DOI: https://doi.org/10.21203/rs.3.rs-1435444/v1
Background: Large-scale data on esophagogastroduodenoscopy (EGD) in China are scarce. This study aimed to assess the indications and diagnostic yield of EGD in children and the relationship between factors (such as age, sex, and indications) and diagnostic yield.
Methods: We performed a prospective cross-sectional observational study involving patients aged < 18 years who underwent diagnostic EGD. The study was conducted in five children’s hospitals, each in a different city. Demographic features, indications for endoscopy, and endoscopic and histopathological findings were collected. Univariable and multivariable ordinal logistic regression analyses of the relationship between the factors and diagnostic yield were performed.
Results: The study included 2268 patients (male/female ratio, 1.3:1) with a median age of 8.68 years. Among the 2268 children, the most frequent indications were abdominal pain in 1954 (86.2%), recurrent vomiting in 706 (31.1%), weight loss in 343 (15.1%), and others. The endoscopic yield was 62.5% and was the highest in patients with dysphagia (90.9%). The histologic yield was 32.4% and was the highest in patients with unexplained anemia (54.6%). On multivariable regression analysis, the endoscopic yield was associated with dysphagia, gastrointestinal (GI) bleeding, and recurrent vomiting, and the histologic yield was associated with weight loss and age. Different groups of patients with abdominal pain had variable probabilities of abnormal endoscopic findings.
Conclusions: The most frequent indication of pediatric EGD is abdominal pain, with variable probabilities of abnormal endoscopic findings in different groups. Endoscopic yield and histologic yield are associated with certain alarming features.
Trial registration: The trial registration number (TRN): NCT03603093 (ClinicalTrials. gov), Date of registration: July 18th, 2018
Pediatric esophagogastroduodenoscopy (EGD) began in the 1970s. Parallel to the growth of pediatric gastroenterology, an increase in the use of EGD has been observed (1). Currently, EGD is a sensitive diagnostic tool with rare complications, which can be performed at any age (2).
EGD in children can either be diagnostic or therapeutic. In 1996, the North American Society of Pediatric Gastroenterology and Nutrition provided general indications for diagnostic upper gastrointestinal (GI) endoscopy in infants, children, and adolescents, which included the presence of symptoms indicative of an underlying organic pathology of the gastrointestinal tract (3). American Society for Gastrointestinal Endoscopy (ASGE) provided modified guidelines for pediatric EGD in 2000, 2008, and 2014 (4–6). In 2015, the European Society for Pediatric Gastroenterology Hepatology and Nutrition (ESPGHAN) and the European Society of Gastrointestinal Endoscopy (ESGE) developed the latest guidelines for pediatric gastrointestinal endoscopy, where the indications for diagnostic EGD were clearly stated (7).
Multiple studies have discussed the indications and diagnostic yield of EGD. Miele et al. found that nearly one-fourth of endoscopic procedures performed were inappropriate (8). Compliance with published guidelines is associated with improved diagnostic efficiency (9). A systematic review found that the diagnostic yield of appropriate endoscopies was higher than that of inappropriate ones (43.3% vs. 35.1%) (10). A recent retrospective study in children found that 47.2% of upper GI endoscopies revealed abnormal findings and that age < 60 months, abdominal pain, dysphagia/odynophagia, and heartburn were predictive of abnormal endoscopy findings (11). Another study showed that EGD was both macroscopically and histologically normal in 80.6% of cases and that unless there are alarming symptoms, younger children do not need EGD. However, studies on pediatric EGD are limited and are mostly retrospective. Based on our prior work in one center, we conducted this multicenter cross-sectional observational study to delineate the indications and diagnostic yield of EGD and depict the relationship between factors and endoscopic and histological diagnostic yields in pediatric patients who underwent their first diagnostic endoscopy.
