OUP user menu

Journal of Crohn's and Colitis: 10 (9)

Editor-in-Chief

Laurence J. Egan, Ireland

Associate Editors

Shomron Ben-Horin, IsraelSilvio Danese, ItalyPeter Lakatos, HungaryMiles Parkes, UKJesús Rivera-Nieves, USABritta Siegmund, GermanyGijs van den Brink, NLSéverine Vermeire, Belgium

6.585
5.586

Published on behalf of

The value of focally enhanced gastritis in the diagnosis of pediatric inflammatory bowel diseases

Kleoniki Roka, Eleftheria Roma, Kalliopi Stefanaki, Ioanna Panayotou, Giannis Kopsidas, Giorgos Chouliaras
DOI: http://dx.doi.org/10.1016/j.crohns.2012.11.003 797-802 First published online: 1 November 2013

Abstract

Background and aims: Focally enhanced gastritis (FEG) has been suggested as a diagnostic marker for patients with Crohn’s disease. In this study we evaluated the prevalence of FEG in children with inflammatory bowel diseases (IBD) and assessed the ability of FEG to distinguish IBD from non-IBD patients.

Methods: A retrospective study of the children who underwent esophagogastroduodenal endoscopy (EGD) during 2004–2011 was performed, after excluding individuals with H. pylori infection and celiac disease. Two groups were studied: patients with IBD (IBD group, n = 185) and non-IBD patients who underwent endoscopy of the upper gastrointestinal tract for various abdominal complaints (non-IBD group, n = 684). Relation of FEG to age and gender was also assessed.

Results: FEG was found significantly more frequently among children with IBD (35.7% vs 3.4%, respectively, p < 0.001). Children with FEG were 15.4 times more likely to have IBD than to belong in the non-IBD group. All types of IBD had significantly higher frequencies of FEG compared to non-IBD individuals (Crohn's disease: 54.1%, ulcerative colitis: 21.6%, IBD unclassified: 18.4%, all three comparisons with the non-IBD group: p-values < 0.001). FEG positivity was more common in females compared to males with Crohn's disease and ulcerative colitis and in children younger than 2 years in the IBD-unspecified group. FEG achieved a sensitivity of 35.7% and specificity of 96.6% in distinguishing between IBD from non-IBD patients.

Conclusions: FEG has significantly higher prevalence in children with IBD, particularly Crohn's disease and can be a valuable supporting finding in cases of indefinite diagnosis.

Keywords
  • Focally enhanced gastritis
  • Inflammatory bowel disease
  • Crohn's disease
  • Ulcerative colitis
  • Inflammatory bowel disease unclassified
  • Prognostic factors
Abbreviations
FEG
Focally enhanced gastritis
IBD
Inflammatory bowel disease
CD
Crohn's disease
UC
Ulcerative colitis
PPV
positive predictive value
NPV
negative predictive value

1 Introduction

The diagnosis and classification of pediatric inflammatory bowel disease (IBD) are a complex task incorporating clinical, endoscopic, radiologic and histologic information.1 In children, a macroscopically normal gastroscopy and ileocolonoscopy do not preclude the diagnosis of IBD. In this context the pathologist's report becomes even more crucial in the diagnostic process. The histopathological features of Crohn's disease (CD) and ulcerative colitis (UC) are well described; however, distinguishing between these two entities in clinical practice can be, occasionally, difficult.1 Especially in children, problems with definitive diagnostic classification are more common considering that 4–30% of pediatric IBD patients are classified as having inflammatory bowel disease unclassified (IBDU).1,2 Today, early and accurate distinction between CD, UC and IBDU is becoming increasingly important, as refinements in treatment are being identified.

