Secondary Logo

Journal Logo

Original Articles: Gastroenterology

Impact of Pediatric Rome III Criteria of Functional Dyspepsia on the Diagnostic Yield of Upper Endoscopy and Predictors for a Positive Endoscopic Finding

Tam, Yuk Him*; Chan, Kin Wai*; To, Ka Fai; Cheung, Sing Tak*; Mou, Jennifer Wai Cheung*; Pang, Kristine Kit Yi*; Wong, Yuen Shan*; Sihoe, Jennifer Dart Yin*; Lee, Kim Hung*

Author Information
Journal of Pediatric Gastroenterology and Nutrition: April 2011 - Volume 52 - Issue 4 - p 387-391
doi: 10.1097/MPG.0b013e31820e2026
  • Free


Much has been changed in the understanding of functional gastrointestinal disorders (FGIDs) in children since the establishment of the first pediatric Rome II criteria in 1999 (1). The term recurrent abdominal pain, which has been used for several decades to describe a common condition in pediatric practice, should now be replaced by abdominal pain-related FGIDs, which include functional abdominal pain, irritable bowel syndrome, abdominal migraine, and functional dyspepsia (FD) (2). FD is reported to have prevalence varying from 3.5% to 27% in children (2). The evolution of Rome III criteria has further enhanced the perception of FGIDs as a distinct disease entity that can be positively diagnosed by symptom-based approach without the prerequisite to perform extensive exclusionary investigations (3). One of the major changes in Rome III criteria in the diagnosis of FD is the elimination of mandatory use of upper endoscopy because of the low diagnostic yield in children. Instead, a symptom-based diagnostic approach has been recommended. In the absence of alarm features with normal physical findings, FD can be diagnosed and empirical treatment initiated without investigations (2,3). We conducted a prospective study in dyspeptic children to investigate the effect on the diagnostic yield of upper endoscopy with the adoption of the new recommendations and to look for any factors that could predict a positive endoscopic finding suggestive of organic disease.



The prospective study was conducted during a period of 30 months on 80 consecutive Chinese children ages 7 to 16, who were referred to the outpatient clinic of a tertiary center for investigation of epigastric pain/discomfort. The center was the only one in the region providing pediatric endoscopy serving a population of around 1.3 million. Each patient was individually assessed by the senior author (Y.H.T.) and children eligible for the study were those who symptomatically fulfilled the Rome III criteria for FD with normal physical findings. Recruited children were divided into 2 groups: a group without any alarm features and the senior author would feel comfortable to make a diagnosis of FD based on Rome III criteria without further investigations, and the other group having alarm features and should undergo investigations including upper endoscopy before arriving at a diagnosis of FD. The study sample size was estimated by having the type I error of 0.05 and power of 80% with the assumptions that dyspeptic children without alarm symptoms should have a <5% prevalence rate of organic diseases diagnosed by upper endoscopy in comparison with >30% in those experiencing alarm symptoms. These assumptions were based on our previous reports that pediatric inflammatory bowel disease is rare in our community, and peptic ulcer diseases (PUD) presenting with chronic epigastric pain occurred in <1% of all of our Chinese children who underwent upper endoscopy, whereas more than two-thirds of our children with PUD presented with gastrointestinal bleeding (4,5). We also assumed that the prevalence of alarm symptoms was around 10% of uninvestigated dyspeptic children (6).

In the present study, the alarm features relevant to dyspepsia included gastrointestinal blood loss (by history, physical examination, and fecal occult blood), dysphagia, persistent vomiting, persistent right upper quadrant pain, nocturnal pain, family history of PUD, and involuntary weight loss (2,3,7). Instead of a yes-or-no answer, we defined in the present study that nocturnal pain was epigastric pain/discomfort that awakened the child from sleep and happened in not less than 25% of the attacks. This was based on the recommendations of the pediatric Rome committee that accompanying symptoms of other FGIDs should be present in not less than 25% of the time (3). Family history of PUD was limited to parents or siblings in the present study. Exclusion criteria included history of upper abdominal surgery, concurrent significant medical or psychiatric illnesses, chronic use of nonsteroidal anti-inflammatory drugs and steroids, predominant heartburn symptoms suggestive of gastroesophageal reflux, history of upper endoscopy for investigation of dyspeptic symptoms, and history of investigations for Helicobacter pylori infection and receiving eradication therapy. Informed consent to the study was obtained from the parents.

