Helicobacter pylori-negative Chronic Gastritis in Children: A Systematic Review

Abstract

Objectives: 

To systematically review the current evidence on Helicobacter pylori-negative chronic gastritis including natural history, available therapies and outcomes.

Methods: 

Articles providing data on the prevalence, treatment or outcomes of Helicobacter pylori-negative gastritis were identified through a systematic search in the MEDLINE and EMBASE databases. All original research articles from human studies until October 31, 2021, were included.

Results: 

A total of 54 studies were included consisted of eosinophilic gastritis (n = 9), autoimmune gastritis (n = 11), collagenous gastritis (n = 16), focally enhanced gastritis (n = 6), lymphocytic gastritis (n = 5) and other causes including idiopathic gastritis and chronic renal failure related (n = 7). Most of the included studies were either cross-sectional or longitudinal cohorts except for collagenous gastritis, which mainly included case reports and case series. The prevalence of paediatric eosinophilic gastritis ranges between 5 and 7/100,000 and patients have generally favourable outcome with 50% to 70% clinical and histological response to either corticosteroids or elimination diets. Autoimmune gastritis and collagenous gastritis are extremely rare entities, commonly present with refractory iron deficiency anaemia, while lymphocytic gastritis is relatively common (10%–45%) in children with coeliac disease. Data on treatments and outcomes of autoimmune, collagenous, and focally enhanced gastritis are lacking with limited data implying poor response to therapy in the former 2 diagnoses.

Conclusions: 

Helicobacter pylori-negative gastritis is uncommonly reported, mainly in small cohorts, mixed adult-paediatric cohorts or as sporadic case reports. As common symptoms are not specific, thus not always result in an endoscopic evaluation, the true prevalence of these distinct disorders may be underestimated, and thus under reported.

What Is Known/What Is New

What Is Known

• Helicobacter pylori is the most common cause of gastritis in adults and children.
• Chronic gastritis may be attributed to other causes, such as eosinophilic, autoimmune, and collagenous gastritis as well as upper gastrointestinal involvement of inflammatory bowel diseases.

What Is New

• Data regarding Helicobacter pylori-negative chronic gastritis is lacking.
• Eosinophilic gastritis is not extremely rare and commonly responds to elimination diets or corticosteroids.
• Autoimmune and collagenous gastritis are rarely reported in children and tend not to respond to available therapies, such as proton pump inhibitors or budesonide.

Gastritis is defined as microscopic evidence of inflammation affecting the gastric mucosa without specific clinical presentations or symptoms defining it. Helicobacter pylori is the most common cause of chronic gastritis, worldwide, in adults and children ⁽¹⁾. The incidence of chronic gastritis in children is not well described but limited evidence implies that the incidence ranges from 2% to 8% ^(2,3) with incidence increasing after the age of 10 years ⁽⁴⁾. H pylori-negative gastritis is considered when, in the presence of histological inflammation, staining of gastric mucosal biopsies for H pylori is negative ⁽⁵⁾. In the absence of medications, which may be the cause of acute gastritis (such as nonsteroidal anti-inflammatory drugs, NSAIDs), other potential causes of gastritis should be considered (Table 1), such as autoimmune or atrophic gastritis, eosinophilic gastritis, collagenous gastritis, and upper gastrointestinal involvement of inflammatory bowel diseases (IBD). Other causes of gastritis may involve infectious microorganisms other than H pylori (mostly causing acute gastritis in immunocompromised patients), chronic renal failure, or lymphocytic gastritis.

TABLE 1 - Differential diagnosis of noninfectious chronic gastritis in children

Diagnosis	Prevalence	Characteristics	Associated disorders	Initial therapies	Outcome
Eosinophilic gastritis	5 to 10/100,000	>30 eosinophils/5 hpf, eosinophilic cryptitis or crypt abscess	Atopy, peripheral eosinophilia	Corticosteroids, elimination diet	Single to multiple flares
Autoimmune/atrophic gastritis	Not reported	Inflammatory infiltrates of plasma cells, lymphocytes, eosinophils, neutrophils, and mast cells in the oxyntic mucosa, patchy oxyntic gland destruction or fibrosis	Refractory iron-deficiency anaemia, autoimmune disorders, primary immune deficiency, presence of anti-parietal cell antibodies	H pylori eradication	Not reported
Collagenous gastritis	2/100,000	Subepithelial collagenous deposition ≥10 μm inflammatory infiltrate of the lamina propria	Refractory iron-deficiency anaemia	PPIs, sucralfate, corticosteroids, Azathioprine	Predominantly chronic course
Focally enhanced gastritis	50 to 70% in Crohn disease, 15% to 20% in ulcerative colitis	One or more foveola/glands surrounded and infiltrated by inflammatory cells (mainly lymphocytes and monocytes)	A common feature of Crohn disease	PPIs, other IBD-related therapies	Usually responds to an effective treatment for distal IBD
Lymphocytic gastritis	5% to 45% of children with coeliac disease	≥ 25 intraepithelial lymphocytes per 100 epithelial cells, variable chronic inflammatory pattern in the lamina propria	Coeliac disease, H Pylori gastritis, Ménétrier disease	Gluten-free diet in coeliac disease	Not reported

H pylori = Helicobacter pylori, IBD = inflammatory bowel disease, PPIs = proton pump inhibitors.

