This article is mentioned in:

Introduction

Helicobacter pylori (H. pylori) is a highly prevalent infectious agent that affects ~40% of the global population (1). H. pylori was first discovered in 1983 by Robin Warren and Barry Marshall in gastric mucosal biopsy samples from patients with chronic active gastritis and peptic ulcer disease (2). Currently, this organism is recognized as the causative agent of gastritis, peptic ulcer disease, gastric adenocarcinoma and gastric B-cell lymphoma (MALT lymphoma) (3).

Several in vivo and in vitro studies have demonstrated that H. pylori can invade epithelial cells and the lamina propria, triggering a more intense mucosal inflammatory response than bacterial attachment to epithelial cells alone. This process is important in the initiation and development of gastric inflammation (4-7). Furthermore, H. pylori that invades the gastric mucosa may translocate to gastric lymph nodes, leading to chronic immune system stimulation (8).

H. pylori in the lamina propria or subepithelial regions may be underdiagnosed without an appropriate immunohistochemical (IHC) study. Ito et al (8) found H. pylori in both the mucous and deeper layers of gastric tissue in patients with gastric cancer using the TMDU antibody. However, the study focused on cancerous conditions and did not provide further descriptions or categorizations of the distribution patterns of the bacteria.

Proton pump inhibitors (PPIs) are among the most frequently prescribed medications to alleviate dyspeptic pain and have demonstrated in vitro anti-H. pylori activity (9). These drugs can reduce the H. pylori load while inhibiting its urease activity (10,11) and may interfere with certain H. pylori detection tests, potentially leading to false-negative results (12).

Following PPI therapy, rod-shaped H. pylori can convert into its coccoid form (13,14). Research has indicated that coccoid H. pylori may contribute to bacterial transmission and the recurrence of infection, even following antimicrobial treatment (13-16). However, the exact mechanisms underlying the pathogenesis of coccoid H. pylori remain poorly understood.

The detection of coccoid H. pylori requires IHC staining because simple histochemical stains cannot reliably detect low numbers of coccoid or intracellular bacteria, nor can they differentiate coccoid H. pylori from artefacts or other bacteria. IHC stains offer high specificity, enabling accurate identification of H. pylori and exclusion of other similar-shaped organisms (17).

It was hypothesized that PPI treatment for acid suppression may modify the gastric mucosal microenvironment, influencing the distribution of H. pylori in both the surface and subepithelial compartments. Using IHC staining, it was aimed to identify distinct infection patterns and their potential associations with clinical outcomes.

Materials and methods

The study protocol complied with the principles of the Declaration of Helsinki and was approved by the Institutional Ethics Committee of Navamindradhiraj University (approval no. 120/2566; Bangkok, Thailand). Gastric tissue biopsies from patients diagnosed with gastritis, who underwent endoscopic examination at Vajira Hospital (Faculty of Medicine, Navamindradhiraj University, Bangkok, Thailand) between October and December 2022, were reviewed retrospectively.

Inclusion criteria were as follows: i) Availability of at least four biopsies obtained from both the antrum and corpus as part of routine clinical practice; ii) diagnosis of gastritis confirmed by histopathological evaluation; and iii) sufficient formalin-fixed paraffin-embedded (FFPE) tissue for both hematoxylin and eosin (H&E) and IHC analysis. Additional exclusion criteria included: vii) inadequate biopsy samples, and viii) poor tissue preservation precluding histological interpretation.

As part of standard clinical practice, at least four biopsies were collected from both the gastric antrum and corpus. Patients were not specifically recruited for the present study, and no additional tissue samples were collected. FFPE blocks were processed for histopathological and IHC analysis. Clinical data were also extracted and recorded between April and May 2024. All identifying information was removed.

The patients were divided into two groups according to PPI therapy: Those receiving PPIs for at least a 14-day period prior to the procedure (PPI-treated group) and those not receiving PPIs or receiving them for a shorter period (PPI-untreated group).

Exclusion criteria

Patients were excluded from the analysis on the basis of the following criteria: i) Previous use of antibiotics, histamine (H2) receptor blockers, statins, aspirin, or non-steroidal anti-inflammatory drugs; ii) a prior history of H. pylori infection or triple therapy eradication; iii) the presence of underlying conditions, including autoimmune diseases, diabetes mellitus, cerebrovascular disease, peptic ulcers, HIV infection, or gastrointestinal cancers; iv) use of immunosuppressive drugs prior to endoscopy; v) diagnosis of Helicobacter heilmannii infection based on morphological criteria (for example, straight appearance, corkscrew-shaped spirals, or exceeding 3 µm in length) (18); and vi) cases of gastritis with specific etiologies other than H. pylori infection, such as reactive gastropathy or autoimmune gastritis.