Upper Endoscopy, Histological Examination, and Diagnosis of H pylori Infection

All of the study subjects underwent upper endoscopies regardless of the presence or absence of alarm features. The procedure was performed under intravenous sedation by a single endoscopist (Y.H.T.). None of the children received H2 blocker, proton pump inhibitor (PPI), or antibiotics within 4 weeks of the endoscopic examination. Biopsies were routinely taken at antrum for histology and rapid urease test (RUT). All of the gastric biopsy specimens were examined by a single pathologist (K.F.T.) who was blinded to the endoscopic findings. Gastritis was graded according to the Sidney classification, and H pylori infection was defined by the presence of H pylori gastritis in antral biopsy and a positive RUT result. Positive endoscopic findings suggestive of organic diseases included peptic ulcer disease, gastritis or duodenitis with erosions, eosinophilic gastroenteropathy, esophagitis, and Crohn disease.

Statistical Analysis

SPSS version 15 (SPSS Inc, Chicago, IL) was used to analyze the data. Dichotomous and continuous variables were analyzed by χ2 and Student t test, respectively. Statistical significance was defined by a P value of <0.05. Logistic regression was used to identify independent predictors for positive endoscopic finding in multivariate analysis. The study protocol was approved by the joint ethical committee of the university and the hospital.


Of the 80 consecutive patients recruited into the study, there were 23 boys and 57 girls (male-to-female ratio 1: 2.5), with mean age 13.1 ± 2.2 years (range 7–16). Mean symptoms duration was 17.5 ± 12 months (range 3–60). There was no difference between the boys and the girls in the mean age, symptom duration, and frequency of attack. Only 9 of the 80 patients (11.3%) had experienced ≥1 of the alarm features. Nocturnal pain and family history of PUD were each reported in 5 children, with 1 child having both of them. None of the children presented with other alarm features.

Five of 80 children (6.3%) had a positive endoscopic finding that suggested organic pathology accounting for the dyspeptic symptoms (the positive group). The remaining 75 patients were considered to be negative and confirmed to suffer from FD because they either had normal endoscopic and histological findings or had nonspecific endoscopic findings (eg, mild erythema) or descriptive histological changes (eg, mild chronic inflammatory cells infiltrate, reactive gastritis). Overall 8 children (10%) were infected with H pylori. There were 5 boys and 3 girls (26.1% vs 3.5%; P < 0.01). Four of the 5 children in the positive group were infected with H pylori (Table 1).

Clinical characteristics and endoscopic findings of study subjects

Among the 9 patients who had alarm features, 3 (33.3%) were found to have organic pathology, compared with only 2 children (2.8%, P < 0.01) of 71 without having any alarm features. The positive and the negative predictive values of the presence of alarm features were 0.33 (95% confidence interval [CI] 0.03–0.64) and 0.97 (95% CI 0.93–1.01), respectively. There was no difference between those with and those without alarm features in the sex distribution, mean age, mean duration of symptoms, and frequency of attacks. The pathological conditions in the positive group included duodenal ulcer (n = 2), duodenitis with erosions (n = 2), and generalized gastritis with gastric erosions (n = 1) (Table 2).

Characteristics of the 5 children with organic diseases

Comparing the 5 children in the positive group with the remaining 75 having negative endoscopic findings and suffering from FD, the positive group had male predominance (80% vs 25.3%, P < 0.01), higher prevalence of the presence of alarm features (60% vs 8%, P < 0.01), and higher prevalence of H pylori infection (80% vs 5.3%, P < 0.01). When nocturnal pain and family history of PUD were studied separately, only nocturnal pain was found to be statistically more prevalent in the positive group (60% vs 2.7%, P < 0.01). Antral biopsies showed more mononuclear cell and neutrophil infiltrate in the endoscopically positive group (P < 0.01). There was no difference between the 2 groups in the mean age, symptom duration, and frequency of attacks (Table 3). Multivariate analysis identified H pylori infection (odds ratio [OR] 23.2; 95% CI 1.5–333) and presence of nocturnal pain (OR 26.3; 95% CI 1.2–500) to be the independent predictors of positive endoscopic evidence of organic pathology to account for the dyspeptic symptoms.