Involvement of the upper gastrointestinal tract was traditionally considered to be a characteristic of Crohn's disease. In nowadays, it is well recognized that UC and IBDU may also affect the upper gastrointestinal tract.3 Several studies in adults have shown that focally enhanced gastritis (FEG) is being recognized with increasing frequency in IBD.3 The term “focally enhanced gastritis” was first introduced by Oberhüber in 1997.4 It is used in order to describe perifoveolar or periglandular mononuclear or neutrophilic infiltrates around gastric crypts.5 The lymphocytes are typically CD3 positive and the histiocytes are CD68R positive. The prevalence of FEG seems to differ among IBD patients, being more commonly identified in CD patients (up to 76%)6 compared to UC (approximately 20%)7 and in younger patients compared to older.3,8 Data in pediatric IBD populations are scarce and the few published reports indicate that FEG could be a useful marker in diagnosing IBD, as well as discriminating between CD, UC and IBDU. The assessment of the diagnostic value of focal gastritis requires knowledge of the likelihood of FEG in non-IBD patients. Pediatric data in this field are even more limited and based on small samples. The prevalence of FEG in the general adult population is approximately 3%, once H. pylori gastritis and reactive gastropathy have been excluded.3

In this study we aimed at estimating the prevalence of FEG in a large number of newly diagnosed, treatment-naive children with IBD and in non-celiac, non-H. pylori children undergoing upper gastrointestinal endoscopy. We also assessed gender and age variations of FEG in both IBD and non-IBD individuals. Finally we evaluated the performance of FEG as a criterion in the distinction of IBD from non-IBD children, as well as between CD, UC and IBDU.

2 Methods

This retrospective study was conducted at the Gastroenterology Unit of the 1st Department of Paediatrics in the University of Athens, in “Aghia Sophia” Children's Hospital. The Ethics Committee of the hospital approved permission for medical review, waiver of informed consent and anonymous publication of data according to the Declaration of Helsinki.

Initially, all endoscopic and pathology reports during an 8-year period (2004–2011) were retrieved. We constrained the retrospective collection of data in that period in order to ensure that all procedures, including endoscopies, biopsies and microscopic analyses, would be complete and comparable for all individuals without any missing data. Subsequently, only children undergoing their first upper gastrointestinal endoscopy were included. Children who were later diagnosed with celiac disease or H. pylori infection, as well as children who had received, prior to the endoscopy, steroids, proton pump inhibitors or H2 antagonists were excluded. Finally two groups were formed: children with a diagnosis of IBD and the control group which consisted of non-celiac, non-H. pylori children undergoing a first upper gastrointestinal endoscopy for investigation of abdominal complaints. Diagnosis of IBD was based on the Porto criteria9 and further classified into three subgroups: CD, UC and IBDU.

The endoscopies were performed by trained pediatric gastroenterologists and the following biopsies were taken in all cases: second part of duodenum (2–4 biopsies), duodenal bulb (1 biopsy), antrum (1 biopsy), body (1 biopsy) and esophagus (3–6 biopsies). Specimens, were fixed in 10% formalin, and embedded in paraffin, and sections were stained with hematoxylin–eosin and modified Giemsa or Masson trichrome and assessed under light microscopy. All histological analyses were performed by a single, highly experienced histopathologist. Classification of gastritis was according to that of the updated Sydney system.10 Focal enhanced gastritis was defined as presence in biopsy material of focal inflammatory lesions composed mainly of lymphocytes and histiocytes, and occasionally neutrophils involving at least one or a few adjacent foveolae/glands.1,5 The definition of FEG was uniform across the study groups and patients with Crohn's disease and gastric involvement were not characterized as FEG-positive unless they fulfilled the described criteria.

In the IBD group, patients with Crohn's disease were classified as proposed by the Paris classification11: L1: distal 1/3 ileum ± limited cecal disease, L2: colonic disease, L3: ileocolonic disease and with regard to the involvement of the upper gastrointestinal tract (L4a+: involvement of the upper gastrointestinal tract, proximal to the ligament of Treitz, L4a −: no involvement). As different imaging modalities were used for detection of small bowel disease, the L4b classification was not applied.

