Introduction

Helicobacter pylori (HP), a Gram-negative bacterium, remains a pivotal focus in clinical gastroenterology, with well-established links to peptic ulcer disease, including gastric and duodenal ulcers1. Our prior research validated the safety and efficacy of four HP eradication regimens and characterized long-term gut microbiota shifts2,3, which has advanced clinical management and solidified HP eradication as a cornerstone of therapy for HP-associated diseases.

Eosinophilic gastrointestinal diseases (EGIDs) encompass a heterogeneous group of disorders defined by abnormal eosinophil (EOS) infiltration in the gastrointestinal tract, which triggers inflammatory cascades that manifest as symptoms ranging from mild discomfort to severe complications (e.g., vomiting, abdominal pain, anemia, and growth retardation)—a condition particularly prevalent in pediatric populations4,5.

Emerging evidence reveals a complex interplay between HP and EGIDs. Epidemiological studies consistently demonstrate an inverse association: lower HP prevalence in patients with eosinophilic esophagitis (EoE) or eosinophilic gastritis (EGE) compared to healthy controls, suggesting a potential protective role of HP in mitigating chronic immune-mediated diseases6,7,8. This “protective effect” is hypothesized to stem from the ability of HP to shift mucosal immunity toward a Th1-dominated response, thereby suppressing Th2-mediated pathways that are central to the pathogenesis of EGIDs.

Conversely, a growing body of reports links successful HP eradication to the de novo onset or exacerbation of eosinophilic conditions, including refractory eosinophilic duodenal bulb ulcers (REDU)9,10. This paradox—where eradication of a pathogen (HP) linked to ulcer disease precipitates a distinct, immune-mediated ulcer phenotype—raises critical questions regarding the long-term consequences of HP clearance, particularly in children with developing immune systems. Despite their rarity, these REDU cases pose significant diagnostic and therapeutic challenges, with their pathophysiology remaining poorly understood—a knowledge gap amplified by the paucity of pediatric data in this field.

To date, research on the post-eradication development of REDU (distinct from HP-associated DBU) remains limited, particularly in pediatric cohorts. This multicenter study aims to address this gap by characterizing the clinical features, endoscopic findings, pathological mechanisms, and management strategies of this understudied condition, with implications for optimizing clinical practice in pediatric gastroenterology.

Methods

Data sources

This study was approved by the Ethics Committee of the Children’s Hospital of Fudan University (Shanghai, China; IRB No.: 2024 (248)). All methods were performed in accordance with the relevant guidelines and regulations. Written informed consent was obtained from the parents of all participants prior to study initiation. Pediatric patients were recruited retrospectively between January 1, 2016, and August 31, 2025, and their electronic medical records were reviewed to assess eligibility.

The inclusion and exclusion criteria

Inclusion criteria:

① Endoscopic confirmation of duodenal bulb ulcer (DBU), a subtype of peptic ulcer, with explicit documentation of ulcer location, size, number, and recurrence interval.

② Pathological evidence of EOS counts ≥ 50 cells per high-power field (HPF) in duodenal bulb biopsy specimens4.

③ Documentation of HP testing using validated methods (e.g., 13C urea breath test [13C UBT], rapid urease test [RUT], histopathological examination); standardized HP eradication therapy was administered to all HP-positive patients.

Exclusion criteria: Patients were excluded if they had a confirmed diagnosis of intestinal tuberculosis, inflammatory bowel disease, parasitic infections, eosinophilic leukemia, primary immunodeficiencies, non-steroidal anti-inflammatory drug (NSAID)-induced mucosal reactions, or other confounding comorbidities.

Definition of REDU: REDU were defined as lesions that either persisted on the first follow-up endoscopy or initially resolved but subsequently relapsed during the follow-up period11.

These REDU cases were divided into 2 subgroups based on their HP infection status:​

Group A: HP eradication-associated REDU group

Definition: Patients with confirmed Hp infection (e.g., positive 13C UBT, positive RUT, or positive histopathological examination), who developed DBU within 8 weeks of completing standardized HP eradication therapy and fulfilled the aforementioned diagnostic criteria for REDU.