Comparison between children with organic diseases and the others with FD after upper endoscopy

The 4 patients infected with H pylori in the positive group were treated with a 7-day course of PPI-based triple eradication therapy using a combination of amoxicillin and clarithromycin. PPI was continued for 3 more weeks in the 2 children with duodenal ulcers. H2 blocker was given for 4 weeks to treat the patient with H pylori–negative duodenitis with erosion. All 5 patients in the positive group had complete resolution of symptoms after treatment. The remaining 75 children with FD confirmed by a normal upper endoscopy showed no evidence of organic disease at a mean follow-up of 13.2 months.


The overall 6.3% of dyspeptic children having organic pathology diagnosed by upper endoscopy was in agreement with previous reports that the diagnostic yield of upper endoscopy in dyspeptic children was low and the vast majority of them were having FD (8–10). We found peptic ulcer disease and erosive duodenogastritis to be the only organic diseases that accounted for the dyspeptic symptoms in this prospective series. However, because random biopsies were not taken at the macroscopically unremarkable mucosa of esophagus, duodenum, and gastric body, we may have missed some organic pathology such as nonerosive microscopic reflux esophagitis, eosinophilic esophagitis or gastroduodenitis, and celiac disease. Little is known about eosinophilic gastroenteritis and celiac disease in Chinese children. Gastroesophageal reflux was found to occur in a significant proportion of adult patients fulfilling Rome criteria for FD with normal macroscopic findings at upper endoscopy (11). We believe that the presence of microscopic reflux esophagitis in the absence of typical reflux symptoms should not necessarily preclude the diagnosis of FD in children and more evidence is required from further studies that investigate this issue.

The American Academy of Pediatrics and the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition have previously stated in their technical report that there is little evidence to support the use of endoscopy and biopsy in the absence of alarm symptoms to give a significant yield of organic disease (12). The recent evolution of Rome III criteria has addressed the same issue by eliminating the mandatory use of upper endoscopy to make a diagnosis of FD. To the best of our knowledge, this is the first prospective study with the intention to validate the effectiveness of using the alarm features to discriminate organic diseases from FD. Two studies have recently reported that the presence of alarm symptoms was not a good predictor for positive diagnostic yield in upper endoscopy performed in children with abdominal pain (13,14). Miele et al (13) found that only deceleration of linear growth was associated with increased diagnostic yield, whereas other alarm symptoms were significantly associated with negative endoscopic findings. Thakkar et al (14) reported that the majority of alarm symptoms except persistent vomiting were not predictive of positive endoscopic findings. Both studies, however, were retrospective in nature and included a heterogeneous group of children who had undergone upper endoscopy for investigation of nondefined chronic abdominal pain, in contrast to our prospective study on only those children with dyspepsia well defined by Rome III criteria. The alarm symptoms analyzed in previous studies did not include family history of PUD and nocturnal pain, and were not defined to be specific for upper gastrointestinal tract because upper and lower gastrointestinal blood loss and persistent right upper and lower quadrant pain were grouped together for analysis (13,14).

We found only 11.3% of our study subjects had alarm features that included nocturnal pain and family history of PUD. Few studies have reported on the prevalence of alarm features among children with functional abdominal pain. A recent study (6) conducted in Norway found that 7% of their children referred for evaluation of abdominal pain had alarm symptoms. The fact that we did not encounter other alarm features may be related to the design of our study in which we recruited the study subjects from only those new patients who were primarily referred for investigation of epigastric pain/discomfort. Following the recent recommendations that FD can be comfortably diagnosed using the symptom-based diagnostic approach in dyspeptic children who have normal physical findings in the absence of alarm symptoms (2,3,7,15), the number of upper endoscopies required in our patients would have been reduced by almost 90% after the first consultation. We have demonstrated in the present study that the diagnostic yield of upper endoscopy was significantly increased to 33.3% in children with alarm features, in contrast to the overall figure of 6.3% and 2.8% in those without. Although only 2 alarm symptoms were encountered in this series, our results are still able to substantiate the Rome III guidelines that screening dyspeptic children for alarm symptoms can effectively reduce the number of fruitless upper endoscopy and increase its diagnostic yield.