Categorical variables are presented as absolute (n) and relative frequencies (%) and compared by Fisher's exact test. Quantification of the effect of categorical variables on the probability of IBD diagnosis and type of IBD diagnosis was done by logistic regression analysis and exact logistic regression analysis. Results are reported as odds ratios (OR) with 95% confidence intervals (95% ci). For the sensitivity analysis of FEG positivity, sensitivities and specificities were determined for the following outcomes: diagnosis of IBD vs non-IBD, CD vs UC, CD vs IBDU and UC vs IBDU. Positive and negative predictive values were calculated and presented based on several a priori probabilities. Τhe acceptable level of statistical significance was set to ≤ 0.05. All analyses were performed using Stata 11.0 MP statistical software (Stata Corp, Texas, USA).

3 Results

A total of 1628 patients were evaluated with esophagogastroduodenal endoscopy (EGD) between 2004 and 2011. After excluding individuals with H. pylori infection, celiac disease, as well as patients who had undergone previous endoscopy or were under treatment, 869 individuals (48.7% females) were included in the analysis. The non-IBD group consisted of 684 children, whereas 185 participants were diagnosed with IBD: 85 (45.9%) had CD, 51 (27.6%) had UC and 49 (26.5%) had IBDU. Distribution of age and gender for all study groups is shown in Table 1. The symptoms and diagnoses for the 684 children in the control group are illustrated in Table 2. Twenty-three children in the control group were FEG-positive and no other concomitant diagnosis was established.

View this table:
Table 1

Age and gender in the study population.

Non-IBDCDUCIBDUp-Value
Gender, females, n (%)317 (46.4%)46 (54.1%)33 (64.7%)27 (55.1%)0.036
Age< 0.001
< 2 years, n (%)166 (24.3%)0 (0%)3 (5.8%)7 (14.3%)
2–10 years, n (%)328 (47.9%)38 (44.7%)24 (47.1%)24 (49.0%)
> 10 years, n (%)190 (27.8%)47 (55.3%)24 (47.1%)18 (36.7%)

IBD: Inflammatory bowel disease.

CD: Crohn Disease.

UC: Ulcerative Colitis.

IBDU: Inflammatory bowel disease unclassified.

    View this table:
    Table 2

    Clinical characteristics (2a) and diagnoses (2b) in the control group.

    2a2b
    Symptomsn (%)aDiagnosesn (%)
    Epigastric pain232 (33.9%)Normal232 (33.9%)
    Chronic persistent abdominal pain43 (6.4%)Non-specific gastritis223 (32.6%)
    Recurrent or persistent vomiting117 (17.1%)Gastric ulcer3 (0.34%)
    Retrosternal pain/burn26 (3.8)Ulcer of the duodenal bulb9 (1.3%)
    Early satiety5 (0.8%)Eosinophilic gastroenteritis40 (5.8%)
    Recurrent abdominal distention8 (1.1%)Esophageal eosinophilia56 (8.3%)
    Chronic diarrhea43 (6.4%)Food allergy enteropathy51 (7.4%)
    Severe refractory constipation11 (1.7%)Esophagitis37 (5.4%)
    Failure to thrive86 (12.6%)Barrett esophagus3 (0.43%)
    Upper gastrointestinal bleeding80 (11.8%)Gastric polyp3 (0.43%)
    Refractory iron deficient anemia29 (4.2%)CMVb gastritis4 (0.58%)
    Dysphagia38 (5.6%)FEGc23 (3.3%)
    Food refusal/feeding difficulties18 (2.6%)
    Ingestion of corrosive substances15 (2.3%)
    • a The sum of absolute numbers and percentages is larger than 684 and 100% respectively because some patients reported more than one chief complaint.

    • b CMV: Cytomegalovirus.

    • c FEG: Focally enhanced gastritis.