Study design
This was a prospective, cross-sectional observational study (ClinicalTrials. gov ID: NCT03603093) in five hospitals, each in a different city in China. These centers were the Children’s Hospital of Fudan University in Shanghai, Children’s Hospital of Fudan University Xiamen Branch in Xiamen, Henan Children’s Hospital (Zhengzhou Children’s Hospital) in Zhengzhou, Wuhan Children’s Hospital in Wuhan, and Hunan Children’s Hospital in Changsha. Ethics approval was obtained from all five hospitals.
At the onset of this study, a steering committee was formed, comprising five directors of the gastroenterology department and 1–2 doctors in charge of data collection from each study center. A case report form (CRF) with inclusion and exclusion criteria, questionnaire, study definitions, and guidelines for data collection were provided to the centers. The questionnaire included the basic characteristics of the patients and indications for EGD. Professor Ying Huang and attendings in the Department of Gastroenterology in the Children’s Hospital of Fudan University designed the CRF and questionnaire based on their previous work.
Patient selection and data collection
Patients aged 0–18 years who underwent diagnostic EGD in any of the five children’s hospitals were included in our study. The indications for EGD were based on the latest guidelines by ESPGHAN and ESGE. Those who did not consent to provide information for the study were excluded from the study.
Recruitment commenced in the five selected hospitals from December 2018 to September 2019. Of the patients who underwent EGD during this period, 2268 fulfilled the inclusion criteria and gave consent to provide data on variables in the questionnaire.
Doctors responsible for data collection administered the questionnaire either to the parents of the patients or the patients themselves. Questions were about the patient’s basic characteristics, indications for EGD, and date of EGD. In addition, endoscopic and pathologic reports were combined with the questionnaires. Biopsy was performed routinely, except in those who declined biopsy.
Indications
Indications were classified as abdominal pain, recurrent vomiting, weight loss or failure to thrive, GI bleeding, unexplained anemia, symptoms of gastroesophageal reflux (GER), diarrhea, caustic ingestion, dysphagia, etc., according to the latest guidelines by ESPGHAN and ESGE. Some patients had more than one indication.
Endoscopic and histologic findings
A positive endoscopic yield was defined as the presence of relevant findings on endoscopy, grouped into categories such as esophagitis, gastritis, duodenitis, peptic ulcer, Helicobacter pylori infection diagnosed through rapid urease test, and others. The histologic findings mainly included moderate or severe inflammation or Helicobacter pylori infection confirmed by immunohistochemical staining.
Data management
The questionnaires and endoscopic and pathologic reports from the other four centers were sent back to the center in Shanghai. Data entry was performed by two people using the EpiData software. After double entry, the data were compared; non-conforming data were confirmed by referring to the original questionnaire.
Statistical analysis
All analyses were performed using Stata 12.0 statistical software (StataCorp LP, College Station, TX, USA). Descriptive statistics were calculated for demographic features (sex and age). After generating descriptive counts and proportions for symptom variables, Chi-square tests were used to compare patients with different demographic features or symptom variables for diagnostic yield. Then, univariable and multivariable logistic analyses were performed to assess the relationship between demographic features or symptom variables and the presence of a positive endoscopic or histologic abnormality. Significance was set at p<0.05.
Study center characteristics
During the study period, 2287 questionnaires were administered. Those that did not conform to the indications and those with missing data were discarded. Eventually, 2268 questionnaires were included in the analyses, comprising 914 questionnaires from Shanghai (population: 24.28 million), 500 from Wuhan (population: 11.21 million), 382 from Zhengzhou (population: 10.35 million), 337 from Xiamen (population: 4.29 million), and 135 from Changsha (population: 8.39 million). The mean age (SD) of the participants was 8.68 years (3.25), and 57% of them were male (Table 1).