H. pylori eradication therapy

H. pylori eradication treatment was defined as the prescription of either a PPI with at least two types of antibiotics initiated simultaneously or a fixed-dose triple therapy. Eradication was considered successful if no second treatment was prescribed after the initial treatment. Treatment failure, defined as the prescription of a second eradication treatment within 12 months, was used as a proxy for antimicrobial resistance.

Histopathology

All the paraffin blocks were sectioned at 3-µm thickness and stained with H&E in August 2023. Staining was performed at room temperature (~22-25˚C), with hematoxylin applied for 5-10 min and eosin for 1-2 min, following standard histological protocols.

Two pathologists (CS and KL) independently reviewed the H&E-stained slides. Cases with discordant results were reviewed together and discussed until consensus was reached. Neutrophilic and mononuclear inflammation, glandular atrophy, and intestinal metaplasia were assessed and graded (0=none, 1=mild, 2=moderate, 3=severe) according to the updated Sydney System (19).

Gastritis was classified on the basis of inflammatory process status (chronic and/or active) and the presence or absence of structural changes (intestinal metaplasia, atrophy, ulcers). This resulted in four distinct groups: i) Chronic nonactive gastritis without structural change, ii) chronic active gastritis without structural change, iii) chronic non-active gastritis with structural change, and iv) chronic active gastritis with structural change.

To analyze the associations between pathologic features and H. pylori presence, the cases were further regrouped as ‘non-active gastritis’ (groups i and iii) and ‘active gastritis’ (groups ii and iv). Additionally, they were regrouped as ‘without structural change’ (groups i and ii) and ‘with structural change’ (groups iii and iv).

IHC

IHC tests were performed on all cases in August 2023 using the Leica Bond-Max automated IHC staining platform (Leica Microsystems, Inc.) with 2-µm-thick consecutive sections prepared from FFPE tissue, in accordance with the manufacturer's instructions. Tissue sections were deparaffinized using Leica Bond Dewax Solution (cat. no. AR9222), and rehydration was performed automatically by the Bond-Max system using a graded ethanol series integrated into the protocol. Antigen retrieval was conducted using Epitope Retrieval Solution 2 (ER2; pH 9.0; cat. no. AR9640) for 20 min at 100˚C under standard automated conditions.

Endogenous peroxidase activity was blocked using the Peroxidase Block reagent included in the Leica Bond Polymer Refine Detection kit (cat. no. DS9800), which contains ~3% hydrogen peroxide. This blocking step was carried out for 5 min at room temperature (~22-25˚), according to the system's default protocol.

The following primary antibodies targeting H. pylori were used: mouse monoclonal anti-H. pylori antibody (1:800; clone TMDU-D8; cat. no. D369-3; MBL International Co.), monoclonal anti-H. pylori antibody (1:300; clone ULC3R; cat. no. MU880-5UCE; BioGenex Laboratories), polyclonal anti-H. pylori antibody (1:100; cat. no. 215A-76; Cell Marque), and polyclonal anti-H. pylori antibody (1:50; cat. no. B0471; Dako; Agilent Technologies, Inc.). Primary antibody incubation was performed at room temperature for 15 min, consistent with Leica's standard program for the Bond-Max system. BioGenex demonstrated superior diagnostic performance for H. pylori detection and was therefore used as the reference standard (gold standard) in the present study (Table SI).

Immunoreactivity was visualized using the Leica Bond Polymer Refine Detection kit (cat. no. DS9800), with incubation performed at room temperature for 15-30 min, as per the manufacturer's protocol. Positive (H. pylori-infected gastric biopsy) and negative (gastric tissue from sleeve gastrectomy patients without gastritis) controls were included. Stained slides were examined using a Nikon Eclipse E200 light microscope.

Detection of H. pylori

Two pathologists (CS and KL) independently evaluated the H. pylori IHC slides. Disagreements were resolved through joint discussion until a consensus was reached. Bacterial density was assessed separately for surface epithelial and subepithelial locations. The surface epithelial H. pylori density was graded using the updated Sydney System's standardized visual analogue scale, which classifies bacteria into four categories (19): S0: normal (no bacteria); S1: individual bacteria or small groups of <1/3 of the mucosal surface; S2: moderate (bacteria count greater than mild but less than severe); S3: severe (large groups of bacteria covering >2/3 of the mucosal surface).

The subepithelial H. pylori location (either intracellular or interstitial) was examined under high magnification (0.196 mm2 area, high-power field) using a Nikon Eclipse E200 light microscope. Subepithelial signals were counted below the lower border of the basement membrane, which served as the histological landmark.

For signal selection, only those exhibiting an intensity comparable to surface-attached bacilli morphology in the external or internal positive controls and showing a negative signal in the negative control were included. Homogeneous, indiscrete clumps were counted as a single signal, whereas heterogeneous, discrete signals within clumps were counted individually.

