable, affords the most significant decrease in symptomatology. Feedings in the prone position - the head at 30° above horizontal - with mush-like formula is the best treatment available clinically. Pharmacologic increase of lower esophageal sphincter (LES) pressure is even more controversial and less clearly beneficial. Nonetheless, the use of bethanechol or metoclopramide provides physicians with a sense of providing maximal treatment. Conversely, drugs that lower LES pressure, such as caffeine and theophylline, should be avoided during treatment for gastroesophageal reflux if at all possible. I7 Esophagitis, when present, is a major hazard with gastroesophageal reflux but is uncommon in infants and children. Therefore H2 blockers seldom have a clear-cut indication in this age group.
In all surgery for infants and children, it is not sufficient that an operation work acutely but that it "stand the test of time" as measured by the longest history and follow-up possible. Antireflux surgery is major surgery and an accurate estimate of functional outcome is essential to proper patient evaluation and care. Accepting that certain patients (eg, those with esophageal stricture, reflux after repair of esophageal atresia, medical treatment failure) need operative relief from gastroesophageal reflux, the choice of operative procedure should be determined by an evaluation of the success rate (immediate and long term) and of mortality and morbidity or dysfunction.
With neurologically intact children who can take food by mouth, two operations are commonly used for the surgical treatment of gastroesophageal reflux in the United States: the Nissen fundoplication as described by Nissen18 and Bettex and Kuffer,19 and the Ashcraft modification of the Thai fundoplication.20 Either operation works well. Characteristics of each are noted in Table 3. Compared with the Thai- Ashcraft fundoplication, the Nissen procedure has a longer history of documented success at follow-up and a higher rate of reflux ablation (99% vs 95%). The Nissen is a more difficult procedure and has a higher rate of postoperative small bowel obstruction and gas bloat syndrome. With either operation mortality is low; 1 death in 1,000 consecutive Thai fundoplications reported by Ashcraft (K. W. Ashcraft, MD, personal communication, 1988) and no deaths in 350 consecutive Nissen fundoplications from the University of Oklahoma College of Medicine (WP. Tunell, MD, unpublished data, 1988).
Characteristics of Fundoplication
The results of surgical treatment are satisfactory; gastroesophageal reflux is ended in over 95% of patients with little or no mortality. Historical morbidity consists principally of repair breakdown in the Thal-Ashcraft fundoplication with return of reflux, and of small bowel obstruction or paraesophageal hernia in the Nissen fundoplication.21"23 In each of these procedures, morbidity has been reduced markedly by technical improvements in operative techniques. The gas bloat syndrome occurring following Nissen fundoplication can be identified preoperatively by measuring gastric emptying, and a pyloroplasty can be done if necessary for significant delay in gastric exit. 14
One additional consideration is germane to the neurologically impaired child. Gastrostomy feedings are associated with a higheT rate of gastroesophageal reflux with protective Thai fundoplication (20%) than with Nissen fundoplication (3%) (W. P. Tunell, unpublished data, 1988). If gastrostomy is required, this rate of postoperative reflux is a legitimate parameter in operative choice.
NEUROLOGICALLY IMPAIRED CHILDREN
The neurologically impaired child with gastroesophageal reflux presents a different set of considerations and problems. There is general agreement that gastroesophageal reflux is extremely difficult to treat medically in the neurologically impaired child.24 In addition, gastrostomy, although often desirable and required for feeding in these children, is associated with an increase in reflux and its complications.25·26 This is true whether the gastrostomy tube is placed operati vely or endoscopically.27 Nevertheless, gastrostomy feedings make care of these children so much easier that this procedure is often requested. Seventy to ninety percent of neurologically impaired children fed by gastrostomy in the absence of an antireflux procedure can be expected to have gastroesophageal reflux.28 Currently, such reflux must be considered as an "all or none" phenomenon. However, Ross has begun to quantify the likelihood of significant gastrostomy induced reflux in these children, using fourchannel esophageal monitoring.
A totally nonrefluxing fundoplication, such as the Nissen procedure, is preferred in this patient group. Effective gastric emptying, as determined by the method of Jolley, will identify those children at risk for delayed gastric emptying.14 In these children, a pyloroplasty may be indicated. On occasion neurologically impaired children with gastrostomy only may present postoperatively with pulmonary esophageal complications of gastroesophageal reflux. For these children fundoplication is indicated if a secure diagnosis is made and operative treatment is necessary.
Gastroesophageal reflux is a frequent occurrence in infancy and childhood. When appropriate symptoms are present, accurate diagnosis and treatment assessment can be obtained by a variety of diagnostic studies, most accurately by esophageal pH monitoring.
Medical, nonoperative treatment usually is indicated initially if no established complication or lifethreatening symptoms exist. When medical treatment is insufficient, operative treatment with fundoplication can be performed with an acceptable complication rate and a high expectation of success.