Table 1. Baseline patient characteristics
Characteristics |
|
Total |
2268 |
Shanghai |
914 |
Wuhan |
500 |
Zhengzhou |
382 |
Xiamen |
337 |
Changsha |
135 |
Male/female |
1.3:1 (1294/974) |
Age (IQR) |
8.68±3.25 (0.28–17.58) |
0–1 y |
16 |
1–5 y |
292 |
5–12 y |
1575 |
Teenagers (13–18 y) |
385 |
Biopsies |
|
Yes |
2242 |
No |
26 |
Indications
Overall, the common indications for endoscopy were abdominal pain (86.2%), persistent vomiting (31.1%), weight loss (15.1%), GER symptoms (8.4%), GI bleeding (4.7%), diarrhea (3.7%), unexplained anemia (0.5%), and dysphagia (0.7%). Many patients had over one recorded indication. The indications were similar in different centers, i.e., abdominal pain was the most common indication.
Endoscopic findings and yield
Some patients had more than one endoscopic finding. The overall prevalence of abnormal endoscopic findings was 62.5%; abnormal findings were found in the esophagus (6.7%), stomach (30.2%), and duodenum (21.6%). The rapid urease test was positive in 32.8% of the 2097 patients. Table 2 shows the differences in endoscopic yield between patients of different age groups or symptom variables. Compared with male patients, female patients had a similar rate of abnormal endoscopic findings. A significant difference was found in the endoscopic yield among the four age groups; infants and teenagers aged 13–18 years were much more likely to have abnormal endoscopic findings than participants in other age groups (p<0.05). The endoscopic yield was the highest in patients with dysphagia (94.1%), followed by unexplained anemia (90.9%), GI bleeding (83.0%), weight loss (68.5%), recurrent vomiting (66.9%), GER symptoms (66.5%), diarrhea (66.3%), and abdominal pain (62.5%).
Table 2. Patient characteristics according to endoscopic and histologic findings
Variable |
Total patients n=2268 (%) |
With endoscopic findings n= (%) |
P value |
With histologic findings n= (%) |
P value |
Sex |
|||||
Male |
1294 |
808 (62.4%) |
0.928 |
415 (32.1%) |
0.732 |
Female |
974 |
610 (62.6%) |
319 (32.8%) |
||
Age |
|||||
0–1 y |
16 |
13 (81.3%) |
0.001 |
11 (68.8%) |
<0.001 |
1–5 y |
292 |
183 (62.7%) |
74 (25.3%) |
||
5–12 y |
1575 |
951 (60.4%) |
506 (32.1%) |
||
Teenagers (13–18 y) |
385 |
271 (70.4%) |
143 (37.1%) |
||
Abdominal pain |
|||||
+ |
1954 |
1215 (62.5%) |
0.401 |
602 (30.8%) |
<0.001 |
- |
314 |
203 (64.7%) |
132 (42.0%) |
||
Recurrent vomiting |
|||||
+ |
706 |
472 (66.9%) |
0.004 |
221 (31.3%) |
0.468 |
- |
1562 |
946 (60.6%) |
513 (32.8%) |
||
Weight loss |
|||||
+ |
343 |
235 (68.5%) |
0.013 |
130 (37.9%) |
0.017 |
- |
1925 |
1183 (61.5%) |
604 (31.4%) |
||
Gastroesophageal reflux symptoms |
|||||
+ |
191 |
127 (66.5%) |
0.236 |
74 (38.7%) |
0.049 |
- |
2077 |
1291 (62.2%) |
660 (31.8%) |
||
Gastrointestinal bleeding |
|||||
+ |
106 |
88 (83.0%) |
0 |
42 (39.6%) |
0.102 |
- |
2162 |
1330 (61.5%) |
692 (32.0%) |
||
Diarrhea |
|||||
+ |
83 |
55 (66.3%) |
0.473 |
28 (33.7%) |
0.786 |
- |
2185 |
1363 (62.4%) |
706 (32.3%) |
||
Unexplained anemia |
|||||
+ |
11 |
10 (90.9%) |
0.051 |
6 (54.6%) |
0.115 |
- |
2257 |
1408 (62.4%) |
728 (32.3%) |
||
Dysphagia |
|||||
+ |
17 |
16 (94.1%) |
0.007 |
5 (29.4%) |
0.794 |
- |
2251 |
1402 (62.3%) |
|
729 (32.4%) |
|
The results of the logistic regression analyses of the symptom variables associated with endoscopic yield are shown in Table 3. On univariate analysis, the endoscopic yield was associated with dysphagia, GI bleeding, weight loss, and recurrent vomiting. On multivariable analysis, dysphagia (p<0.05), GI bleeding (p<0.01), and recurrent vomiting (p<0.05) were independently associated with endoscopic yield.