When the signal density varied across the slide, the region with the highest density (hot spot) was designated for counting. The subepithelial bacterial density was then graded on the basis of the number of immunopositive signals: I0: no signal; I1: 1-20 signals/HPF; I2: 20-50 signals/HPF; and I3: >50 signals/HPF.

Patterns of H. pylori

The antibodies demonstrating the highest diagnostic accuracy served as the gold standard for classifying H. pylori patterns. The classification, derived from combining bacterial presence in both surface epithelial and subepithelial locations, resulted in the following five groups: Group 1: Isolated surface epithelial pattern (S1-S3 and I0); Group 2: Isolated subepithelial pattern (I1-I3 and S0); Group 3: Predominant surface epithelial pattern (S3 and only I1); Group 4: Predominant subepithelial pattern (I3 and only S1); Group 5: Non-specific pattern (S1 and I1-I2, S2 and I1-I3, or S3 and I2-I3).

Statistical analyses

All the statistical analyses were conducted using Stata (version 13; StataCorp LP). The diagnostic performance of each IHC study was assessed by calculating the sensitivity, specificity, positive predictive value and negative predictive value.

Categorical variables were compared using the chi-square test or Fisher's exact test, as appropriate. Univariate and multivariate analyses utilized the Cox proportional hazards model. The time to clinical improvement (cumulative clinical improvement rate) was calculated as the duration between the onset of upper abdominal symptoms at diagnosis and either clinical improvement or the last follow-up. These data were compared between groups via Kaplan-Meier plots and the log-rank test.

To assess the associations between H. pylori infection patterns and treatment success, modified Poisson regression was used to estimate the relative risk (RR), with robust standard errors used to calculate 95% confidence intervals (CIs). Statistically significant difference was set at P<0.05. Only patients with consistent follow-up visits every 1-2 months for at least 6 months were included in the analysis.

Results

Demographic and histopathological characteristics of the study population

The present study included 255 patients with chronic gastritis diagnosed from tissue biopsies. The mean age of the patients was 47 years, and 130 (51.0%) were female. Nearly half (115 patients, 45.1%) had received PPI therapy, with a median treatment duration of 2.7 years. There were no significant differences in age or sex between the PPI-treated and untreated groups.

The demographic and histopathological characteristics of the study population are summarized in Table I. Among all participants, the most common type of gastritis, considering both inflammatory activity and structural changes, was chronic non-active gastritis without structural changes (68.2%). This was followed by chronic nonactive gastritis with structural changes (15.7%), chronic active gastritis without structural changes (12.2%), and chronic active gastritis with structural changes (3.9%). These trends were similar in patients receiving PPI treatment. However, among patients not using PPIs, chronic nonactive gastritis without structural changes was most common (66.4%). Chronic non-active gastritis with structural changes and chronic active gastritis without structural changes were equally prevalent (14.3%), whereas chronic active gastritis with structural changes was the least common (5.0%).

Table I Classification of 255 patients according to age, sex, histopathological findings, and H. pylori status by BioGenex (n=255).

Table I

Classification of 255 patients according to age, sex, histopathological findings, and H. pylori status by BioGenex (n=255).

Variables	Total, (n=255)		PPI positive (n=115)		PPI negative (n=140)		P-value
Age, years
<60	159	(62.4)	68	(59.1)	91	(65.0)
≥60	96	(37.6)	47	(40.9)	49	(35.0)	0.336a
Sex, n (%)
Female	130	(51.0)	66	(57.4)	64	(45.7)	0.063a
Male	125	(49.0)	49	(42.6)	76	(54.3)
Histology
Chronic active gastritis with structural change	10	(3.9)	3	(2.6)	7	(5.0)	0.519b
Chronic non-active gastritis with structural change	40	(15.7)	20	(17.4)	20	(14.3)	0.497a
Chronic active gastritis without structural change	31	(12.2)	11	(9.6)	20	(14.3)	0.251a
Chronic non-active gastritis without structural change	174	(68.2)	81	(70.4)	93	(66.4)	0.494a
H. pylori status
H. pylori positive	83	(32.5)	31	(27.0)	52	(37.1)	0.107a
Any surface	60	(23.5)	26	(22.6)	34	(24.3)	0.753a
Any subepithelium	75	(29.4)	25	(21.7)	50	(35.7)	0.015a

[i] Data are presented as number, (%). P-value corresponds to

[ii] ^aChi-square test or

[iii] ^bFisher's exact test. PPI, proton pump inhibitors; H. pylori, Helicobacter pylori.

H. pylori detection rate

Among the 255 gastric biopsies, the H. pylori detection rate by BioGenex was 32.5%. Bacteria were detected in the surface epithelium in 23.5% of the cases (72.3% of positive cases) and in the subepithelial location in 29.4% of the cases (90.4% of positive cases). The performance of all four antibodies is illustrated in Fig. 1.