    The prevalence of FEG in the entire study population was 10.2% (95% ci: 8.3–12.4) and it was found significantly more frequently in the IBD group (IBD vs non-IBD: 35.7% ci: 28.8–43.9 vs 3.4% ci: 2.1–5.0, respectively, p < 0.001). All types of IBD had significantly higher frequencies of FEG compared to non-IBD individuals (CD: 54.1% ci: 42.9–65.0, UC: 21.6% ci: 11.3–35.3, IBDU: 18.4% ci: 8.8–32.0, all three comparisons with the non-IBD group: p-values < 0.001). The distribution of the probability of FEG in sub-groups defined by gender or age is shown in Table 3. The analysis showed that FEG positivity is more common in females compared to males CD and UC patients, although the results, marginally, did not reach statistical significance. In addition, in the IBDU group younger (< 2 years) children had significantly higher chance of being FEG (+), whereas in CD and UC the probability appears to be invariably distributed across all age groups. For 70 patients with Crohn's disease a full ileocolonoscopy was available. The probability of FEG positivity was not related to the involvement of the lower gastrointestinal tract (FEG (+): L1: 44.4%, L2: 69.3%, L3:54.8%, L0 (normal ileocolonoscopy: 25%), p = 0.27), nor to the involvement of the upper gastrointestinal tract (FEG (+): L4a+: 52.4%, L4a −: 54.8%, p = 0.99). One out of the 4 patients with normal ileocolonoscopy (25%) had FEG.

    View this table:
    Table 3

    FEG positivity according to gender (2a) and age (2b) stratifications.

    3a.
    Gender
    FEG (+), n (%)OverallFemalesMalesp-Value
    non-IBD23/684 (3.4%)10/317 (3.1%)13/367 (3.5%)0.8
    CD46/85 (54.1%)29/46 (63.0%)17/39 (43.6%)0.08
    UC11/51 (21.6%)10/33 (30.3%)1/18 (5.6%)0.07
    IBDU9/49 (18.4%)4/27 (14.8%)5/22 (22.7%)0.7
    All IBD66/185 (35.7%)43/106 (40.6%)23/79 (29.1%)0.12
    3b.
    Age
    FEG (+), n (%)Overall< 2 years2–10 years> 10 yearsp-Value
    Non-IBD23/684 (3.4%)4/166 (2.4%)13/328 (4.0%)6/190 (3.2%)0.7
    CD46/85 (54.1%)21/38 (55.3%)25/47 (53.2%)0.9
    UC11/51 (21.6%)1/3 (33.3%)5/24 (20.8%)5/24 (20.8%)0.8
    IBDU9/49 (18.4%)4/7 (57.1%)2/24 (8.3%)3/18 (16.7%)0.027
    All IBD66/185 (35.7%)5/10 (50.0%)28/86 (32.6%)33/89 (37.1%)0.5

    FEG: Focally enhanced gastritis.

    IBD: Inflammatory bowel disease.

    CD: Crohn Disease.

    UC: Ulcerative Colitis.

    IBDU: Inflammatory bowel disease unclassified.

      Logistic regression analysis estimated that children with FEG were 15.4 times more likely to have IBD compared to children without (OR: 15.4, 95% ci: 8.9–26.4), after adjusting for age and gender which were significant confounders in the final model as shown in Table 4. Within the IBD group, the frequency of FEG varied significantly with CD patients showing increased prevalence compared to both UC and IBDU individuals (p < 0.001). Individuals with FEG were 4.3 times more likely (OR = 4.3, 95% ci: 1.9–9.5) to have CD instead of UC and 5.2 times more likely (OR = 5.2, 95% ci: 2.3–12.1) to have CD instead of IBDU. No differences were observed between UC and IBDU.

      View this table:
      Table 4

      Multiple logistic regression model on the probability of having IBD vs non-IBD in the study population.

      Odds ratio95% cip-Values
      FEG (+)15.48.9–26.4< 0.001
      Gender, females vs males1.51.02–2.140.038
      Age
      2–10 years vs < 2 years4.42.1–9.1< 0.001
      > 10 years vs < 2 years7.83.7–16.2< 0.001
      > 10 years vs 2–10 years1.81.2–2.60.004

      FEG: Focally enhanced gastritis.