1. Randolph JG: Experience with the Nissen fundoplication for correction of gastroesophageal reflux in infants. Am Surg 1983; 198:579-584
2. Carre IJ: The natural history of the partial thoracic stomach (hiatus hernia) in children. Ardi Ois CMd 1959; 34:344-353.
3. Jolley SG, Johnson DG. Herbst JJ. et al: The significance of gastroesophageal reflux patterns in children- J ftdiarr Surg 1981; 16:859-865.
4. Allen JL, Wohl MEB: Pulmonary function in older children and young adults with gastroesophageal reflux. Curi Pediatr 1986; 25:541-546.
5. Tunell WP, Smith EI, Carson JA: Gastroesophageal reflux in childhood. Ann Surg 1983; 197:560-565.
6. Johnson DG, Jolley SG: Gastroesophageal reflux in infants and children. Surg CIm North Am 1981; 61:1101-1115.
7. Meyers WF, Roberts CC, Johnson DG, et al: Value of tests for evaluation of gastroesophageal reflux in children. J Pediatr Surg 1985; 20:515-520.
8. DeMeesterTR, Wang Cl, Wemly JA, et al: Technique, indications, and clinical use of 24 hour esophageal pH monitoring. ) Thorac Cardiowisc Surg 1980; 79:656-670.
9. Euler AR. Byrne WJ: Twenty-four hour esophageal intraluminal pH probe testing: A comparative analysis. Gastroenterology 1981; 80:957-961.
1 0. Sondheimer JM: Continuous monitoring of distal esophageal pH: A diagnostic test for gastroesophageal reflux in infents. J Pediatr 1980; 96:804-807.
11. NeuhauserEBD, Berenberg W: Cardio-esophageal relaxation as a cause of vomiting in infants. Radiology 1947; 48:480-483.
12. Fillet RE. Randolph JG, Gross RE: Esophageal hiatus hernia in infants and children. J Thorac Cordiouuc Surg 1964: 47:551-565.
13. Lilly JR. Randolph jG: Hiatal hernia and gastroesophageal reflux in infants and children. ) Thorac Cardiotasc Surg 1968; 55:42-54.
14. Jolley SG, Tunell WP, Leonard JC, et al: Gastric emptying in children with gastroesophageal emptying: II. The relationship to retching symptoms following anti reflex surg. J Pediatr Surg 1987; 22:927-930.
15. Hassall E, Weinstein WM, Ament ME: Barrett's esophagus in childhood. Gastroenterology 1985;89:1331-1337.
16. Ryan P, Lander M, Ong TH, et al: When does reflux oesophagitis occur with gastrooesophageal reflux in infants? A clinical and endoscopic study, and correlation with outcome. AuM Ffedlatr J 1983; 19:90-93.
17. VandenplasY, DeWoIfD, Sacre L: Influence of xanthines on gastroesophagel reflux in infants at risk for sudden infant death syndrome, ftdiarrics 1986; 77:807-810.
18. Nissen R: Gastropexy and "fundoplication" in surgical treatment of hiatal hernia. American Journal of Digesave Diseases 1961; 6:954-961.
19. Bettex M, Kuffer F: Long term results of fundoplication in hiatal hernia and cardioesophageal chalazia in infants and children. ] Pediatr Surg 1969; 4:526-530.
20. Ashcraft KW, Holder TM, Arnoury RA: Treatment of gastroesophageal reflux in children by Thai fundoplication. J Tliorac Cardiofasc Surg 1981; 82:706-712.
21. Festen C: Post-operative small bowel obstruction in infants and children. Ann Surg 1982; 196:580-583.
22. Jolley SG, Tunell WP, Hoelzcr DJ, et al: Fbst-operative small bowel obstruction m infants and children: A problem following Nissen fundoplication. J Pediatr Surg 1986; 21:407-409.
23. Festen C: Paraesophageal hernia: A major complication of Nissen's fundoplication. ) Pediatr Surg 1981; 16:496-499.
24. Wilkinson JD, Dudgeon DL. Sondheimer JM: A comparison of medical and surgical treatment of gastroesophageal reflux in severely retarded children. J Pediatr 1981; 99:202-205.
25. Wesley JR, Coran AG, Sarahan TM, et al: The need for evaluation of gastroesophageal reflux in brain damaged children referred for feeding gastrostomy. / ftdiotf Sun; 1981; 16:866-871.
26. Sondheimer JM, Morris BA: Gastroesophageal reflux among severely retarded children. J Pediatr 1979; 94:710-714.
27. Grunow JE, Al Hafidh A, Tunell WP: Gastroesophageal reflux following percutaneous endoscopic gastrostomy in children. ] Pediatr Surg 1989; 24:42-45.
28. Jolley SG. Smith El, Tunell WP: Protective anti reflux operation with feeding gastrostomy. Ann Surg 1985; 201:736-740.
Presenting Problems of Gastroesophageal Reflux
Summary of Diagnostic Studies for Gastroesophageal Reflux
Characteristics of Fundoplication