Figure 1 Topographic distribution of immunoreactive H. pylori according to surface epithelial and subepithelial locations, using BioGenex, TMDU, Cell Marque and DAKO. Although the detection frequency and staining intensity varied, the immunoreactive H. pylori did not differ morphologically among these antibodies when detected in the surface epithelium. Remarkable differences were observed in the IHC results for subepithelial H. pylori in the lamina propria. (A-D) Scattered dot-like particles were found with (A) BioGenex and, to lesser extent, with (B) TMDU, a few particles with (C) Cell Marque, and no reactivity with (D) DAKO. (E-H) In the isolated subepithelial pattern, (E) BioGenex and (F) TMDU identified subepithelial bacteria while (G) Cell Marque and (H) DAKO did not identify any bacteria. This also clarifies that all four markers were negative for surface bacteria (original magnification, x60). H. pylori, Helicobacter pylori.

PPI use and H. pylori detection

A significantly lower H. pylori detection rate was observed among patients not using PPIs than among those in the PPI-treated groups: 21.7% vs. 35.7% for overall subepithelial staining (P=0.015) and 3.8% vs. 19.4% for the isolated epithelial pattern (P=0.047), respectively. However, no statistically significant associations were found between PPI treatment and other mixed histopathology patterns of gastritis. The results are summarized in Table II.

Table II Comparison of histopathological characteristics and group classification between PPI and non-PPI users in the H. pylori-positive cases, using BioGenex as the gold standard.

Table II

Comparison of histopathological characteristics and group classification between PPI and non-PPI users in the H. pylori-positive cases, using BioGenex as the gold standard.

Variables	Total (n=83)		PPI positive (n=31)		PPI negative (n=52)		P-value
Histology
Chronic active gastritis with morphological change	10	(12.0)	3	(9.7)	7	(13.5)	0.737b
Chronic nonactive gastritis with morphological change	17	(20.5)	8	(25.8)	9	(17.3)	0.353a
Chronic active gastritis without morphological change	26	(31.3)	10	(32.3)	16	(30.8)	0.888a
Chronic nonactive gastritis without morphological change	30	(36.1)	10	(32.3)	20	(38.5)	0.569a
HP positive	83	(100)	31	(100)	52	(100)	NA
Any surface staining	60	(72.3)	26	(83.9)	34	(65.4)	0.069a
Any subepithelial staining	75	(90.4)	25	(80.6)	50	(96.2)	0.047b
Isolated surface epithelial pattern	8	(9.6)	6	(19.4)	2	(3.8)	0.047b
Isolated subepithelial pattern	23	(27.7)	5	(16.1)	18	(34.6)	0.069a
Predominant surface epithelial pattern	17	(20.5)	7	(22.6)	10	(19.2)	0.879a
Predominant subepithelial pattern	8	(9.6)	2	(6.5)	6	(11.5)	0.704b
Non-specific pattern	27	(32.5)	11	(35.5)	16	(30.8)	0.657a

[i] P-value corresponds to

[ii] ^aChi-square test or

[iii] ^bFisher's exact test. PPI, proton pump inhibitors; NA, not applicable.

Mucosal density and location of H. pylori and grades based on the Sydney system visual analogue scale

Using BioGenex as a reference, a weak correlation (r=0.36) was observed between the surface epithelial density of H. pylori (evaluated via IHC) and active inflammation in the epithelial compartment (gastric mucosa) (Table III). However, H. pylori density was moderately correlated (r=0.45) with chronic inflammation and not correlated with active inflammation in the subepithelial compartment.

Table III Correlation between H. pylori-positive cases and histologic grades according to the updated Sydney System, using BioGenex as the gold standard.

Table III

Correlation between H. pylori-positive cases and histologic grades according to the updated Sydney System, using BioGenex as the gold standard.

	Surface		Subepithelium
Sydney system	+	-	+	-
Chronic inflammation
0	0	0	0	0
1	0	0	0	0
2	4	1	2	3
3	22	4	23	3
ra	0.046		0.451
P-value	0.805		0.011
PMN infiltration
0	13	5	13	5
1	0	0	0	0
2	3	0	2	1
3	10	0	10	0
ra	0.363		0.296
P-value	0.045		0.105
Glandular atrophy
0	18	3	18	3
1	4	1	3	2
2	3	1	3	1
3	1	0	1	0
ra	-0.065		-0.149
P-value	0.727		0.424
Intestinal metaplasia
0	22	3	20	5
1	0	1	1	0
2	4	1	4	1
3	0	0	0	0
ra	-0.200		0.027
P	0.281		0.887
H. pylori density
Surface
0	0	5	5	0
1	11	0	5	6
2	7	0	7	0
3	8	0	8	0
ra	0.663		0.285
P-value	<0.001		0.120
Subepithelium
0	6	0	0	6
1	10	4	14	0
2	6	1	7	0
3	4	0	4	0
ra	0.021		0.727
P-value	0.912		<0.001

[i] ^aSpearman's rank correlation coefficient (r). H. pylori, Helicobacter pylori.