Group B: Idiopathic REDU group

Definition: Patients with no confirmed evidence of HP infection (all HP detection assays yielded negative results), no other identifiable etiological factors for ulceration (e.g., non-NSAID exposure, autoimmune disorders), and who met the diagnostic criteria for REDU as defined above.

Group C: Control group.

Thirty-eight age-matched children with newly diagnosed HP-associated DBU, who had received standardized HP eradication therapy and achieved complete ulcer healing, were randomly enrolled during the same study period. These patients were matched to Group A by age (± 2 years), gender, and HP eradication regimen to ensure identical exposure to HP eradication therapy, but with a divergent clinical outcome (i.e., absence of post-eradication ulcer recurrence). Notably, routine duodenal biopsy was not performed in Group C.

Laboratory and genetic assessments

During episodes of recurrent/refractory DBU, the following parameters were evaluated: peripheral blood EOS count, erythrocyte sedimentation rate (ESR), serum albumin level, serum specific IgE level, immunoglobulin G4 (IgG4) level, and humoral/cellular immunity profiles. For the REDU group, thirty-four serum cytokines were measured using a Luminex 200 instrument (Millipore) with a ThermoFisher ProcartaPlex multiplex immunoassay kit (cat. no.: EPXR340-12167-901), including macrophage inflammatory protein-1α (MIP-1α), MIP-1β, stromal cell-derived factor-1α (SDF-1α), interleukin-27 (IL-27), IL-1β, IL-2, IL-4, IL-5, interferon gamma-induced protein 10 (IP-10), IL-6, IL-7, IL-8, IL-10, Eotaxin, IL-12p70, IL-13, IL-17 A, IL-31, IL-1 receptor antagonist (IL-1RA), CCL5/regulated on activation, normal T cell expressed and secreted (RANTES), interferon-γ (IFN-γ), granulocyte–macrophage colony-stimulating factor (GM-CSF), tumor necrosis factor-α (TNF-α), TNF-β, IFN-α, monocyte chemoattractant protein-1 (MCP-1), IL-9, growth-related oncogene alpha (GRO-α), IL-1α, IL-23, IL-15, IL-18, IL-21, and IL-22. Cytochrome P450 family 2 subfamily C member 19 (CYP2C19) gene polymorphisms were analyzed via polymerase chain reaction (PCR). The metabolic phenotypes of CYP2C19 included extensive metabolizers (EM), intermediate metabolizers (IM), poor metabolizers (PM), rapid metabolizers (RM), and ultrarapid metabolizers (UM).

Endoscopic and histopathological evaluation

Upper gastrointestinal endoscopy (gastroscopy) and lower gastrointestinal endoscopy (colonoscopy) were performed. Biopsies of the duodenal bulb and gastric antral mucosa were collected for pathological examination and EOS counting. Colonoscopy with mucosal biopsies was conducted to exclude secondary duodenal ulcers caused by systemic diseases. Duodenal stenosis was diagnosed based on objective endoscopic findings: either the inability of a standard-diameter endoscope (outer diameter: 9.8 mm, consistent with adult upper gastrointestinal endoscopy specifications) to pass smoothly through the duodenal bulb without endoscopic dilation, or the identification of a stenotic ring. As described in previous studies, DBU with an endoscopic diameter of ≥ 2 cm were defined as giant DBU12.

Statistical analysis

Continuous variables were expressed as mean ± standard deviation (SD) or median (interquartile range [IQR]) depending on their distribution, while categorical variables were expressed as frequencies (percentages). Differences between the HP eradication-associated REDU group (Group A) and the HP eradication-control group (Group C) were analyzed using the chi-square (χ2) test for categorical variables and the independent samples t-test and the Mann–Whitney test for continuous variables. Binary logistic regression was employed to explore the risk factors of REDU.Statistical significance was set at P < 0.05. All statistical analyses were performed using SPSS software version 26.0 (IBM Corp., Armonk, NY, USA).