Most literature describes the epidemiology, characteristics, treatment and outcomes of H pylori gastritis; however, H pylori-negative chronic gastritis is not well characterised, particularly in children.

Therefore, we thought to perform a systematic review on the prevalence, treatment and outcomes of H pylori-negative chronic gastritis in the paediatric population with an emphasis on 5 main entities: eosinophilic gastritis, autoimmune/atrophic gastritis, collagenous gastritis, focally enhanced gastritis, and lymphocytic gastritis.

METHODS

Articles providing data on the prevalence, treatment or outcomes of non-H pylori gastritis were identified through a systematic search in the MEDLINE and EMBASE databases (including cross referencing) by using various combinations of the terms ‘gastritis’, ‘gastric’, ‘ulcer’, ‘eosinophilic’, ‘collagenous’, ‘atrophic’, ‘atrophy’, ‘autoimmune’, ‘lymphocytic’, ‘chronic’, ‘inflammatory bowel disease’, ‘Crohn’, ‘Crohn's’, ‘colitis’, ‘focally enhanced’, ‘celiac’, ‘coeliac’, ‘idiopathic’, ‘renal failure’, ‘children’ and ‘pedi-atric’/‘paediatric’.

All original research articles from human studies published in English until October 31, 2021 were included. We did not limit the start date in-order to allow complete representation of these rare entities. Titles and abstracts retrieved were screened for eligibility by 2 independent researchers. Due to the scarcity of evidence, all studies reporting any exposure or outcome of the selected topics were included. At full text review, each text was screened by 2 independent researchers to produce a final shortlist. Data were extracted from each article by 1 researcher, and checked by at least 1 other researcher.

The data retrieved from the eligible studies included: first author, year of publication, characteristics of the study population, study design, study intervention, and study outcomes.

This systematic review was not registered and it was performed according to an internal protocol.

RESULTS

The selection process of included studies is depicted in supplementary Figure 1, https://links.lww.com/MPG/C692. The 54 eligible studies included eosinophilic gastritis (n = 9), autoimmune or atrophic gastritis (n = 11), collagenous gastritis (n = 16), focally enhanced gastritis (n = 6), lymphocytic gastritis (n = 5), and other causes of paediatric gastritis including idiopathic gastritis and chronic renal failure-related gastritis (n = 7). Most of the included studies were either cross-sectional or longitudinal cohorts except for collagenous colitis, which mainly included case-reports and caseseries. Nonspecific gastritis associated with inflammatory bowel disease is described in numerous paediatric studies, thus, the systematic literature search included only the more specific form of focally enhanced gastritis.

Eosinophilic Gastritis

Eosinophilic gastrointestinal disorders (EGIDs) constitute a group of diseases characterised by abnormal eosinophilic infiltration, which may involve any part of the gastrointestinal tract, in the absence of known causes of intestinal eosinophilia. The differential diagnosis includes parasite infestation (such as Strongyloides and Toxocara species), drug reactions (such as NSAIDs), immune deficiencies (including human immunodeficiency virus, HIV), inflammatory conditions (such as IBD or autoimmune enteropathy), systemic mastocytosis, and neoplastic hypereosinophilic syndromes ⁽⁶⁾. EGIDs embrace eosinophilic esophagitis, eosinophilic gastritis (EoG), eosinophilic duodenitis, eosinophilic enteritis, and eosinophilic colitis ⁽⁶⁾. In EGID that spare the oesophagus, the stomach and the duodenum are the most frequently affected organs with different layers’ involvement: predominantly mucosal, muscle layer, and subserosal ^(7,8).

Nonoesophageal EGIDs are considered rare disorders and data regarding the incidence or prevalence of these conditions are limited ^(9,10). A recent population-based analysis described the overall prevalence of these disorders in children to be 5.3/100,000, slightly higher than in adults (5.1 per 100,000) ⁽¹¹⁾. Another recent study reported a prevalence of EoG and eosinophilic duodenitis/ enteritis as 6.3/100,000 and 8.4/100,000, respectively ⁽¹²⁾. Non-oesophageal EGID subjects seem to be frequently white with a history of atopy; however, no male predominance has been demonstrated, as in eosinophilic oesophagitis. Multisite inflammation was more common in children than in adults ⁽¹³⁾.

Unlike the oesophagus, eosinophils are a normal finding in other segments of the gastrointestinal tract and are believed to play a role in the mucosal immune response ⁽¹⁵⁾. Currently, the exact mechanism that underlies EGIDs is not completely understood, falling between a pure IgE-mediated and a delayed TH2-type response, which notably includes chemokines like eotaxin and cytokines like IL-5 ⁽¹⁴⁾. It seems to be related to several factors including genetic predisposition, environmental risk factors (such as food allergens), intestinal dysbiosis and impaired epithelial barrier ^(14,15). Activated eosinophils infiltrate the tissues, release cytotoxic granules, and induce an inflammatory response causing damage to the involved intestinal layer ⁽¹⁴⁾.

The clinical presentation of EGIDs depends on site, extent, and depth of disease in the gastrointestinal tract. EGIDs have been classified into 3 variants: mucosal, muscular, and subserosal, with the mucosal variant being the most common (approximately 90%) ⁽¹⁶⁾. In EoG, symptoms at presentation are highly variable. Abdominal pain is the most common symptom, followed by vomiting. Other main symptoms/signs include anaemia, melaena, food aversions, failure to thrive, oedema, and ascites. Severe gastric eosino-philia has also been associated with a protein-losing enteropathy even without duodenal eosinophilia ⁽⁹⁾ and rarely with gastric outlet obstruction in patients with the muscular type ⁽¹⁷⁾.