Factors predicting H. pylori distribution patterns

Multivariate analysis (Table IV) revealed that PPI use and a histologically active pattern were independently associated with the isolated surface epithelial pattern [hazard ratio (HR), 8.02 and 3.55, respectively] and inversely associated with the isolated subepithelial pattern (HR, 0.24 and 0.05, respectively). Additionally, a histologically active pattern was independently associated with the predominant surface epithelial pattern (HR, 5.18) and the non-specific pattern (HR, 3.12). No statistically significant associations were observed between PPI treatment (treated vs. untreated groups) and age, sex, or the histopathology of gastritis.

Table IV Multivariable analyses of factors associated with Helicobacter pylori distribution patterns.

Table IV

Multivariable analyses of factors associated with Helicobacter pylori distribution patterns.

	Multivariable analysis (G1)			Multivariable analysis (G2)			Multivariable analysis (G3)			Multivariable analysis (G4)			Multivariable analysis (G5)
Variables	ORadj2	95% CI	P-value	ORadj2	95% CI	P-value	ORadj2	95% CI	P-value	ORadj2	95% CI	P-value	ORadj2	95% CI	P-value
Age, years
<60	1.00	Reference		1.00	Reference		1.00	Reference		1.00	Reference		1.00	Reference
≥60	1.01	(0.34-3.02)	0.989	1.04	(0.32-3.33)	0.953	1.30	(0.41-4.06)	0.657	2.09	(0.38-11.56)	0.398	0.80	(0.30-2.12)	0.655
Sex
Male	1.00	Reference		1.00	Reference		1.00	Reference		1.00	Reference		1.00	Reference
Female	1.08	(0.38-3.12)	0.881	0.68	(0.21-2.16)	0.509	3.82	(1.25-11.68)	0.019	0.77	(0.16-3.65)	0.738	0.73	(0.28-1.89)	0.518
Treatment
PPI negative	1.00	Reference		1.00	Reference		1.00	Reference		1.00	Reference		1.00	Reference
PPI positive	8.02	(2.62-24.50)	<0.001	0.24	(0.07-0.84)	0.026	0.59	(0.19-1.86)	0.366	0.60	(0.11-3.40)	0.565	1.09	(0.41-2.89)	0.869
Histology
Active inflammation
Non-active patterna	1.00	Reference		1.00	Reference		1.00	Reference		1.00	Reference		1.00	Reference
Active patterna	3.55	(1.18-10.72)	0.024	0.05	(0.01-0.25)	<0.001	5.18	(1.73-15.51)	0.003	0.36	(0.07-1.96)	0.237	3.12	(1.22-8.00)	0.018
Structural alteration
No structural changeb
With structural changeb

[i] ^aRegardless of structural changes;

[ii] ^bRegardless of gastritis activity. NA, data not applicable; PPI, proton pump inhibitors; OR, odds ratio; OR_adj, adjusted OR; CI, confidence interval; Group 1 (G1), Isolated surface epithelial pattern; Group 2 (G2), isolated subepithelial pattern; Group 3 (G3), Predominant surface epithelial pattern; Group 4 (G4), Predominant subepithelial pattern; Group 5 (G5), Non-specific pattern.

Clinical data and follow-up

Of the 255 patients included in the histopathological and IHC studies, 83 patients with H. pylori infection had complete clinical follow-up for a minimum of 3 months and were included in the outcome analyses.

Overall, 55.4% (46 patients) achieved clinical improvement (Table V). The clinical outcomes of patients on the basis of the clinical and pathological features of gastritis, as well as the H. pylori pattern, are presented in Table VI. Univariable analysis revealed that patients with chronic gastritis with active inflammation experienced greater clinical improvement than those with non-active infections did (improvement rate: 80.6; 95% CI: 66.42-91.44; P=0.005), whereas patients with the isolated subepithelial pattern had the lowest improvement rate compared with the other groups (improvement rate: 8.7; 95% CI: 2.25-30.51; P<0.001). However, only the isolated subepithelial pattern was confirmed as an independent factor for clinical outcomes through multivariate analysis (HR, 0.25; 95% CI: 0.12-0.50; P<0.001). Notably, PPI treatment was not found to be associated with clinical outcomes. The Kaplan-Meier survival curves of patients according to their H. pylori distribution pattern is included in Fig. 2.

Figure 2 Kaplan-Meier survival curves for H. pylori-infected patients in relation to PPI use and bacterial distribution patterns. (A) Clinical improvement rate based on PPI users and non-users. (B) Clinical improvement rate based on group classification (G1 + G3 + G5 vs. G2 + G4). G1, surface epithelial pattern; G2, subepithelial pattern; G3, predominant surface epithelial pattern; G4, predominant subepithelial pattern; G5, non-specific pattern; PPI, proton pump inhibitors.