Results

Patient baseline characteristics

Of 53 initially reviewed cases of suspected REDU, 30 were enrolled after applying the exclusion criteria (23 cases were excluded: 12 with insufficient pathological data, 9 with subthreshold EOS counts [< 50 cells/HPF], and 2 with comorbidities [Crohn’s disease and Henoch-Schönlein purpura]). A total of 30 enrolled patients with REDU were divided into two subgroups: Group A (HP eradication-associated REDU, 19 cases) and Group B (idiopathic REDU, 11 cases). For group A, although patients received standardized HP eradication therapy, the HP-induced ulcers initially persistently failed to heal or recurred thereafter. An additional 38 children with HP-induced DBU who had achieved successful cure after HP eradication therapy were randomly selected as the control group (Group C).

The REDU cohort (n = 30) comprised 27 males and 3 females (male-to-female ratio: 9:1; mean age: 10.0 ± 3.4 years; age range: 5–15 years). Symptom duration varied from 1 day to 6 years, and abdominal pain was the most prevalent primary presenting symptom, with a median duration of 24.0 [IQR:7.5, 30.0] months (Table 1). Nineteen patients (63.3% of the REDU cohort) presented with anemia, among whom three had severe anemia (with the lowest hemoglobin level of 57 g/L) accompanied by melena, requiring blood transfusion. REDU were the primary reason for repeated hospital visits.

Table 1 The symptom for REDU patients first visit hospital.
Full size table

For the REDU cohort, the median body mass index (BMI, kg/m2) was 15.6, with an IQR of [15.0, 17.0]. Malnutrition was prevalent in the REDU cohort: 10 patients (33.3%) had a BMI z-score between − 1 and − 2 SD, 6 patients (20.0%) had a BMI z-score between − 2 and − 3 SD, and 1 patient (3.3%) had a BMI z-score < -3 SD.

Comparison of characteristics between HP eradication-associated REDU (group A) and HP-associated DBU control (group C)

The characteristics of the HP eradication-associated REDU group (Group A) and the HP-associated DBU control group (Group C) are compared in Table 2. Abdominal pain was also the most common chief complaint in both groups. In Group A, the time interval from HP eradication therapy completion to REDU onset was 14.7 ± 8.8 months, and the interval between initial HP infection diagnosis and subsequent ¹³C UBT re-examination showing negative results was 9.3 ± 7.0 months. For Group C, the median duration of the last follow-up was 3.0 (IQR:2.0, 11.5) months. Following HP eradication and endoscopic re-evaluation, no gastrointestinal discomfort was reported in Group C; therefore, electronic medical record data regarding further gastroenterological follow-up were unavailable.

Table 2 Comparison of characteristics between HP- eradication-associated REDU group and HP -associated DBUcontrol group.
Full size table

After gastroscopic diagnosis of HP-induced DBU, patients in Group A and Group C immediately received eradication therapy. Compared with the Group C, Group A had a significantly higher incidence of anemia, giant DBU, and elevated peripheral blood EOS counts (all P < 0.01). Additionally, Group A had a higher need for repeated HP eradication therapy than Group C (42.1% vs. 2.6%, P < 0.01). The number of endoscopic examinations was also compared between the two groups. Compared with Group C, Group A required a significantly higher number of endoscopic examinations (7.8 ± 4.3 vs. 2.6 ± 1.3, P < 0.001). For Group A, the median number of HP eradication courses per patient was 1.0 [IQR: 1.0, 3.0] (Table 2).

Binary logistic regression analysis revealed that an elevated peripheral blood EOS percentage (P = 0.008; odds ratio [OR] = 0.08, 95% confidence interval [CI]: 0.008–0.289) and the number of endoscopic examinations (P = 0.019; OR = 0.428, 95% CI: 0.210–0.869) were significant factors influencing the cure outcome of REDU.

HP infection was confirmed in 8 patients via serum HP antibody testing, 9 patients via RUT, 3 patients via pathological examination of gastric mucosa biopsies, and 4 patients via ¹³C UBT. The HP eradication regimens administered to Group A patients included amoxicillin-based triple therapy (6 patients), bismuth-based quadruple therapy (16 patients), and high-dose amoxicillin-proton pump inhibitor (PPI) dual therapy (1 patient).