Table 3. Univariate and multivariate regression analyses of endoscopic findings
Variable |
Univariate analysis |
Multivariate analysis |
||||||
|
OR |
95% CI |
P value |
OR |
95% CI |
P value |
||
Age |
1.02 |
0.99–1.05 |
0.125 |
|||||
Sex |
1.01 |
0.85–1.20 |
0.928 |
|||||
Abdominal pain |
0.9 |
0.70–1.15 |
0.402 |
|||||
Recurrent vomiting |
1.31 |
1.09–1.58 |
0.004 |
1.21 |
1.00–1.47 |
0.046 |
||
Weight loss |
1.36 |
1.07–1.74 |
0.013 |
1.27 |
0.99–1.64 |
0.057 |
||
Gastroesophageal reflux symptoms |
1.21 |
0.88–1.65 |
0.237 |
|||||
Gastrointestinal bleeding |
3.06 |
1.83–5.12 |
0 |
2.85 |
1.70–4.79 |
0 |
||
Diarrhea |
1.18 |
0.75–1.88 |
0.473 |
|||||
Unexplained anemia |
6.03 |
0.77–47.19 |
0.087 |
|||||
Dysphagia |
9.69 |
1.28–73.19 |
0.028 |
9.61 |
1.27–72.7 |
0.028 |
Pathologic findings and yield
Biopsy samples were obtained from 2235 patients and pathologically examined. The overall prevalence of histological abnormalities was 32.3%. Table 2 shows the differences in histologic yield between patients of different age groups or symptom variables. Compared with male patients, female patients had a similar rate of occurrence of histologic abnormalities. A significant difference in histologic yield was found among the four age groups, with infants most likely to have histological abnormalities (p<0.05). The histologic yield was the highest in patients with unexplained anemia (54.6%), followed by GI bleeding (39.6%), GER symptoms (38.7%), weight loss (37.9%), diarrhea (33.7%), recurrent vomiting (31.3%), abdominal pain (30.8%), and dysphagia (29.4%).
The results of the logistic regression analyses of the symptom variables associated with histologic yield are shown in Table 4. On univariate analysis, the histologic yield was positively associated with factors such as age and weight loss. Patients who had the indication of abdominal pain appeared to have less likelihood of histologic abnormality than those without. On multivariable analysis, weight loss (p<0.05) and age (p<0.05) remained positively correlated with histologic yield.
Table 4. Univariate and multivariate regression analyses of histologic findings
Variable |
Univariate analysis |
Multivariate analysis |
||||
OR |
95% CI |
P value |
OR |
95% CI |
P value |
|
Age |
1.04 |
1.01–1.07 |
0.009 |
1.05 |
1.02–1.08 |
0.001 |
Sex |
1.03 |
0.86–1.23 |
0.732 |
|||
Abdominal pain |
0.61 |
0.48–0.78 |
0 |
0.59 |
0.46–0.76 |
0 |
Recurrent vomiting |
0.93 |
0.77–1.13 |
0.468 |
|||
Weight loss |
1.33 |
1.05–1.69 |
0.018 |
1.37 |
1.07–1.74 |
0.01 |
Gastroesophageal reflux symptoms |
1.36 |
1.00–1.84 |
0.05 |
|||
Gastrointestinal bleeding |
1.39 |
0.93–2.08 |
0.103 |
|||
Diarrhea |
1.07 |
0.67–1.70 |
0.786 |
|||
Unexplained anemia |
2.52 |
0.77–8.29 |
0.128 |
|||
Dysphagia |
0.87 |
0.31–2.48 |
0.794 |
The indication of abdominal pain
A significant difference was found in the endoscopic yield among the patients with abdominal pain (Table 5). In these patients, those with any other additional symptoms such as vomiting, weight loss, GER symptoms, gastrointestinal bleeding, diarrhea, unexplained anemia, and dysphagia were much more likely to have abnormal endoscopic findings than those without (p<0.05). Additionally, patients with abdominal pain and greater than three other symptoms had the highest rate of abnormal endoscopic findings than other groups (p<0.05). Patients with abdominal pain and evidence of Helicobacter pylori infection were much more likely to have abnormal endoscopic findings than those without (p<0.05). There was no significant difference between patients with family members having Helicobacter pylori infection and patients without.