According to the criteria proposed by Talley et al, the diagnosis of EoG should be based on both clinical symptoms and histological evidence of significant gastric eosinophilia (>30 eosinophils/5 high power fields, hpf) in the gastric antrum and/or body, whereas endoscopic findings are not essential ⁽¹⁶⁾. Although the number of 30 eosinophils/hpf as the upper limit of normal for stomach and duodenum has been proposed, others have suggested lower or higher thresholds ⁽¹⁸⁾. Given the patchy nature of EoG, multiple random biopsies are warranted per site (at least 5–6 biopsies) ⁽¹⁹⁾. Supportive histologic findings include: eosinophils in clusters in the superficial lamina propria, eosinophils in the muscular propria, glandular destruction, the presence of intra-epithelial eosinophils and eosinophilic cryptitis or crypt abscess ^(14,20). Endoscopic findings in EoG may vary from normal appearance (most patients) to nodularity, erythema, erosions, and ulcera-tive lesions ^(10,21). About 70% of EGID patients present with peripheral eosinophilia (>500 cells/mL), and in two-thirds of eosinophilic duodenitis/enteritis patients, total serum IgE was reported to be ≥100IU/mL ^(22,23).

Treatment of EoG is predominantly based on dietary and corticosteroid therapies, as first-line treatment ⁽¹⁵⁾. In contrast to eosinophilic oesophagitis, the efficacy of dietary treatment in EGIDs has been described only in small case series. In a retrospective study ^(24,30) children with nonoesophageal EGIDs received different diets resulting in 82% clinical response and 78% histo-logical response. Nevertheless, only 6 children underwent histo-logic re-evaluation after treatment.

More recently, in a retrospective multicentre study on 376 patients (315 children) affected by EGID, a clinical improvement with dietary therapy has been demonstrated in all types of EGID with a significant reduction in eosinophilic gastric infiltration ⁽¹²⁾. Corticosteroids still represent the mainstay treatment for EoG and have been demonstrated efficacy in children and adults, though type, duration, and dosage vary, and no RCTs exist. A clinical remission has been demonstrated in 50% to 90% of patients with EGIDs receiving 20 to 40 mg corticosteroids per day for 2 weeks, although higher doses (0.5–1 mg/kg) are suggested in some reports ⁽²⁵⁾. Maintenance treatment with a low-dose corticosteroid may be required in patients with disease relapse during or after drug tapering ⁽²⁶⁾. Corticosteroid-sparing maintenance therapy, such as mesalazine, azathioprine, and antitumor necrosis factor a agents has been inconsistently reported, with variable efficacy. Other options include mast-cell stabilisers, such as sodium cromoglycate, the leukotriene receptor antagonist montelukast, the interleukin 5-blocking agent mepolizumab and omalizumab, an anti-IgE agent and anti-CCR3, an eotaxin receptor expressed by eosinophils ⁽²⁷⁾. Recently, a phase 2 trial of AK002 (lirentelimab), an anti-sialic acid-binding Ig-like lectin-8 (Siglec-8) antibody that depletes eosinophils and inhibits mast cells, showed a significant reduction of gastrointestinal eosinophils and symptomatic improvement in both treatment doses compared with placebo ⁽²⁸⁾.

Natural history and prognosis of primary EoG are still not well studied, especially in children. In a small cohort of adult patients, recurrence rate was high (∼60%) whereas a high blood eosinophil count was associated with relapse ⁽²²⁾. A systematic review of original studies, which included paediatric patients with EoG is presented in Table 2.