Laboratory findings

At the time of HP diagnosis, the peripheral blood EOS count in Group A was 460.0 ± 311.0 cells/µL (reference range: 60–300 cells/µL). Among the entire REDU cohort (n = 30), 19 patients (63.3%) had anemia, with a mean hemoglobin (Hb) level of 96.6 ± 25.6 g/L. When comparing Group A (HP eradication-associated REDU) and Group B (idiopathic REDU), no significant difference in mean Hb level was observed (Group A: 96.8 ± 26.9 g/L vs. Group B: 106.1 ± 23.0 g/L, P = 0.029). In the REDU cohort, the median serum IgE level was 393.0 kU/L (IQR: 61.7–655.8 kU/L; reference range: 0–300 kU/L), and the mean peripheral blood EOS count was 518.0 ± 349.0 cells/µL (reference range: 60–300 cells/µL). With respect to serum IgE levels and peripheral blood EOS counts, no significant differences were detected between Group A and Group B (all P > 0.05). In the REDU cohort, the mean erythrocyte ESR was 17.6 ± 14.0 mm/h (reference range: 0–21 mm/h), the mean serum IgG4 level was 1.64 ± 1.29 g/L (reference range: 0.012–1.699 g/L), the mean serum albumin level was 42.6 ± 3.3 g/L (reference range: 35–55 g/L), and the mean serum prealbumin level was 201.5 ± 41.8 mg/L (reference range: 200–400 mg/L). No significant differences in ESR, serum IgG4 level, serum albumin level, or serum prealbumin level were observed between Group A and Group B (all P > 0.05).

All 30 patients in the REDU cohort underwent specific allergen testing, among whom 28 (93.3%) tested positive. Of these 28 allergen-positive patients, 24 (85.7%) had food allergies—specifically, 15 (62.5% of food-allergic patients) were sensitized to cow’s milk protein, 13 (54.2%) to egg white, and 8 (33.3%) to egg yolk—and 11 (39.3% of allergen-positive patients) had dust mite sensitivity. For the HP-associated DBU control group, allergen information is lacking due to disease testing and medical ethics restrictions.

Serum cytokines were assayed in the REDU cohort(n = 19). The median serum IL-8 level was significantly elevated to 361.2 pg/mL (IQR: 92.8–613.4 pg/mL; reference range: 0.0–47 pg/mL). Nine patients (30.0% of the REDU cohort) had elevated serum levels of both MIP-1α (reference range: 0.0–21 pg/mL) and MIP-1β (reference range: 0.0–147 pg/mL), while five patients (16.7%) had elevated serum levels of both IL-10 (reference range: 0.0–3.0 pg/mL) and TNF-α (reference range: 0.0–14 pg/mL). Additionally, one patient (3.3%) had an elevated serum IL-4 level (84.8 pg/mL; reference range: 0–45 pg/mL), and another patient (3.3%) had an elevated serum IL-5 level (41.2 pg/mL; reference range: 0–17 pg/mL).

CYP2C19 gene polymorphisms were analyzed in 24 patients from the REDU cohort, and the metabolic phenotypes were as follows: 12 (50.0%) IM, 8 (34.8%) NM, 3 (13.0%) PM, and 1 (4.3%) RM.

Endoscopic and pathologic findings

Endoscopic findings of patients in the REDU cohort are summarized in Table 3; Fig. 1. Among the 30 REDU patients, 13 (43.3%) had ulcers located in the postbulbar duodenum, 12 (40.0%) in the anterior wall of the duodenal bulb, and 5 (16.7%) in the greater curvature of the duodenal bulb; three of these patients had “kissing ulcers”.

Table 3 The endoscopic findings in all REDU patients.
Full size table
Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Endoscopic images of DBU and description of treatment outcomes. (A) Image of the duodenal bulb ulcer (DBU) caused by prior HP infection in Case 2 of Group A; (B) Healing of DBU after HP eradication; (C,D) Recurrence of DBU following omeprazole dose reduction. (E) Image of the 3rd DBU recurrence in Group A Case 3 after omeprazole dose reduction; (F) Healing of DBU after 40 mg omeprazole treatment; (G) Recurrence of DBU after omeprazole dose reduction; (H) Healing of DBU under maintenance therapy with low-dose omeprazole combined with azathioprine; (I) Image of DBU caused by prior HP infection in Case 10 of Group A; (J) Scar formation after omeprazole administration; (K) Recurrence of DBU after omeprazole dose reduction; L. Healing of DBU under maintenance therapy with low-dose omeprazole combined with azathioprine; M-N. Image of the DBU in Group B and healing after treatment. O-P: Image of the DBU in Group C and healing after HP eradication.