Table 5. Endoscopic findings in different groups of patients with abdominal pain
groups |
with endoscopic findings, n= (%) |
P value |
without any other symptoms |
473 (43.0%) |
|
with any other symptoms |
437 (51.2%) |
0 |
1 |
285 (48.0%) |
|
2 |
115 (56.7%) |
|
3 |
37 (66.1%) |
0.007 |
Helicobacter pylori infection evidence |
||
+ |
758 (44.2%) |
|
- |
152 (63.9%) |
0.000 |
family members with Helicobacter pylori infection |
||
+ |
750 (46.1%) |
|
- |
160 (48.8%) |
0.379 |
Pediatric EGD aids the understanding of the pathophysiology of common GI disorders in children and plays an important role in the management of some disorders. Parallel to the utilization of endoscopy in pediatric patients, the volume of EGD being performed has increased. When using this tool, we need to review its use to maximize its efficacy.
In this multicenter study, we found that abdominal pain was the most common indication, followed by recurrent vomiting, weight loss, GER symptoms, GI bleeding, diarrhea, and others. According to studies performed in large children’s hospitals or pediatric clinics, abdominal pain is also the most common indication for upper GI endoscopy in UK and US cohorts (12–14). A retrospective analysis (carried out over 20 years from 1985 to 2005) of children and adolescents who underwent EGD at a single center revealed that the proportion of patients with abdominal pain increased from 23–43%, while that of patients with GI bleeding declined from 34–5% over the 20-year interval (1). Studies in some small countries have found the most frequent indication in their centers was surveillance for esophageal varices and suspected celiac disease (15–17).
Currently, guidelines provide the indications for EGD; the guidelines developed and newly modified in 2014 by ASGE provide indications for EGD (6). The guidelines developed in 2015 by ESPGHAN and ESGE also elaborate indications for EGD (7). These two guidelines have the following in common: abdominal pain, weight loss, failure to thrive, unexplained anemia, dysphagia or odynophagia, caustic ingestion, recurrent vomiting with unknown cause, GI bleeding, diarrhea/malabsorption (chronic), and intractable or chronic symptoms of GERD. However, unexplained irritability, anorexia, and suspicion of graft versus host disease differentiate them.
In previous retrospective studies, the diagnostic yield of EGD was significantly different, varying from 18.9–79% (13–18). Several more recent studies have identified certain basic patient characteristics including age that affect diagnostic yield. The diagnostic yield was found to be higher in teenagers (11, 13, 14, 19) and lower in those aged < 7 years, especially in those without alarming symptoms (20). Our findings suggest that EGD is valuable for evaluating patients with digestive symptoms; the prevalence of abnormal endoscopic findings was 62.5% and that of abnormal histological findings was 32.4%. In this study, we found that the infant group had a higher yield of abnormal endoscopic and histological findings. In addition, the teenager group appeared to have a higher yield of abnormal endoscopic and histological findings, similar to the results of previous studies (11, 13, 14, 19).
