Symptoms suggestive of gastroesophageal reflux disease (GERD) are relatively common, occurring in 2% to 7% of children (1) (2–4) (5,6) (7,8) (9,10) (11,12) (13)
The prevalence of symptoms of pediatric GERD has been documented. In a recent cross-sectional survey of 16 pediatric group practices in the Chicago area, Nelson et al (1) (1)
We recently carried out a single-center retrospective cross-sectional study using chart review (N = 402) to examine the prevalence of endoscopic manifestations of GERD in neurologically normal children without esophageal congenital abnormalities (11)
Although the prevalence symptoms are well studied, the more severe manifestations of GERD in children are not known. There is a relative scarcity of information on the prevalence of endoscopic findings of GERD in children. The purpose of this study was to examine the prevalence and determinants of erosive esophagitis in children undergoing upper endoscopy.
PATIENTS AND METHODS
This was a retrospective cross-sectional study using information collected in the Pediatric Endoscopy Database System-Clinical Outcomes Research Initiative (PEDS-CORI) (14) pediatric endoscopy procedures and to promote scientific inquiry into the need, efficacy, costs, and outcomes of pediatric endoscopy . PEDS-CORI has 12 children's hospital sites across the United States, with more than 100 pediatric gastroenterologists (Fig. 1 ). The data are contained in a MySQL database, which is a commercial program that allows data storage similar to Microsoft Access.
FIG. 1: The 12 active PEDS-CORI sites.
Study Population
All of the patients who underwent procedures that were recorded in PEDS-CORI during the time period from 1999 to 2002 were eligible to be included in this study. Inclusion criteria were upper endoscopic procedures between 1999 and 2002 and age at the time of diagnosis between 0 months and 17 years, 11 months. Exclusion criteria were incomplete reports and reports that omitted demographic features, indication for endoscopy, or endoscopic findings.
Data Collection
For all of the patients who satisfied the inclusion and exclusion criteria, the following information was accrued: demographic features (age, sex, ethnicity, and geographic location of the endoscopy center) and the listed indication for the upper endoscopy procedure. The PEDS-CORI database contains at least 1 indication per procedure. Listed indications include evaluation of a condition (eg, anemia, iron deficiency anemia, graft-versus-host disease, feeding refusal, Crohn disease), evaluation of a suspected condition (eg, varices, upper gastrointestinal bleed, malabsorption, Barrett esophagus, inflammatory bowel disease), presence of symptoms (eg, weight loss, chest pain, dysphagia, nausea, anorexia, odynophagia, pulmonary symptoms, dyspepsia, abdominal pain, reflux symptoms <1 year, hematemesis, melena, hematochezia, vomiting), surveillance (Barrett esophagus, gastric polyps, varices, gastric ulcer, duodenal ulcer, Helicobacter pylori ), and screening (eg, varices, Barrett esophagus, family history of cancer). In the case of 1 person with multiple procedures, only the information recorded in the reports of the index endoscopy were considered. The data were converted from MySQL format into SAS datasets and transferred to the Houston Center for Quality of Care & Utilization Studies for evaluation. The primary analyses involved the indications listed above. We assigned several indications as surrogates for case definitions used in our primary analyses (eg, noncomplicated reflux = reflux symptoms <1 year, no dysphagia, no Barrett esophagus, no pulmonary symptoms, no anemia; complicated GERD = reflux symptoms <1 year, dysphagia, in the absence of Crohn disease or graft-versus-host disease). The presence of erosive esophagitis was indicated either by descriptive terms (eg, erosion, ulcer) or according to the Los Angeles classification (15)
0 = No breaks or healed breaks
A = 1 or more mucosal breaks on folds, <5 mm
B = 1 or more mucosal break >5 mm, noncontiguous
C = At least 1 contiguous mucosal break between 2 folds
D = 1 or more circumferential mucosal breaks
Other findings that were specifically collected include hiatal hernia, ring, stricture, postsurgical changes, and either established or suspected Barrett esophagus.