      IBD: Inflammatory bowel disease.

        The diagnostic performance of FEG in distinguishing between IBD patients from non-IBD individuals achieved a sensitivity of 35.7% (95% ci: 28.8–43.9) and specificity of 96.6% (95% ci: 95.0–97.8). Sensitivities and specificities after stratification according to age and gender are shown in Table 5. As shown above, FEG-positivity differed significantly between CD and UC patients as well as between CD and IBDU patients. The sensitivity and specificity for discriminating CD from UC patients were 54.1% (95% ci: 43.0–65.0) and 78.4% (95% ci: 64.7–88.7) respectively, while between CD and IBDU the corresponding values were 54.1% (95% ci: 43.0–65.0) and 81.6% (95% ci: 68.0–91.2) respectively. Fig. 1 presents positive and negative predictive values of FEG for a priori clinical probabilities of 10%, 25%, 50%, 75% and 90%.

        Figure 1

        Positive (ppv) and negative (npv) predictive values derived from the sensitivity analysis of FEG in distinguishing between IBD from non-IBD (1a), CD from UC (1b) and CD from IBDU (1c). Five distinct levels of a priori clinical probability (prevalence) have been hypothesized: 10%, 25%, 50%, 75% and 90%.

        FEG: Focally enhanced gastritis.

        IBD: Inflammatory bowel disease.

        CD: Crohn Disease.

        UC: Ulcerative colitis.

        IBDU: Inflammatory bowel disease unclassified.

        View this table:
        Table 5

        Diagnostic performance of FEG-positivity in distinguishing between IBD from non-IBD patients after stratification according to gender and age. Number in brackets indicate 95% confidence intervals.

        SensitivitySpecificity
        Gender
        Males29.1% (19.4–40.4)96.5% (94.0–98.1)
        Females40.6% (31.1–50.5)96.9% (94.3–98.5)
        Age
        < 2 years50.0% (18.7–81.2)97.6% (93.9–99.3)
        2–10 years32.6% (22.8–43.5)96.0% (93.3–97.9)
        > 10 years37.1% (27.1–48.0)96.8% (93.2–98.8)

        FEG: Focally enhanced gastritis.

        IBD: Inflammatory bowel disease.

          4 Discussion

          In this study, we analyzed, in the largest pediatric population so far, the value of FEG in the diagnosis of IBD and a strong association between these two entities was observed. The prevalence of FEG in children with IBD was 35.7% generating a relative risk of 15.4 compared to non-IBD individuals. Within the IBD group FEG-positivity supported the diagnosis of CD vs both UC and IBDU, but did not reliably differentiate between UC and IBDU. These conclusions are in agreement with Sharif et al.7, as well as with Hummel's recent study,12 which are the only available data in children, although in smaller samples. The ratio of FEG-positivity in IBD children in our analysis was lower compared to the abovementioned two articles where FEG was found in 50% and 60% of IBD children, respectively. This difference is mainly due to the reduced frequency of FEG in our CD sample compared to these two studies (54.1% vs 65.1% and 69.0%), whereas in the UC subgroup the findings were similar. The observed discrepancy could be attributed to the different timings of endoscopy and probable treatment effects, as we only included treatment-naive patients undergoing their first GED which implies relatively short disease duration. In adults, significantly more data have been published. Overall, most studies have shown similar trends in the distribution of FEG in IBD adult patients, although absolute numbers are not directly comparable to pediatric populations. Focally enhanced gastritis is commonly seen in adult Crohn's disease (43%)13 but less frequently found in ulcerative colitis (29%).14