Except for the basic characteristics, the composition of the patients in the study also affected the overall diagnostic yield, because diagnostic yield varies in patients with different indications for EGD. In our multivariate regression analyses, we observed that endoscopic yield was associated with dysphagia, GI bleeding, and recurrent vomiting, and histological yield was associated with weight loss and age. Abdominal pain was the most common indication; however, patients with abdominal pain had a lower rate of abnormal endoscopic or histological findings than those without. This finding is similar to that of a large retrospective study of 1,000 children in 2013, in which the most common indication (abdominal pain) had lower rates of abnormal endoscopic findings (28.9%) than other indications including stricture which was confirmed on upper GI series (100%), foreign body (88%), GI bleeding (57%), dysphagia (56%), and positive celiac screening (52%) (14). Other studies have also shown that patients with generalized abdominal pain had a lower rate of abnormal endoscopic findings (36%) than those with UGI bleeding (71.3%), variceal surveillance (54.8%), recurrent persistent vomiting (38%), and dyspepsia (37.8%) (17). Unexplained chronic abdominal pain (CAP) remains one of the most common indications for EGD in children, but the role of EGD in these patients has been debated because the majority of them have functional disorders. The guidelines developed in 2015 by ESPGHAN and ESGE included, as an indication, abdominal pain with suspicion of an organic disease, but this organic disease needs to be better defined. In the procedures performed for CAP, the rates of diagnostic abnormalities varied between 26.2% and 87.4% (21–24). However, the diagnostic yield could be higher in CAP patients with alarming symptoms (22) as well as a previously positive noninvasive Helicobacter pylori test (25) than in those without. This study might have tried to explain which CAP patient should be suspected of having an organic disease, but further studies are needed to confirm this hypothesis. In our study, we found that those patients with abdominal pain and any other symptoms, or evidence of Helicobacter pylori infection before EGD had higher rates of abnormal endoscopic findings, and seemed to have more necessity to receive EGD.
In conclusion, EGD is valuable for diagnosis in children with digestive symptoms, especially those with alarming features such as dysphagia, GI bleeding, and recurrent vomiting. Biopsy and histological examinations should be performed more aggressively in infants, those with weight loss or without abdominal pain.
EGD esophagogastroduodenoscopy
GI gastrointestinal
ASGE American Society for Gastrointestinal Endoscopy
ESGE European Society of Gastrointestinal Endoscopy
ESPGHAN European Society for Pediatric Gastroenterology Hepatology and Nutrition
CRF case report form
GER gastroesophageal reflux
CAP chronic abdominal pain
Ethics approval and consent to participate: The research protocol was assessed and ethics approval was obtained from all five hospitals, including Research Ethics Committee of Children’s Hospital of Fudan University, Research Ethics Committee of Wuhan Children’s Hospital, Research Ethics Committee of Henan Children’s Hospital (Zhengzhou Children’s Hospital), Research Ethics Committee of Hunan Children’s Hospital, Research Ethics Committee of Children’s Hospital of Fudan University Xiamen Branch. Informed consent was obtained from a parent or other legal guardians of all individual participants included in the study. All the methods were performed in accordance with the relevant guidelines and regulations.
Consent for publication: Not applicable.
Availability of data and materials: The datasets used and analyzed during the current study are not publicly available due not all of the researchers wish to share the data with public at present, but available from the corresponding author on reasonable request.
Competing interests:The authors declare that they have no competing interests.
Funding: There is no funding for the research.
Authors’ contributions: All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Shengnan Wang, Xiaoxia Qiu, Ying Huang, Jingfang Chen, Hong Mei, Haiyan Yan and Jieyu You. The first draft of the manuscript was written by Shengnan Wang and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Acknowledgments: The authors would like to thank Chongfan Zhang (Associate Editor-in-Chief, Chinese Journal of Evidence-Based Pediatrics) for his professional direction in choosing the design of this study. We also thank Yi Zhang (Department of Clinical Epidemiology, Children’s Hospital of Fudan University, Shanghai, China) for his assistance with the statistical analysis.