Table V Univariable analyses of the clinical improvement rate at 3 months in the Helicobacter pylori-positive cases according to various clinicopathologic factors (n=83).

Table V

Univariable analyses of the clinical improvement rate at 3 months in the Helicobacter pylori-positive cases according to various clinicopathologic factors (n=83).

			Clinical improvement rate at 3 months
Variables	N	Number of improvements	Rate	(95% CI)	P-value
Total	83	46	55.42	(45.14-66.28)
Age, years
<60	31	14	45.16	(29.74-64.03)	0.474
≥60	52	32	61.54	(48.66-74.57)
Sex
Female	39	23	58.97	(44.24-74.31)	0.521
Male	44	23	52.27	(38.55-67.49)
Treatment
PPI negative	52	31	59.62	(46.74-72.89)	0.243
PPI positive	31	15	48.39	(32.64-66.96)
Histopathology
Active inflammation
Non-active patterna	47	17	36.17	(24.27-51.57)
Active patterna	36	29	80.56	(66.42-91.44)	0.005
Structural alteration
No structural changeb	56	31	55.36	(42.98-68.58)
With structural changeb	27	15	55.56	(38.25-74.44)	0.841
Topographic distribution
Isolated surface or predominant surface epithelial or non-specific patterns	52	39	75.00	(62.74-85.73)	<0.001
Isolated subepithelial pattern	23	2	8.70	(2.25-30.51)
Predominant subepithelial pattern	8	5	62.50	(32.56-91.3)

[i] Significant at P<0.05 (P-value corresponds to Log-rank test).

[ii] ^aRegardless of structural changes;

[iii] ^bRegardless of gastritis activity. CI, confidence interval; PPI, proton pump inhibitors.

Table VI Univariable and multivariable Cox proportional hazards regression survival analyses of factors associated with clinical improvement in Helicobacter pylori-positive cases.

Table VI

Univariable and multivariable Cox proportional hazards regression survival analyses of factors associated with clinical improvement in Helicobacter pylori-positive cases.

	Univariable analysis			Multivariable analysis
Variables	HR	95%CI	P-value	HRadj	95%CI	P-value
Age (years)
<60	1.00	Reference		1.00	Reference
≥60	0.85	(0.54-1.35)	0.497	0.97	(0.58-1.63)	0.921
Sex
Male	1.00	Reference		1.00	Reference
Female	1.15	(0.74-1.78)	0.541	1.00	(0.64-1.57)	0.984
Treatment
PPI negative	1.00	Reference		1.00	Reference
PPI positive	1.30	(0.82-2.06)	0.268	0.78	(0.44-1.37)	0.388
Histopathology
Active inflammation
Non-active patterna	0.55	(0.35-0.86)	0.009	1.13	(0.66-1.93)	0.664
Active patterna	1.00	Reference		1.00	Reference
Structural alteration
No structural changeb	1.00	Reference
With structural changeb	1.05	(0.66-1.66)	0.848
Topographic distribution
Isolated surface or predominant surface epithelial or non-specific patterns	1.00	Reference		1.00	Reference
Isolated subepithelial pattern	0.30	(0.17-0.51)	<0.001	0.25	(0.12-0.50)	<0.001
Predominant subepithelial pattern	0.56	(0.26-1.22)	0.147	0.49	(0.22-1.12)	0.093

[i] Significant at P<0.05 (P-value corresponds to Log-rank test).

[ii] ^aRegardless of structural changes.

[iii] ^bRegardless of gastritis activity; HR, hazard ratio; HR^adj, Adjusted HR; CI, confidence interval.

H. pylori eradication therapy

Among the H. pylori-positive patients, 73.5% (61 patients) received eradication therapy, while 26.5% (22 patients) were left untreated. Among the untreated patients, the majority (81.8%, 18 patients) had an isolated subepithelial pattern. Compared with the isolated surface epithelial pattern or mixed patterns, the isolated subepithelial pattern was associated with a lower eradication success rate. Modified Poisson regression analysis revealed that compared with the isolated subepithelial pattern, the surface epithelial pattern or mixed patterns had a greater risk ratio (RR) of successful H. pylori eradication, with RR values ranging from 1.46-1.67 (RR=1) (Table VII).

Table VII Factors associated with the success rates of H. pylori eradication.

Table VII

Factors associated with the success rates of H. pylori eradication.

Patterns of H. pylori	Number of patients treated	Eradication rate (%)	Treatment failure (%)	RR	95% CI	P-value
Isolated surface epithelial pattern	8	100	0	1.67	(0.81-3.43)	0.165
Isolated subepithelial pattern	5	60	40	1.00	Reference
Predominant surface epithelial pattern	15	93.3	6.7	1.56	(0.75-3.24)	0.238
Predominant subepithelial pattern	8	87.5	12.5	1.46	(0.68-3.14)	0.336
Non-specific pattern	25	92	8	1.53	(0.74-3.18)	0.252

[i] H. pylori, Helicobacter pylori.