Although nocturnal pain has been considered to be one of the alarm symptoms, recent literature reviews do not find enough evidence to support its role to distinguish organic from functional abdominal pain (12). There are no clear definitions on the frequency of nocturnal pain in international guidelines. We have arbitrarily defined nocturnal pain in the present study that the alarm symptom was only considered to be present if it happened in not less than 25% of the attacks. By adopting such a tighter definition, only 6.3% of our children reported this symptom, in contrast to 45% to 70% reported by others (8,16). Our results have shown that using such a definition, nocturnal pain was an independent predictor for organic pathology. A too-loose definition would inevitably include too many children to require upper endoscopy because of the presence of alarm symptoms while majority of these children are just having FD.

Two of the 71 children (2.8%) who did not report any alarm symptoms were found to have organic pathology by upper endoscopy. Others have reported that 8 of 152 children (5.3%) were found to have organic diseases despite the absence of any alarm symptoms and would have otherwise fulfilled the criteria for various abdominal pain-related FGIDs based on symptoms and physical examinations (6). Our results have demonstrated that male sex and the status of H pylori infection were associated with organic diseases, in addition to the alarm features. Our previous epidemiological study reported that the prevalence of H pylori infection among our asymptomatic children ages 6 to 19 years was 13.1% (17). The overall infection rate of 10% among the dyspeptic children in the present study appeared to support the general belief that H pylori is not associated with abdominal pain in children because of a similar prevalence in symptomatic and asymptomatic children (18). Nevertheless, we found in the present study that 80% of the children with organic pathology were H pylori infected, in comparison to only 5.3% in those with FD. Our findings suggest that H pylori is associated with the dyspeptic symptoms in a small proportion of dyspeptic children who have H pylori–associated organic diseases instead of FD, and this cannot be reflected simply by comparing the prevalence between symptomatic and asymptomatic children.

The Rome III subcommittee recommends upper endoscopy in children who need a diagnosis of H pylori–associated disease to be confirmed but does not provide guidelines on how to identify these potential children (3). It would be worth considering investigation in dyspeptic children for H pylori by noninvasive methods such as urea breath test, not intending to adopt a test-and-treat strategy, but to supplement alarm symptoms to identify those who are at higher risk for having organic diseases and should undergo upper endoscopy before making a diagnosis of FD. We believe that it may even be more cost-effective to perform screening tests of H pylori on male patients, who constitute only a minority among dyspeptic children but have a higher association with organic diseases and higher prevalence of H pylori, as demonstrated in our results. In contrast, H pylori infection has no sex preponderance in asymptomatic children as reported by us and others (17,19).

We have demonstrated in the present study that the Rome III recommendations to screen dyspeptic children for alarm features and investigation for H pylori infection are effective to discriminate organic diseases from FD, and can significantly increase the diagnostic yield of upper endoscopy by reducing the number of fruitless endoscopic procedures. A recent study reporting the opinions of a panel of experts on different hypothetical scenarios found that upper endoscopy was perceived as an appropriate investigation in dyspeptic children with a family history of PUD and/or H pylori infection, older than 10 years of age, with symptoms persisting for more than 6 months and severe enough to affect activities of daily living, including sleep (20). The expert opinions appear to be in agreement with our results. Although limited by the small case number and the geographical variations in the infection rates of H pylori in children, our findings may still be applicable to most of the tertiary referral settings in developed countries where, in general, the prevalence rates of H pylori in children are low.