Eosinophilic DBU (in both Groups A and B) were characterized by large, shallow, pan-like lesions covered with a clean, thin white exudate (Fig. 1C, D, E, G, K and M)—distinct from the prior HP-associated DBU cases in Group A (Fig. 1A and I) and Group C(Fig. 1O), both of which typically featured a dirty ulcer base .

Regarding giant DBU, 17 patients (89.5%) in Group A and 6 patients (54.5%) in Group B had this finding, with a statistically significant difference between the two groups (P < 0.001). Twenty-eight patients (93.3% of the REDU cohort) underwent colonoscopy, which revealed terminal ileitis in 13 cases (46.4% of colonoscopy recipients) and proctocolitis in 8 cases (28.6%). Fourteen patients (46.7% of the REDU cohort) underwent capsule endoscopy, with findings including enteritis (6 cases, 42.9% of capsule endoscopy recipients), duodenal ulcer (2 cases, 14.3%), normal findings (4 cases, 28.6%), and capsule impaction secondary to duodenal bulb stenosis (2 cases, 14.3%).

All patients in the REDU cohort underwent repeated gastroscopy due to persistent abdominal pain or for ulcer surveillance. Follow-up gastroscopy revealed that 21 patients (70.0%) had persistent DBU, 13 patients (43.3%) had duodenal bulb stenosis (9 cases in Group A), and 4 patients (13.3%) had duodenal mucosal erosion. Pathologically, pre-treatment duodenal bulb biopsies from the REDU cohort showed EOS counts of ≥ 100 cells/HPF in 17 patients (56.7%) and ≥ 50 cells/HPF in 13 patients (43.3%). Gastric antrum biopsies revealed EOS counts of ≥ 100 cells/HPF in 3 patients (10.0%) and ≥ 30 cells/HPF in 8 patients (26.7%). No statistically significant difference in EOS counts per HPF was observed between Group A and Group B, either in the duodenal bulb (P = 0.16) or in the gastric antrum (P = 0.70).

In the last follow-up, after post-treatment, duodenal EOS counts decreased significantly (pre-treatment: 100.0 [IQR: 60.0, 100.0]cells /HPF vs. post-treatment: 20.0 [IQR: 2.0, 88.0]cells /HPF; P < 0.01), while gastric antrum EOS counts showed no significant change (pre-treatment: 2.0 [IQR: 0.0, 52.0]cells /HPF vs. post-treatment: 0.0 [IQR: 0.0, 4.3]cells /HPF; P = 0.08). The typical pathological picture of these patients was shown in Fig. 2.

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Histological findings of duodenal bulb biopsy specimens from patients with DBU. (A) Case 2 of Group A: HE staining demonstrating HP-associated chronic inflammation (red oval indicates HP-like Gram-negative bacilli; magnification: 400×); (B) Case 2 of Group A: Pre-treatment duodenal bulb mucosa with marked EOS infiltration (≈ 90 /HPF; magnification: 200×); (C) Case 2 of Group A: Post-treatment duodenal bulb mucosa with reduced EOS count per HPF (200×); (D,E) Case 3 of Group A: Pre-treatment duodenal mucosa with severe EOS infiltration: D (≈ 100 /HPF) and E (≈ 50 /HPF); magnification: 200× for both; (F) Case 3 of Group A: Post-treatment duodenal mucosa (omeprazole 40 mg/day + azathioprine 1 mg/kg/day) with near-resolution of EOS infiltration (≈ 2 /HPF; magnification: 200×); (G) Case 10 of Group A: Methylene blue staining of gastric antral glandular crypts showing small short rod-shaped HP (red oval, 400×); (H) Case 10 of Group A: Pre-treatment duodenal mucosa with severe EOS infiltration (≈ 100 /HPF; magnification: 200×); (I) Case 10 of Group A: Post-treatment duodenal mucosa (omeprazole + azathioprine) showing reduced EOS infiltration (400×).

Radiological findings

Twenty-four patients underwent enhanced abdominal CT: 23 showed duodenal bulb wall thickening with abnormal mucosal enhancement, and 8 had gastric antral mucosal thickening. The typical imagological picture of these patients was shown in Fig. 3.