TABLE 2 - Table of evidence for eosinophilic gastritis

First author, Publication year	Population	Country	Study period	N	Design	Presentation	Treatment	Outcome
Tien et al, 2011 (23)	Children (0–15 years)	Taiwan	12y	14	Case series	Abdominal pain, anaemia, hypoalbuminemia	Corticosteroids (n = 1), montelukast (n = 4), Combination of corticosteroid/montelukast/ketotifen (n = 8), elimination diet (n = 1)	Only 2 of 9 on any corticosteroid combination had sustained corticosteroid-free clinical remission. Montelucast: 4/4 clinical response. Median FU of ly
Ko et al, 2014 (9)	Children (0–15 years)	USA	6y	30	Cohort study, longitudinal	Abdominal pain, vomiting, failure to thrive, anaemia, protein losing enteropathy, 43% concomitant eosinophilic esophagitis, 21% eosinophilic enteritis	Elimination diets (n = 17), PPIs (n = 2), cromolyn sodium (n = 2)	Elimination diets: 82% clinical and 78% histologic response. PPI: 1/2 histologic response. Cromolyn sodium: no histologic response. Median FU time not reported
Choi et al, 2015 (29)	Children (0–18 years)	Korea	21y	24	Cohort study, longitudinal	Abdominal pain, diarrhoea, anaemia, hypoalbuminemia	Systemic corticosteroids (n = 19), elimination diet (n = 9)	Corticosteroids: 58% sustained clinical response. Elimination diet: 55% clinical response. Median FU time not reported
Reed et al, 2015 (30)	Children (0–18 years), adults	USA	13y	33	Cohort study, longitudinal	Abdominal pain, nausea, vomiting diarrhoea, dysphagia. 45% (n = 18) with eosinophilic gastritis and 30% with associated duodenal eosinophilia	80% corticosteroids (mostly budesonide), 47% elimination diets	Out of 40 with a repeat endoscopy (median FU of 26 months): 60% clinical remission, 51% endoscopic resolution, 69% histologic resolution. More significant improvement with corticosteroids than elimination diets
Pesek et al, 2019 (14)	Children (0–18 years), adults	USA	lOy	317	Multicenter cohort study, longitudinal	Abdominal pain, nausea, vomiting, diarrhoea. 38% eosinophilic gastritis, 32% eosinophilic gastroenteritis, 29% eosinophilic colitis, 41% with concomitant eosinophilic esophagitis	Out of 142 patients with eosinophilic gastritis: 61% PPI, 23% topical steroids, 20% systemic steroids and 13% budesonide. 57% were treated with elimination diets	Out of 40 with a repeat endoscopy (FU of 6 months): 75% clinical improvement, 53% endoscopic improvement, 67% histologic improvement. More significant improvement with topical steroids and elimination diets
Votto et al, 2021 (11)	Children (0–18 years)	Italy	4y	5	Cohort study, longitudinal	Abdominal pain, nausea, vomiting	NR (for patients with gastric eosinophilia)	NR
Gurkan et al, 2021 (31)	Children (0–18 years)	Turkey	lOy	17	Cohort study, longitudinal	Abdominal pain, vomiting	PPIs alone (n = 7), elimination diets (n = 6), systemic corticosteroids (n = 5)	Response to treatment (clinical and histologic) of 50% (4/8 patients). Median FU of 3.3 years
Yamamoto et al, 2021 (32)	Children (0–18 years), adults	Japan	4y	25	Cohort study, longitudinal	Abdominal pain, vomiting	PPIs (42%), systemic corticosteroids (40%), elimination diets (19%)	Comparable symptom resolution rates between corticosteroids and elimination diets (52% and 53%, respectively). FU time not reported.
Higuchi et al, 2021 (33)	Children (6–14 years)	Japan	8y	4	Case series	Abdominal pain, diarrhoea	Elemental diet therapy with following elimination diet	Clinical remission in all 4 patients. One patient relapsed upon food introduction. Median FU of up to 2 years.

FU = follow-up, GI = gastrointestinal, N = number of children with gastric involvement in the study, NA = not available, NR = not reported, PPI = proton pump inhibitor.

Autoimmune Atrophic Gastritis

H pylori-negative atrophic gastritis is characterised by atrophy of the gastric mucosa, which is caused by an autoimmune mechanism, mainly affecting the gastric corpus ⁽³⁴⁾.

Autoimmune gastritis (AIG) is defined as an inflammatory disease, mainly affecting the corpus and the fundus of the stomach, characterised by atrophy of the gastric mucosa induced by auto-antibodies against parietal cells or intrinsic factor ⁽³⁵⁾. AIG usually develops in elderly adults and is considered a rare condition in children. Several cases are, however, reported also in the paediatric population especially associated with refractory iron-deficiency anaemia (IDA) or autoimmune disorders or immune dysregulation ^(36–41).

The aetiology of AIG is unknown. It is characterised by the presence of autoantibodies against the proton pump H+/K+ adenosine triphosphatase of the parietal cells of the stomach (APCA) or, less frequently, against the intrinsic factor, which induces a chronic inflammation that leads to gastric cells destruction up to mucosal atrophy ⁽³⁵⁾.

As a consequence of gastric mucosal atrophy, patients develop hypochlorhydria with hypergastrinaemia and a reduced secretion of intrinsic factor that may lead to vitamin B12 malabsorption and pernicious anaemia ^(42,43).

Hydrochloric acid, in fact, is essential for iron absorption and its absence may result in iron malabsorption and IDA ⁽³⁶⁾. Interestingly, refractory IDA in adults can precede the onset of pernicious anaemia by about 20 years and may be, therefore, prevalent in the paediatric population ^(44,45).

AIG needs a high index of clinical suspicion as it can be asymptomatic. Upper endoscopy with biopsies obtained from the gastric body and antrum/incisura (at minimum) is the gold standard for diagnosis ⁽⁴⁶⁾. The classic endoscopic features of AIG include pale-appearing gastric mucosa, loss of rugal folds, increased visibility of submucosal blood vessels because of the thinning of gastric mucosa, and frequently, the presence of a visible atrophic border. In the early stages, the endoscopic appearance of the stomach may be normal ⁽⁴⁷⁾. The Kimura-Takemoto classification categorises AIG in adults as closed (C) or open (O) type, based on the extent of the atrophic border. In the closed-type 1 (C-1), atrophy is limited to the antrum, in the C-2 type, atrophy is extended to the lesser curvature of the distal gastric body, and in the C-3 type, atrophy is present in the antrum and lesser curvature of the proximal gastric body. In the open-type (0-1 to 0-3), the atrophic border further spreads from the anterior wall (O-1) to the greater curvature (O-3) ⁽⁴⁸⁾.

APCA and/or intrinsic factor antibodies, which more often appear positive later in course of the disease can also precede the clinical presentation of AIG in patients with other autoimmune disorders ^(43,49). APCA is the most sensitive serum biomarker for AIG but may be falsely positive in H pylori infection and other autoimmune diseases ^(39,50). A recent meta-analysis revealed that the combination of pepsinogen, gastrin-17, and anti-H pylori antibody serological assays appears to be a reliable tool for the diagnosis of atrophic gastritis in adults ⁽⁵¹⁾.