Discussion

In the present study, the prevalence of H. pylori infection, as determined by IHC in 255 patients, was 32.5%, which is lower than the 40-45% reported in previous studies (1,20,21). This discrepancy may be attributed to differences in study periods, with more recent research indicating a decline in prevalence compared with earlier studies (1,20,21). This reduction in infection rates could be due to demographic variations, improved personal hygiene, improved living conditions, increased public awareness, and active H. pylori eradication efforts (1,20,21).

The ability to detect H. pylori and its specific localization may vary depending on the antibody used for IHC staining. Certain antibodies typically detect H. pylori only in superficial areas and not beneath the epithelium. Ito et al (8) examined the presence and invasive behavior of H. pylori via various methods, including IHC, PCR, bacterial culture and immunoelectron microscopy. They found H. pylori in both the mucous layer and the lamina propria in gastric cancer patients using the TMDU-mAb. In the present study, the diagnostic accuracy of BioGenex was comparable with that of TMDU, with BioGenex showing greater sensitivity in detecting H. pylori in deeper tissue, where TMDU was detected in only 68.9% of cases.

H. pylori was initially considered a non-invasive pathogen that resides solely in the gastric lumen and affects only gastric epithelial cells. However, most previous studies focused primarily on the surface epithelium. Subsequent in vivo and in vitro studies have shown that H. pylori is invasive and can also be found in the lamina propria (4-8). Expanding the focus to include the subepithelial location may provide a more accurate representation of H. pylori prevalence.

The presence of H. pylori in the subepithelial region correlated more strongly with the severity of chronic inflammation than surface bacteria in active gastritis, which aligns with the findings of Ito et al (8). This suggests a location-dependent immunological response, potentially involving macrophage interaction and contributing to the induction and maintenance of chronic inflammation in H. pylori-infected gastric mucosa. While surface epithelial H. pylori typically induces IL-12 production and Th1 cell accumulation (22), subepithelial bacteria, interacting with lamina propria macrophages and T cells, may elicit a distinct chronic inflammation. The present study observed a significantly higher prevalence of isolated subepithelial H. pylori (27.7%) compared with Ito et al's (2%). Notably, Ito et al (8) reported 46 PCR-positive H. pylori cases in the stomach; however, 4 of these cases exhibited negative IHC results for both surface and subepithelial bacteria, with only 1 showing isolated subepithelial positivity. The differences in findings may be explained by several factors. Firstly, Ito et al (8) employed more thorough tissue sampling, which may have led to an increased detection of surface bacteria.

In contrast to surface-attached bacilli, subepithelial H. pylori presented as dot-like signals, which may represent bacterial casts or coccoid forms. Additionally, the enhanced sensitivity of BioGenex used in the present study may explain the discrepancies. BioGenex utilizes the ULC3R clone to identify specific epitopes of H. pylori that are resistant to intracellular digestion and capable of detecting antigenic forms of H. pylori resembling the bacillary type, even in coccoid forms. BioGenex detected all TMDU-positive cases in the subepithelium, indicating greater sensitivity. Furthermore, differing positivity cutoffs (1 in the present study vs. ≥10 signals in Ito et al's) could contribute to the observed variation. Consequently, the precise prevalence of isolated subepithelial H. pylori remains inconclusive, necessitating further validation.

A significantly lower detection rate of H. pylori in the subepithelial compartment was observed in patients treated with PPIs than in those not treated with PPIs. These findings suggest that PPI use may inhibit H. pylori infiltration into the lamina propria.

The underlying mechanism may involve inhibiting MMP-2/TIMP-3 interactions. These effects could protect against gastric mucosal injury, thereby reducing the degree of epithelial damage that facilitates bacterial translocation (23).

In addition to the diagnostic challenges of detecting subepithelial bacteria, their presence may have implications for clinical outcomes. First, subepithelial particles could signify residual bacteria from the surface that remain viable in unbiopsied areas of the stomach, potentially causing persistent symptoms. Second, immunoreactive signals in subepithelial regions may represent debris from dead bacteria, which can continue to trigger inflammation and contribute to ongoing inflammatory responses.

Third, subepithelial immunoreactive particles may reflect a transformation from the usual bacillary form of H. pylori to an atypical coccoid form under stressful conditions. These coccoid forms could revive and revert to culturable bacilli. This transformation may contribute to treatment failure and relapse, as coccoid forms may evade immune detection while remaining susceptible to antibiotics (13-16). These findings may explain our results, which revealed that the eradication rate of H. pylori treatment was greater in patients with isolated surface epithelial or mixed infection patterns than in those with an isolated subepithelial pattern. The presence of the latter pattern may provide indirect evidence supporting the consideration of a step-up eradication therapy regimen. However, establishing a definitive link between the isolated subepithelial pattern and treatment success or failure remains challenging due to the small sample size.