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

Abdominal enhanced CT examination of eosinophilic DBU. (A,B) Thickening of the duodenal bulb (large arrow) and descending part (small arrow) in case 10, with improvement after treatment. (C) Thickening of the duodenal bulb (large arrow) in case 3. (D) Thickening of the duodenal bulb (large arrow) in case 18.

Treatment and follow-up

For these patients with REDU, treatment included dietary avoidance (100%), PPIs (100%), glucocorticoids (46.7%), immunosuppressants (azathioprine/methotrexate, 33.3%), antihistamines (16.7%), and montelukast (40%) (Table 3). All patients received PPIs for a median duration of 20.5 [IQR: 11.3, 35.0] months; Seven patients received glucocorticoids for a median duration of 2.5[IQR:2.0,4.5] months; Thirty patients (100%) received food avoidance plus PPI treatment, 14 patients (46.7%) received food avoidance plus PPI plus glucocorticoid treatment, and 5 patients (17.7%) received food avoidance plus PPI plus glucocorticoid plus immunosuppressive agent treatment. Recurrence occurred in 7 patients, 6 of whom had PPI dose reduction or discontinuation (median interval from adjustment to recurrence: 50.0 [IQR: 35.0, 120.0] days).

For patients in the REDU cohort, at the last follow-up visit, 29 remained on PPI therapy, while 3 had persistent positive HP antibodies. Eleven patients achieved ulcer healing, with 8 of these patients responding to treatment consisting of prednisolone tapering combined with immunosuppressants and PPI maintenance therapy. In contrast, 8 patients had persistent ulcers despite the use of multiple treatment strategies.

Discussion

HP infection is highly prevalent in children, yet post-eradication EGIDs remain underreported. As far as we know, this study presents the largest pediatric cohort to date of REDU following HP eradication, shedding light on this underrecognized phenomenon.

Fujita et al. first reported a 14-year-old boy with exacerbation of eosinophilic duodenal ulcer after HP eradication, managed with PPIs10,13. Our study expands this evidence to 30 pediatric cases, 19 of whom developed persistent or worsening symptoms (including malnutrition, anemia, and hemorrhagic ulcers) after HP eradication. Thirteen cases (13/30) of REDU were located in the postbulbar duodenum. This finding emphasizes the need to consider EGID in pediatric patients with persistent or recurrent ulcers post-HP eradication, particularly those with eosinophilic infiltration of the duodenal bulb14,15.

The severity of HP infection-induced pediatric DBU may be closely associated with adverse outcomes such as REDU. Our study found that compared with Group C (children with HP-induced DBU who achieved cure), Group A (REDU post-HP eradication) had higher rates of giant DBU, anemia, elevated peripheral EOS, and subsequent duodenal bulb stenosis. Additionally, Patients in Group A received more frequent HP eradication therapy. This was because we previously misattributed the recurrence of DBU to failed HP eradication, and thus administered repeated anti-HP treatment. It has only been acknowledged in recent years that these recurrent lesions are in fact EGIDs and eosinophilic DBU16. Notably, even with negative post-eradication HP tests, the patients in Group A had persistent EOS infiltration and recurrent ulcers in the duodenal bulb, posing significant therapeutic challenges. Therefore, HP eradication in pediatric patients with HP-associated DBU requires prudence, and the optimal regimen should be selected based on individual patient characteristics.

Interestingly, we observed striking phenotypic differences between HP-associated ulcers and eosinophilic post-eradication ulcers—differences that have clinical implications for diagnosis. Eosinophilic post-eradication ulcers differ from HP-associated ulcers in three key ways: (1) clean vs. dirty ulcer base17. (2) shallow, pan-shaped vs. deep, irregular lesions; (3) less surrounding mucosal edema18. Histopathologically, HP-associated peptic ulcers are characterized by a marked neutrophilic infiltrate, with concurrent eosinophilic infiltration observed in some cases. In contrast, REDU or EGIDs are pathologically defined by an eosinophil-predominant inflammatory infiltrate, in which neutrophilic involvement is minimal or even absent. Capsule endoscopy is limited here due to stenosis-related capsule impaction and inability to obtain biopsies, reinforcing upper gastrointestinal endoscopy with biopsy as the gold standard for diagnosis.