Currently, only a few case reports and series have been published for AIG in children ^{(35,37,52–57)}. Therefore, no treatment recommendations exist for paediatric patients with AIG, regardless of the absence or presence of metaplastic changes, while established guidelines for surveillance, staging, and endoscopic management are available only for adults ⁽⁴⁸⁾. In adults, in contrast to children, most cases of atrophic gastritis are related to H pylori infection, and as such, patients with AIG (including children) should always be tested for H pylori, and if positive, H pylori should be eradicated with appropriate confirmation ⁽⁵⁸⁾ as eradication was shown to be sufficient to restore normal gastric mucosa or improve histological grading ^(59,60). Two other meta-analyses in adults of H pylori eradication demonstrated significant risk reduction for gastric cancer but not chronic atrophic gastritis ^(61,62).

In all AIG patients, treatment should be targeted to manage potential associated micronutrient deficiencies. There should be a low threshold for evaluation of other possible associated autoimmune disorders based on clinical presentation. Adult patients with H pylori-related metaplastic atrophic gastritis are at risk of developing gastric dysplasia and subsequent adenocarcinoma. Despite the fact that autoimmune metaplastic atrophic gastritis (AMAG) is not considered a part of the metaplasia-dysplasia-carcinoma cascade, the incidence of gastric carcinoma is higher in patients with AMAG compared with the general population ⁽⁶³⁾. AMAG is associated with increased risk of the development of type 1 gastric neuroendocrine tumor (NET), hence various strategies have been proposed in adult studies such somatostatin as the use of netazepide ⁽⁶⁴⁾ to reduce the elevated circulating gastrin.

There is no data available on the outcome of children with paediatric-onset AIG.

A systematic review of original studies, which included paediatric patients with autoimmune/atrophic gastritis is presented in Table 3.

TABLE 3 - Table of evidence of autoimmune/atrophic gastritis

First author, publication year	Country	Population	Study period	N	Design	Presentation	Treatment	Outcome
Nardone et al, 2001 (65)	Italy	Children (3–15 years old)	ly	5	Cohort, cross-sectional	Dyspepsia, abdominal pain	NA	NA
Segni et al, 2004 (47)	Italy	Children (0–18 years old)	NR	5	Cohort, cross-sectional	Autoimmune thyroid disease with positive APCA	NA	All children with atrophic gastritis had active H pylori infection
Ricuarte et al, 2005 (48)	Korea and Colombia	Children (4–18 years)	NR	16	Cohort, cross-sectional	Chronic abdominal pain	NA	All patients with atrophic gastritis had active H pylori infection
Greenwood et al, 2008 (52)	Australia	Children (8 and 9 years old)	NR	2	Case reports	Addison and autoimmune thyroid disease with anaemia	Cobalamin supplement	Improvement in intestinal permeability in both patients
Miguel et al, 2014 (31)	Portugal	Children (4–18 years)	3y	8	Case series	Iron deficiency anaemia in AIG	H pylori eradication in 4 patients	No improvement in histologic appearance in all 8 patients
Gonçalves et al, 2014 (30)	Portugal	Children (7–16 years)	6y	5	Case series	Iron deficiency anaemia and dyspepsia	NA	NA
Besançon et al, 2017 (66)	France	Children (11 and 13 years old)	8y	2	Case reports	Epigastric pain and anaemia in 2 children with type 1 diabetes	PPIs and iron supplement	NR
Saglietti et al, 2018 (49)	Swiss	Children (12 and 17 years old)	7 and 15 months	2	Case reports	Iron deficiency anaemia	H pylori eradication in 1 patient	No histologic improvement following H pylori eradication
Mitsinikos et al, 2020 (32)	USA	Children (10, 15 and 17 years old)	NR	3	Case reports	Iron deficiency anaemia in 2 and megaloblastic anaema in 1	Iron and B12 supplement	Restoration of iron and B12 deficiencies
Demir et al, 2020 (67)	Turkey	Children (10–18 years old)	NR	10	Cohort, cross-sectional	Positive APCA and anaemia	NA	NA
Kulaketal, 2021 (37)	USA	Children (3 months to 18 years old)	9y	22	Cohort, cross-sectional	Abdominal pain, anaemia,	NA	60% had other autoimmune disorders, 50% had anti-parietal cell antibodies

APCA = antiparietal cell antibody, N = number of children with autoimmune/atrophic gastritis in the study, NA = not applicable, NR = not reported, PPI = proton pump inhibitor.

Collagenous Gastritis

Collagenous gastritis (CG) is a rare form of chronic gastritis first described in 1989 by Colletti and Trainer ⁽⁶⁸⁾. The childhood-onset type has an incidence rate of 0.25/100,000 person-years, anda female-male ratio of 4.2 ⁽⁶⁹⁾. The prevalence of CG has been described as 2.1/100,000 in children below 18 years ⁽⁶⁹⁾. Regarding its aetiology, collagen deposition may be a reparative process in response to an inflammatory, infectious or toxic damaging process ⁽⁷⁰⁾, and autoimmune abnormalities, medication effects (such as angiotensin receptor blockers), and infection have been theorised ⁽⁷¹⁾. CG was shown to be associated with other lymphocytic disorders of the upper gastrointestinal tract, including coeliac disease ⁽⁷²⁾. There are 2 age peaks of CG, namely 10 to 19 years and > 60 years ⁽⁷²⁾, although the youngest patient described was 1 year old ⁽⁷³⁾. Paediatric patients mostly present with isolated CG, although in a minority of paediatric patients, concurrent collage-nous colitis/enteritis is found ^(69,74,75).