1. Rasquin-Weber A, Hyman PE, et al. Childhood functional gastrointestinal disorders. Gut 1999; 45(Suppl 2):1160–1168.
2. Yacob D, Di Lorenzo C. Functional abdominal pain: all roads lead to Rome (criteria). Pediatr Ann 2009; 38:253–258.
3. Rasquin A, Di Lorenzo C, Forbes D, et al. Childhood functional gastrointestinal disorders: child/adolescent. Gastroenterology 2006; 130:1527–1537.
4. Tam YH, Lee KH, Chan KW, et al. Colonoscopy in Hong Kong Chinese children. World J Gastroenterol 2010; 16:1119–1122.
5. Tam YH, Lee KH, To KF, et al. Helicobacter pylori-positive versus Helicobacter pylori-negative idiopathic peptic ulcers in children with their long-term outcomes. J Pediatr Gastroenterol Nutr 2009; 48:299–305.
6. Helgeland H, Flagstad G, Grotta J, et al. Diagnosing pediatric functional abdominal pain in children (4–15 years old) according to the Rome III criteria: results from a Norwegian prospective study. J Pediatr Gastroenterol Nutr 2009; 49:309–315.
7. Noe JD, Li BU. Navigating recurrent abdominal pain through clinical clues, red flags, and initial testing. Pediatr Ann 2009; 38:259–266.
8. Hyams JS, Davis P, Sylvester FA, et al. Dyspepsia in children and adolescents: a prospective study. J Pediatr Gastroenterol Nutr 2000; 30:413–418.
9. Spiroglou K, Chatziparasidis G, Paroutoglou G, et al. Functional dyspepsia in children. J Pediatr Gastroenterol Nutr 2001; 33:519.
10. Schurman JV, Singh M, Singh V, et al. Symptoms and subtypes in pediatric functional dyspepsia: relation to mucosal inflammation and psychological functioning. J Pediatr Gastroenterol Nutr 2010; 51:298–303.
11. Xiao YL, Peng S, Tao J, et al. Prevalence and symptom pattern of pathologic esophageal acid reflux in patients with functional dyspepsia based on the Rome III criteria. Am J Gastroenterol 2010; 105:2626–2631.
12. Di Lorenzo C, Colletti RB, Lehmann HP, et al. Chronic abdominal pain in children: a technical report of the AAP and the NASPGHAN. J Pediatr Gastroenterol Nutr 2005; 40:249–261.
13. Miele E, Giannetti E, Martinelli M, et al. Impact of the Rome II paediatric criteria on the appropriateness of the upper and lower gastrointestinal endoscopy in children. Aliment Pharmacol Ther 2010; 32:582–590.
14. Thakkar K, Chen L, Tatevian N, et al. Diagnostic yield of oesophagogastroduodenoscopy in children with abdominal pain. Aliment Pharmacol Ther 2009; 30:662–669.
15. Di Lorenzo C, Colletti RB, Lehmann HP, et al. Chronic abdominal pain in children: a clinical report of the AAP and the NASPGHAN. J Pediatr Gastroenterol Nutr 2005; 40:245–248.
16. Machado RS, Reber M, Patricio FR, et al. Gastric emptying of solids is lower in functional dyspepsia unrelated to Helicobacter pylori infection in female children and teenagers. J Pediatr Gastroenterol Nutr 2008; 46:403–408.
17. Tam YH, Yeung CK, Lee KH, et al. A population-based study of Helicobacter pylori infection in Chinese children resident in Hong Kong: prevalence and potential risk factors. Helicobacter 2008; 13:219–224.
18. Jones NJ, Sherman P, Fallone CA, et al. Canadian Helicobacter study group consensus conference: update on the approach to Helicobacter pylori infection in children and adolescents—an evidence-based evaluation. Can J Gastroenterol 2005; 19:399–408.
19. Brown LM. Helicobacter pylori: epidemiology and routes of transmission. Epidemiol Rev 2000; 22:283–297.
20. Guariso G, Meneghel A, Visona Dalla Pozza L, et al. Indications to upper gastrointestinal endoscopy in children with dyspepsia. J Pediatr Gastroenterol Nutr 2010; 50:493–499.

children; functional dyspepsia; Helicobacter pylori; Rome III criteria; upper endoscopy

Copyright 2011 by ESPGHAN and NASPGHAN