Data Analysis
The overall number of patients with any erosive esophagitis, and with each grade of erosive esophagitis (Los Angeles classification) was determined. The prevalence rates of each of these lesions were calculated for the entire group. Age-specific prevalence rates were calculated for each year from 0 to 17 years, 11 months. Similarly, sex-specific (male and female) and ethnicity-specific (white, black, Hispanic, other) prevalence rates were calculated. For each prevalence rate, the accompanying 95% confidence intervals were calculated. For every indication (eg, reflux symptoms <1 year), univariate analyses to test for differences between patients with and those without erosive esophagitis with regard to demographic characteristics, and the presence of hiatal hernia, gastritis, and gastric/duodenal ulcers was performed. Similar comparisons were performed for prespecified constellations of indications (eg, a combination of reflux symptoms <1 year, dysphagia, weight loss, and absence of Crohn disease or graft-versus-host disease to indicate severe GERD) and for each grade of erosive esophagitis. The χ2 test was used for dichotomous variables, unpaired t tests to compare continuous variables with normal distribution, and Wilcoxon tests for nonparametric variables.
Esophageal biopsy was not evaluated. The presence of erosive esophagitis or the presence of each grade of erosive esophagitis was examined separately as outcome variables in logistic regression analyses in which demographic features, geographic location, and indication or constellation of indications served as explanatory variables. Parameter estimates, adjusted risk ratios, and their accompanying 95% confidence intervals were calculated for each explanatory variable. Hierarchical models with random effects were used to adjust for clustering within referral areas. All of the models were checked for satisfying the regression assumptions and tested for interactions. The study was approved by the Investigational Review Board for Human Subjects at the Baylor College of Medicine.
RESULTS
Between December 1999 and December 2002, 7188 children who underwent upper endoscopy fulfilled the inclusion and exclusion criteria. Of those, 888 (12.4%) had erosive esophagitis. The median age of children with children with erosive esophagitis was 12.7 ± 4.9 years, compared with 10.0 ± 5.1 years in those without erosive esophagitis (P ≤ 0.0001). Of those with erosive esophagitis, 490 (55.2%) were male, compared with 3040 (48.2%) in those without erosive esophagitis (P = 0.001). The ethnicity of those children with erosive esophagitis was 77% white, 12% black, and 11% other, compared with children without erosive esophagitis, of whom 79% were white, 9% were black, and 13% other (P = 0.14) (Table 1 ). Erosive esophagitis was seen in 29 of 532 (5.5%) children ages 0 to 1 years and progressively increased to 106 of 542 (19.6%) children at age 17 (Figs. 2 and 3 ). Hiatal hernia was seen in 68 of 888 (7.7%) children with erosive esophagitis, compared with 228 of 6300 (2.5%) children without erosive esophagitis (P ≤ 0.0001) (Table 1 ). The prevalence of Barrett esophagus, esophageal stricture, ulcer, previous surgery, nodules, foreign body or retained food, and anatomic abnormalities was not significantly different between children with erosive esophagitis and those without (data not shown).
TABLE 1: Clinical characteristics of children with and without erosive esophagitis
FIG. 2: Frequency (by absolute number) of erosive esophagitis in children undergoing upper endoscopy by age group.
FIG. 3: Percentage of children with erosive esophagitis undergoing upper endoscopy by age group.
DISCUSSION
To our knowledge, this is the first large population-based study to suggest that erosive esophagitis is common in children undergoing upper endoscopy. Erosive esophagitis is slightly more common in male individuals and increases with age (Figs. 2 and 3 ). Hiatal hernia was the only endoscopic observation that predicted the presence of erosive esophagitis.
There are notable similarities and differences in the endoscopic manifestations of GERD between children and adults. For example, the prevalence of erosive esophagitis (12.4%) is significantly lower than that that reported in adults (16) (17)
We found a lower prevalence of erosive esophagitis in this study (12.4%) than that reported in our previous single-center study (34.6%) (18)
There are many important limitations of the data collected in PEDS-CORI. The accuracy of the data is assumed, inasmuch as the PEDS-CORI report serves as the medical record in clinical practice. However, clinicians are required by payers to provide an “acceptable” indication for a procedure, and in some cases they may feel pressured to provide an endoscopic diagnosis other than normal. Interestingly, this does not seem to be the case of the data currently stored in PEDS-CORI, inasmuch as “normal” was a common endoscopic impression. Another important limitation is the endoscopic lexicon used in the database. The terminology was designed for the practice of endoscopy in adults and reflects the minimum standard terminology developed in Europe (19). Individual users may resist lexicons that restrict them, making them feel limited to a particular finding or indication. PEDS-CORI provides a free-text field to allow the endoscopist to enter miscellaneous data; this field also provides a means for future revision of the endoscopic lexicon. Another potential limitation is the inability to collect data from all of the consecutive procedures at all of the sites. It is possible that some procedures are missed if the computer crashes or if procedures are performed offsite where there is no access to the database. There has been a diligent effort to obtain participation of a cross-section of children's hospitals across North America, including both academic centers and private practitioners (Fig. 1 ). Additionally, there was no comparison between the endoscopic findings and histological findings. It is possible that the endoscopic interpretation does not correlate with biopsy-proven esophagitis. The PEDS-CORI database also does not require a final diagnosis, presenting symptoms, or prescribed treatment, which would aid in data analysis. Finally, we included the Los Angeles criteria for the endoscopic diagnosis of erosive esophagitis (15)
In conclusion, erosive esophagitis, a complication of GERD, is slightly more common in male individuals and increases with age. Hiatal hernia was the only endoscopic finding that predicted erosive esophagitis. These data suggest that the duration of GERD is a risk factor for the development of erosive esophagitis.