Affected children typically present with severe IDA (often long-standing) and abdominal pain ⁽⁷⁶⁾ but other symptoms, such as nausea, recurrent vomiting, weight loss, and diarrhoea have been described ⁽⁶⁹⁾. Anaemia is most likely attributed to damage of dilated capillaries entrapped in the subepithelial collagen band ^(70,76). All patients with chronic diarrhoea were found to have the adult-onset phenotype with coexisting collagenous colitis. Perforation of the stomach is a rare complication of CG ⁽⁷⁷⁾.

Although a family history of autoimmune disorders and auto-antibodies are commonly detected in children with CG, early development of autoimmune comorbidities has rarely been observed. The most common autoantibody detected is anti-nuclear antibody (∼20% of patients) ⁽⁶⁹⁾.

The most characteristic endoscopic finding is a diffuse gastric mucosal nodularity that can spare the antrum. Other reported findings range from normal mucosa to diffuse erythema, haemorrhage, erosions, ulcerations, and exudates ^{(69,71,73–76)}. By definition, histopathology revealed a sub-epithelial collagenous deposition of ≥10 μm and an inflammatory infiltrate of the lamina propria ⁽⁶⁸⁾. The collagen deposition is patchy, with a mean thickness of 77 μm (13–184) ⁽⁷¹⁾. It appears at haematoxylin-eosin staining but it can be enhanced by Masson staining. The inflammatory infiltrate (which decreases in areas of collagen deposits is composed of lymphocytes, plasma cells, and eosinophils, and the number of intraepithelial lymphocytes is ≤25/100 surface epithelial cells ⁽⁷¹⁾. It is important to take biopsies from and around the nodules ⁽⁷⁶⁾.

CG is a rare condition and no therapeutic guidelines exist. Medications used are mainly proton pump inhibitors (PPIs), sucral-fate, and budesonide but even when symptoms improve, histologic damage usually remains ^(72,74,78).

A systematic review of original studies, which included paediatric patients with eosinophilic gastritis is presented in Table 4.

TABLE 4 - Table of evidence of collagenous gastritis

First author, publication year	Country	Population	Study period	N	Design	Presentation	Treatment	Outcome
Colletti et al, 1989 (60)	USA	Child (15 years old)	ly	1	Case report	Recurrent abdominal pain, acute upper GI bleeding	H2 blockers, sucralfate, furazolidine	No clinical improvement, no change in histologic features (ly FU)
Lagorce-Pages et al, 2001 (79)	France	Child (11 years old), adults	8y	1	Case series	Anaemia	NA	Histologic improvement (8y FU)
Ravikumara et al, 2007 (62)	United Kingdom	Child (9 years old)	4y	1	Case report	Epigastric pain, vomiting, diarrhoea, iron deficiency anaemia	PPI, iron	Clinical remission but no change in histologic features (50 months FU)
Leung et al, 2009 (70)	USA	Children (13,14 and 15 years old), adults	14y	3	Case series	Anaemia, nausea, vomiting, epigastric pain	Budesonide (n = l), PPI (n = 1), sucralfate (n = l)	Clinical improvement on budesonide and sucralfate, endoscopic FU on budesonide (n = 1) with no change in histologic features
Camarero Salces et al, 2011 (80)	Spain	Child (9 years old)	ly	1	Case report	Previously diagnosed collagenous colitis; epigastric pain, anaemia	NA	NA
Ma et al, 2015 (63)	USA	Children (3–18 years old), adults	23y	10	Cohort study, longitudinal,	Iron deficiency anaemia, epigastric pain, diarrhoea,	Budesonide (n = 2), sucralfate (n = 2), Gluten-free diet (n = 1)	Resolution of symptoms (budesonide), no change in histologic features (all, n = 5), median FU of 17 months
Hijaz et al, 2013 (67)	USA	Children (7 and 11 years old)	3y	2	Case reports	iron deficiency anaemia, dyspepsia	PPIs and budesonide for both	clinical improvement, no change in histologic features (1-year FU)
Bajwa et al, 2015 (81)	USA	Child (13 years old)	3y	1	Case report	Diffuse abdominal pain, diarrhoea, weight loss	Gluten free diet	Clinical remission, histologic improvement (3-year clinical FU, 6 months endoscopic FU)
Koide et al, 2015 (82)	Japan	Child	6y	1	Case report	Hematemesis	H2 receptor antagonist	Clinical improvement, No change in histologic features (6-year FU)
Appelman et al, 2016 (69)	The Netherlands	Child (15 years old)	iy	1	Case report	Acute epigastric pain, iron deficiency anaemia, gastric perforation,	Surgery, iron supplement,	No change in histologic features (1-year FU)
Matta et al, 2018 (66)	Australia, Lebanon,	Children (5–18 years old)	19y	12	Case series	Abdominal pain, iron deficiency anaemia, diarrhoea, vomiting, 3 (25%) with associated collagenous colitis	PPIs (n = 4), dietary modulation (gluten free diet, elimination diet, n = 4) steroids, azathioprine (n = 1)	Resolution of iron deficiency anemia in all (n = 9), histological improvement (steroids and azathrioprine (n = 1), no change in histologic features (n = 7), median FU of 30 months
Lee et al, 2019 (68)	Korea	Children (8,11 and 15 years old)	2019	3	Case reports	Vomiting, iron deficiency anaemia, abdominal pain	PPI (n = 2), corticosteroids (n = 1), iron supplement	Clinical improvement in all 3 patients. FU time not reported
Eke et al, 2019 (83)	South Africa	Child (15 months old)	iy	1	Case report	Oedema, diarrhoea, anaemia	PPI, iron, then corticosteroid, azathioprine	Clinical remission, no endoscopic re-evaluation (8 months FU)
Lee et al, 2019 (76)	Korea	Children (8,11, and 15 years old)	NR	3	Case reports	Anaemia, vomiting	Iron, PPI and corticosteroids (n = 1)	Clinical remission, no endoscopic re-evaluation
Käppi et al, 2020 (61)	Sweden	Children (0–18 years old)	lly	15	Cohort study, longitudinal	Iron deficiency anaemia, recurrent abdominal pain, diarrhoea	PPI (n = 5), gluten-free diet (n = 2), elimination diets (n = 12), budesonide (n = 1)	Normalization of iron status, no change in histologic features (n = 11/11), median FU of 4.4 years
De Ronde et al, 2020 (84)	Belgium	Children (10 and 15 years old)	iy	2	Case reports	Anaemia, epigastric pain	PPI, iron supplement for both	Resolution of iron deficiency anaemia, No change in histologic features under PPIs (n = l)