          The sensitivity analysis demonstrated that FEG is a very specific marker in identifying IBD from non-IBD cases, which results in high positive predictive values, especially in patients with increased clinical suspicion. Practically, whenever the a priori clinical probability (prevalence) is over 25%, finding FEG in GED produces positive predictive values higher than 82%. Furthermore, the probability of an IBD diagnosis exceeds 91% when the clinical suspicion is over 50%. Approximately 10% of the children with Crohn's disease have exclusive involvement of the small intestine with normal terminal ileum.15,16 In such cases ileocolonoscopy is macroscopically normal and biopsy findings may not be that of typical Crohn's disease. Clinicians often are based on small bowel imaging studies to set the diagnosis. In patients with those specific features (compatible clinical picture and inconclusive ileocolonoscopy) where there is strong suspicion of CD, the presence of FEG significantly increases the probability and has clearly an important diagnostic value. Reversely, the relatively low sensitivity reduces the ability of the absence of FEG to exclude IBD. This implies that clinicians cannot depend on a negative-for-FEG pathology report to rule out IBD. Within the IBD group, there are some evidence supporting that FEG enhances the probability of CD against UC and IBDU. However, it is clear that the capacity of FEG as a diagnostic marker in that field is limited and could be used only as a surrogate index. Hummel et al.12 reported that FEG exhibited a specificity of 87.1% and positive predictive value of 78.6% for CD, although the authors did not specify vs which group these figures were achieved and on what a priori probability the positive predictive value was calculated. Nevertheless, their results also indicate that FEG may be useful in supporting, rather than excluding, a probable IBD diagnosis.

          According to our findings, FEG was detected in 3.4% of non-IBD patients. This result corresponds with previous studies in children reporting that FEG ranges from 3 to 5%.3,17 The fluctuations between different researchers are probably due to the inconsistent definition of the control groups. Hummel et al.12 found FEG in 7% of their non-IBD patients. The higher prevalence could be the result of different characteristics among non-IBD individuals. In that particular study, the non-IBD group consisted of “suspected IBD cases”, implying that it was a selected group, with specific clinical features. Reviewing the medical records in our non-IBD children ensured that no atypical cases were included and therefore a more appropriate control group was formed.

          An important observation was the age and gender-dependent variations of FEG in IBDU patients. It seems that FEG is much more common in females and younger patients with IBDU. A direct consequence is the increased sensitivity of FEG in diagnosing IBDU individuals with those specific features. A similar finding occurred in CD patients in relation to gender, but not age. Females, again, had higher probability of FEG positivity compared to males, although the result was marginally not significant. To our knowledge these outcomes have not appeared in literature before. Further studies are needed to confirm whether these two parameters should be taken into account when evaluating the results of an EGD in suspected cases.

          The advantage of this retrospective study is the large number of children included, the uniformity in terms of gastroenterologic and histopathological assessments and the clear definition of patients and non-IBD individuals. However, our results should be interpreted taking under consideration the following limitations. The non-IBD group is a clinical sample and not an average of the general pediatric population. The retrospective nature and clinical-observational design of the study did not allow the histopathologists to be blind with respect to the clinical information and the endoscopic finding of the patients. In addition, the lack of immune-histochemical staining to identify or confirm borderline histological findings and the number of biopsies taken may have been suboptimal to detect FEG especially in IBD cases. The cross sectional, at the time of first GED, assessment of the participants, reduces the power of this study to record the long term outcome of non-IBD children with FEG and to identify whether this group is at increased risk for developing gastrointestinal pathology in the future.

          In conclusion, the presence of FEG in pediatric individuals strongly suggests underlying IBD, especially when considering FEG-positive females and children younger than 2 years of age.

          Conflict of interest statement

          All the authors have no conflict of interest to declare.

          Acknowledgments

          All authors were involved in the design of the study, interpretation of results, preparation and final approval of the manuscript. Niki Roka and Giannis Kopsidas were involved in data entry. Giorgos Chouliaras and Giannis Kopsidas performed the statistical analysis. Eleftheria Roma supervised the project and reviewed the manuscript.

          References

          View Abstract