REFERENCES
1. Nelson SP, Chen EH, Syniar GM,
et al . Prevalence of symptoms of gastroesophageal reflux during childhood: a pediatric practice-based survey. Arch Pediatr Adolesc Med 2000; 154:150–154.
2. Nebel OT, Formes MF, Castell DO. Symptomatic gastroesophageal reflux: incidence and precipitating factors. Am J Dig Dis 1976; 21:953–956.
3. Gallup Organization.
Gallup Survey on Heartburn Across America . Princeton, NJ: Gallup Organization, 1988.
4. Locke GR III, Talley NJ, Fett SL,
et al . Prevalence and clinical spectrum of gastroesophageal reflux: a population-based study in Olmstead County, Minnesota. Gastoenterology 1997; 112:1448–1456.
5. Sonnenberg A, El-Serag HB. Epidemiology of gastroesophageal disease. In: Buchler MW, Farthmann EH, editors. Progress in Surgery, Vol. 23. Basel: Karger; 1997. pp. 20–36.
6. Shepard RW, Evans JWS, Lander M,
et al . Gastroesophageal reflux disease in children. Clin Pediatr 1987; 26:55–60.
7. Marcon MA. Advances in the diagnosis and treatment of gastroesophageal reflux disease. Curr Opin Pediatr 1997; 9:490–493.
8. Hassall E. Wrap session: is the Nissen slipping? Can medical treatment replace surgery for severe gastroesophageal reflux disease. Am J Gastroenterol 1995; 90:1212–1220.
9. Krug E, Bergmeijer JHLJ, Dees J,
et al . Gastroesophageal reflux and Barrett's esophagus in adults born with esophageal atresia. Am J Gastroenterol 1999; 94:2825–2828.
10. Lindahl H, Rintala R, Sariola H. Chronic esophagitis and gastric metaplasia are frequent late complications of esophageal atresia. J Pediatr Surg 1993; 28:1178–1180.
11. El-Serag HB, Gilger MA, Kuebler M,
et al . Extraesophageal associations of gastroesophageal reflux disease in children without neurologic disease. Gastroenterology 2001; 121:1294–1299.
12. Taylor LA, Weiner T, Lacey SR,
et al . Chronic lung disease is the leading risk factor correlating with failure (wrap disruption) of antireflux procedures in children. J Pediatr Surg 1994; 29:161–166.
13. Gilger MA. Pediatric otolaryngologic manifestations of gastroesophageal disease. Curr Gastroenterol Rep 2003; 5:247–252.
14. Gilger MA, Gold BD.
Pediatric endoscopy : new information from the PEDS-CORI project. Curr Gastroenterol Rep 2005; 7:234–239.
15. Lundell LR, Dent J, Bennett JR,
et al . Endoscopic assessment of esophagitis: clinical and functional correlates and further validation of the Los Angeles Classification. Gut 1999; 45:172–180.
16. Ollyo JB, Monnier P, Fontolliet C,
et al . The natural history, prevalence and incidence of reflux esophagitis. Gullet 1993; 3(Suppl 1):3–10.
17. El-Serag HB, Bailey NR, Gilger MA,
et al . Endoscopic manifestations of gastroesophageal reflux disease in patients between 18 months and 25 years without neurological deficits. Am J Gastroenterol 2002; 97:1635–1639.
18. Delvaux M, Schubbe H, Stettin J,
et al . Report generation in digestive endoscopy using a computerized endoscopic database: results of a four year experience. Gastrointest Endosc 1995; 41:301–309.