FU = follow-up, GI = gastrointestinal, N = number of children with collagenous gastritis in the study, NA = not available, NR = not reported, PPI = proton pump inhibitor.

Gastritis in Inflammatory Bowel Disease

Although upper gastrointestinal tract involvement has been classically considered as a variable, which differentiates between Crohn disease (CD) and ulcerative colitis (UC) (limited to the colon), different studies have reported gastric involvement in both entities. Initial work up for IBD in children ⁽⁸⁵⁾ should include upper endoscopy as part of the recommended algorithm. Apart from the presence of granulomata in the stomach, which is highly suggestive of CD and reported in 20% of children with this entity ⁽⁸⁶⁾, other microscopic findings can be observed both in CD and UC. The commonest of these upper findings is chronic active gastritis, also termed focally enhanced gastritis (FEG). FEG has been defined as the presence in biopsy material of 1 or more foveola/glands surrounded and infiltrated by inflammatory cells that consist mainly of lymphocytes and monocytes, thus, in the absence of other findings along the gastrointestinal tract may overlap with lymphocytic gastritis ⁽⁸⁷⁾. The area may also contain plasma cells, eosinophils, and neutrophils and the surrounding mucosa may be relatively normal. FEG or any other form of chronic active gastritis was 6 to 10 times more common in patients with IBD than in non-IBD controls ⁽⁸⁷⁾. The presence ofFEG (predominantly involving the antrum and body) has been associated with IBD in paediatric patients with a higher rate in CD, oscillating between 54% and 68% but is present also in 16% to 21% of UC, and hence is generally a positive predictor for CD ^(88–90). All the studies comparing FEG in IBD patients to non-IBD patients, show statistically significant differences, with around 3.4% in the non-IBD group ⁽⁹¹⁾. FEG is considered to have a sensitivity of 36% to 41% and specificity of 94% to 97% in distinguishing between paediatric IBD from non-IBD patients ^(89,91). Therefore, the finding of FEG in gastric biopsy specimens in children is very suggestive of IBD. Multifocal involvement is common and the total number of glands involved is usually higher in UC than in CD (6.4 ± 5.1 vs 4.0 ± 3.0 glands, respectively) ⁽⁹²⁾. Clinical manifestations in children with FEG ranges from no symptoms to epigastric and generalised abdominal pain, nausea, vomiting, and weight loss ^(90,93,94). Abnormal endoscopic findings (including aphthoid ulcers, granularity, and submucosal bleeding) are commonly identified in patients with FEG (49%–68.7%) ^(92,93). Patients with CD and FEG are more likely to have active ileitis and granulomata elsewhere in the gastrointestinal tract. Meanwhile, there is no correlation reported between FEG and other gastrointestinal findings in UC ⁽⁹²⁾.

Other nonspecific gastric findings can also be observed in children with IBD including mild chronic (nonfocal) gastritis in up to 50% ⁽⁸⁸⁾. Whether this finding is a manifestation of IBD per se or because of perhaps an altered immune response is unclear ⁽⁸⁷⁾.

The prognostic value of gastric findings in IBD are minimal to date, although 1 recent study suggests that gastropathy at initial IBD diagnosis may be associated with an increased risk of postoperative recurrence in paediatric CD ⁽⁹⁵⁾.

The literature lacks controlled studies evaluating the effects of drugs available for the treatment of upper gastrointestinal CD, thus therapy is based on the concomitant activity of the distal disease. PPIs relieve symptoms and are commonly prescribed for the short-term but have not shown to have an effect on the chronic inflammatory process ⁽⁹⁶⁾. Anti-tumor necrosis factor a agents are effective for the treatment of gastroduodenal CD with a reported rate of 73% mucosal healing in a small cohort of adult patients ⁽⁹⁷⁾.