Unlike treated patients, most untreated H. pylori-positive patients exhibited an isolated subepithelial pattern, likely due to underdiagnosis. Most treated patients are diagnosed using the rapid urease test (RUT), H&E staining, Giemsa, or polyclonal DAKO IHC, which are routinely employed in Vajira Hospital. By contrast, untreated patients predominantly presented with an isolated subepithelial pattern, which typically yields negative results for H&E, Giemsa, or polyclonal DAKO IHC. A total of 5 patients were treated on the basis of positive RUT results. However, RUT is rarely performed at our institution; treatment decisions are typically based on histopathology (H&E), special stains, or IHC results. This underscores the potential for significant underdiagnosis in real-world clinical practice.

Whether this staining represents live subepithelial bacteria or merely debris from dead bacteria remains unclear. Nevertheless, our results suggest that this immunoreactive signal should not be overlooked and warrants further investigation.

The current reporting system for H. pylori may benefit from the incorporation of immuno-stained subepithelial bacteria as a prognostic factor. The authors propose reporting H. pylori-IHC as an ancillary study alongside the Updated Sydney System to provide a clearer summary of H. pylori density and improve clinical communication. Specifically, infection patterns should be reported under the heading ‘H. pylori-positive cases’, including specific location details. This would enable clinicians to accurately identify infection types and guide treatment decisions. The isolated subepithelial pattern may suggest the necessity for step-up therapy, closer monitoring, or additional testing, such as urea breath tests or stool H. pylori antigen tests, in accordance with current guidelines (24). Using IHC, H. pylori density grading can be easily summarized on the basis of the groups described in the methodology for clinical implications.

The overall pattern should be reported under the heading ‘H. pylori-positive cases’ followed by a description of the specific location. This approach would improve communication with clinicians, for example, ‘isolated surface epithelial pattern (surface bacteria 3, negative subepithelial signal)’ or ‘non-specific pattern (surface bacteria 3, subepithelial signal 2)’. This pattern group system aims to improve description of H. pylori behavior and could guide further management, particularly for the isolated subepithelial pattern group. In such cases, patients should undergo a urea breath test or a stool H. pylori antigen test, according to guidelines. The presence of this pattern suggests that step-up therapy and close follow-up or further investigation should be considered. By contrast, the surface-attached groups had a greater likelihood of successful treatment than did the isolated subepithelial pattern group.

Although the present study provides valuable insights into H. pylori, it has several limitations. The primary limitations are its retrospective design and the small number of patients with complete clinical follow-up data, which may affect the generalizability of the findings. Additionally, since only omeprazole is included in the national reimbursement program, other PPIs could have varying effects on the outcomes. Further prospective studies with larger cohorts and broader medication coverage are needed to validate and expand upon these findings.

In conclusion, H. pylori, particularly in subepithelial locations, can be underdiagnosed even with IHC tests. The detection rate is influenced by the sensitivity of the primary antibody used and the location of the bacteria. The effect of PPIs might prevent surface bacterial translocation. This feature should be included in pathological assessments and reports as an ancillary study alongside the updated Sydney System. Subepithelial H. pylori infection could serve as an independent prognostic factor for clinical outcomes and could guide further patient management.

Supplementary Material

Sensitivity, specificity, positive predicted value, negative predicted value, accuracy, and kappa measure of agreement for BioGenex, TMDU, Cell Marque and DAKO.

Acknowledgements

The authors would like to thank Mrs Unaporn Sitthivilai (Department of Anatomical Pathology, Faculty of Medicine, Vajira Hospital, Navamindradhiraj University, Bangkok, Thailand) for her technical assistance and Professor Siriwan Tangjitgamol (Department of Obstetrics and Gynecology, Faculty of Medicine, Vajira Hospital, Navamindradhiraj University, Bangkok, Thailand; Obstetrics and Gynecology Section, MedPark Hospital, Bangkok, Thailand) for her support in manuscript preparation.

Funding

Funding: No funding was received.

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

KL and CS conceived and designed the study, conducted the literature review, analyzed and interpreted the data. CS wrote the manuscript. KL revised the manuscript and performed the statistical analysis. CS and KL confirm the authenticity of all the raw data. Both authors read and approved the final version of the manuscript.

Ethics approval and consent to participate

The study protocol was approved by the ethics committee of the Navamindrajhiraj University (approval no. 120/2566; Bangkok, Thailand) before study initiation. The Ethics Committee/Institutional Review Board waived the requirement of written informed consent for participation from the participants or the participants' legal guardians/next of kin because due to the retrospective nature of the study.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References