The pathogenesis of EGIDs- beyond eosinophilic involvement- remains incompletely understood, but current research highlights multifactorial mechanisms involving IgE-mediated and cell-mediated responses19. These include Th2 cytokines (IL-4, IL-5, IL-13) and chemokines, which collectively promote eosinophil recruitment and activation in the gastrointestinal tract20. Thirty-four cytokines involved in Th1, Th2, and Th17 pathways were measured in this study and we found a dramatically increase of IL-8 level in these patients with eosinophilic DBU. Eosinophils can activate neutrophils, inducing the release of superoxide and IL-821. A prior study reported a 12.2-fold upregulation of IL-8 in the inflamed esophageal mucosa of adult patients with EoE (P < 0.001)22. The detection of IL-8 levels may serve as an indicator for judging the exacerbation or alleviation of diseases.

The exact mechanism underlying REDU post-HP eradication remains unclear, but one proposed hypothesis involves HP’s role in modulating gastrointestinal mucosal immunity23. HP infection typically elicits a Th1-dominated immune response, which suppresses Th2 cytokines associated with EoE. HP eradication may shift this balance toward Th2 dominance, fostering EOS infiltration and triggering EGID symptoms24. In EGID patients, elevated serum IgE levels are commonly observed25. The impact of HP effect the production of IgE produced by B Cells. HP infection may also influence IgE production by B cells: HP-induced Th1 responses (predominantly secreting IFN-γ) inhibit Th2 responses, thereby reducing B cell differentiation into IgE-secreting plasma cells This also explains why IgE levels are usually not high in HP infections and may be inversely correlated with allergic reactions24,26. Our findings suggest HP eradication may disrupt immune homeostasis, promoting eosinophilic inflammation in the duodenal bulb. Our findings—elevated serum IgE, peripheral blood EOS count and high allergen positivity in REDU patients—support the hypothesis that HP eradication disrupts immune homeostasis, promoting eosinophilic inflammation in the duodenal bulb.

Upper gastrointestinal endoscopy with duodenal bulb biopsy is essential for diagnosing REDU post-HP eradication. Most of our REDU patients had giant DBU, which developed or persisted after HP eradication. Notably, our REDU cohort had higher peripheral blood EOS counts (median: 491.2 cells/µL) than previously reported in pediatric EGID studies.

No definitive treatment guidelines for eosinophilic duodenal ulcers exist15,27. In our cohort, PPIs relieved symptoms but had a high recurrence rate (56.2%, 9/16) upon dose reduction or discontinuation. This is consistent with our finding that 50% of REDU patients carried CYP2C19 IM/PM genotypes—genotypes associated with reduced PPI metabolism and efficacy. Meta-analyses have also linked CYP2C19 IM/PM status to reduced PPI efficacy in peptic ulcers. Dietary avoidance is a standard conservative treatment for EGIDs28. In our study, all REDU patients received dietary avoidance plus PPIs, so we could not assess the efficacy of dietary avoidance alone. Glucocorticoids provided rapid symptom relief but required combination with immunosuppressants to prevent relapse—mirroring management strategies for inflammatory bowel disease29.

This study has several limitations. First, despite being a multicenter study, the rarity of REDU resulted in a relatively small sample size, which may introduce selection bias. Second, as a retrospective study, some data (e.g., long-term follow-up of growth parameters, the allergic history and EOS count in duodenal bulb at the time of HP diagnosis in Group A and C) were incomplete. Third, we did not conduct in-depth mechanistic studies (e.g., gut microbiota profiling or Th1/Th2 cytokine signaling pathway analysis). These limitations should be addressed in future prospective, large-cohort studies18,30.

Conclusion

This multicenter study establishes a clinical link between HP eradication and REDU in children, mediated by immune dysregulation—characterized by elevated serum IL-8, peripheral blood EOS, and marked EOS infiltration of the duodenal bulb. Clinicians should monitor EOS levels in post-eradication refractory duodenal bulb ulcers. Further research into HP-immune-microbiota interactions may identify novel targets for EGID therapy.