A systematic review of original studies, which included paediatric patients with focally enhanced gastritis is presented in Table 1, Supplementary Digital Content, https://links.lww.com/MPG/C693.

Lymphocytic Gastritis

Lymphocytic gastritis (LG) is an unusual form of gastritis characterised by ≥25 intraepithelial lymphocytes per 100 consecutive epithelial cells, associated with a variable chronic inflammatory pattern in the lamina propria ⁽²⁹⁾. LG has been found in association with a variety of disorders, such as coeliac disease, H pylori gastritis, HIV, Meanetrier's disease, NSAIDs, and also with angiotensin II receptor blockers suggesting an abnormal local immunologic response to different antigens as the underlying pathophysiologic mechanism ^(98,99). A recent large retrospective study in adults has, however, questioned the pathogenic role of H pylori infection ⁽¹⁰⁰⁾. The prevalence of LG in children with coeliac disease ranges between 10% and 45% ^(101–105). The pattern of involvement of the gastric mucosa might be predictive of duodenal villous atrophy, as an antrum-predominant or diffuse LG is associated with more severe duodenal villous atrophy ⁽¹⁰⁶⁾. Children with LG at the time of diagnosis of coeliac disease are significantly younger, have a more severe disease and greater neutrophilic infiltration of the duodenum compared with those without gastritis ^(107,108). A systematic review of original studies, which included paediatric patients with lymphocytic gastritis is presented in Table 2, Supplementary Digital Content, http://links.lww. com/MPG/C694.

Other Causes of Gastritis

After excluding aetiologies such as eosinophilic, collage-nous, autoimmune, focally enhanced gastritis and lymphocytic gastritis, there are still children with histological diagnosis of idiopathic H pylori-negative gastritis. Elitsur ⁽¹⁰⁹⁾ reported a higher prevalence of H pylori-negative gastritis in children (∼50%) compared with adults (∼20%). In children, idiopathic gastritis has no explanation, and its long-term prognosis is not known ⁽¹⁰⁹⁾. In contrast, in adults, this finding may be more commonly attributed to environmental risk factors (such as alcohol and smoking), or common drugs (NSAIDs, PPIs) ⁽¹¹⁰⁾. It was hypothesised that the gastritis may be related to microorganisms located in the stomach other than H pylori⁽¹¹¹⁾. A subsequent study in adults found that the prevalence of idiopathic gastritis was 10.6% with no association with gender, age, or geographic distribution, and that intestinal metaplasia was more common in patients with H pylori gastritis (13% vs 6.1%) ⁽¹¹²⁾. A relatively recent paediatric study demonstrated that 60.7% of children with chronic gastritis were H pylori-negative (after excluding all other known causes). These children were younger, and were less likely to present with abdominal pain. Nodular gastritis was found in 6.8% of children with H pylori-negative gastritis compared with H pylori-positive gastritis (35.4%). The inflammatory infiltrate was less severe in H pylori-negative gastritis but in 51.3% of these patients, lymphoid follicles were present ⁽¹¹³⁾. In a follow-up study carried by the same group, out of 8 children with a mean follow-up of 25.6months, 3 had chronic gastritis with lymphoid follicles, 4 had resolution of the gastritis but still had follicles, and 1 patient had resolution of both findings ⁽¹¹⁴⁾. Still, the clinical relevance and prognosis of this entity are unknown and warrant further research.

Gastropathies are commonly found in both adults and children with chronic renal failure (CRF) or end-stage renal disease (also termed uraemic gastritis) and in the post-renal transplant period ^(115–118). The mechanism resulting in gastritis involves elevated gastrin level, increased production of gastric acid, delayed gastric emptying, complex disorder of gastrointestinal motility, and uraemic toxicity ⁽¹¹⁹⁾. In children with end-stage kidney disease, the prevalence of gastro-duodenal lesions ranges between 46% and 81%, irrespective of the symptomatology and type of dialysis (peritoneal dialysis or haemodialysis) ^(115–119), suggesting the importance of an endoscopic examination in all transplant candidates. The most frequent mucosal lesions are diffuse oedema and erythema, with or without erosions (detected in ∼1/3 of patients). Postrenal transplant gastritis is commonly associated with immu-nosuppressive medications, such as tacrolimus, mycophenolate mofetil, and corticosteroids ⁽¹²⁰⁾.

CONCLUSIONS

Similar to adults, infection with H pylori represents the main aetiologic factor for the development of chronic inflammation of the gastric mucosa in children. Nevertheless, in the absence of H pylori infection or following its successful eradication, the presence of gastritis may be attributed to a variety of othercauses. Among these causes, special attention should be given to treatable entities, such as eosinophilic, autoimmune, and collagenous gastritis. Findings compatible with focally enhanced or lymphocytic gastritis should raise suspicion of IBD and coeliac disease, respectively. Data on potential under-detection rate, long-term outcome, and effective treatment inthese pathologies is lacking. Corticosteroid therapy and various elimination diets may be effective in children with eosino-philic gastritis but there is very little evidence of effective therapies in autoimmune or collagenous gastritis.

This systematic review is limited by the lack of specific research question because of the scarcity of high-quality studies or even cohort studies with large sample size. Therefore, we sought to present the entire available data on the selected topics shedding more light on this poorly described entity. As ever, further collaborative research is necessary in this area, in order to inform for future effective diagnosis and therapy.