In 1994, the International Agency for Research on Cancer and the World Health Organization classified Helicobacter pylori as a class 1 carcinogen, making it the only bacterium to receive this classification.¹ It has been reported as a causative factor in many malignant and benign gastrointestinal diseases, such as chronic gastritis, peptic ulcer^2,3 and gastric cancer.^4-7 Considering its pathology, there is a need for widespread testing and eradication of H. pylori in the population. To identify individuals possessing severe gastric pathologies, mass screening is required but upper gastrointestinal endoscopy and direct biopsy tests are unsuitable because they are expensive, time-consuming and invasive.⁸ Instead, upper gastrointestinal endoscopy can be employed as a second confirmatory step for individuals identified by mass screening.

To identify individuals requiring such advanced diagnosis (e.g., upper gastrointestinal endoscopy), a simple serum test that could be completed within 10 to 20 minutes would be optimal. For the last two decades, clinicians have been using a range of invasive and noninvasive methods to detect H. pylori. However, there is an increasing demand for noninvasive methods but advances in such techniques have been limited. Noninvasive diagnostic tests for H. pylori include an enzyme-linked immunosorbent assay (ELISA) and a latex agglutination turbidity (LA) assay using serum samples, a stool antigen test, and a urea breath test. These tests each have their own advantages and disadvantages. The urea breath test detects the non-radioactive but isotopic ¹³CO₂ product of the H. pylori-specific urease reaction after the uptake of ¹³C-urea, and requires a specific ¹³C-detector device. It is thereby specific but costly. Some stool antigen tests use a sandwich ELISA to detect H. pylori-specific protein from H. pylori bacteria in stool samples, which contain an array of complex proteins from all gastrointestinal microbiomes. Such tests require specific discrimination of H. pylori protein and can be labor-intensive (taking a few hours to complete), although simple immunochromatography of stool antigen tests has recently been developed.^9,10

Serum is an ideal diagnostic sample because it comprises well-studied, relatively simple components, mainly albumin and globulins. It has previously been reported that ELISA and LA assays based on serum samples have similar potency to detect H. pylori infection using anti-H. pylori antibodies as assays based on blood samples.^11,12 These tests cannot detect infections before humoral immunity induces an increase in anti-H. pylori antibodies and these antibodies persist in the blood even after the eradication of H. pylori, with some patients remaining seropositive for 1 to 4 years.^13-15 Despite the disadvantages of such tests for diagnostic purposes, antibody tests can indicate previous exposure to a specific bacterium as a result of humoral immunity memory. Antibody tests report not only exposure to H. pylori in the stomach but may also serve as an indicator of the disease status of an individual via humoral immunity. A recently developed automated LA assay takes just 10 to 20 minutes to complete and allows for high-throughput analysis which uses total H. pylori lysate antigens on the latex bead surface, where numerous antibodies (immunoglobulin [Ig]G, IgA, and IgM) against H. pylori antigens are trapped.^12,16-18 The LA assay could potentially be applied for mass screening of H. pylori infection in the general population.¹⁹ Such a test may provide information on the infection status and may predict other disease status information that may be useful for clinicians prior to upper gastrointestinal endoscopy.

In Mongolia, the H. pylori infection rate is high among all ages (~83.8%),²⁰ particularly among the younger generation. In fact, Mongolia has reported the highest rates of morbidity and mortality from gastric cancer of any country consistently over the last decade. Diagnostic tools for H. pylori infection are vital in Mongolia and need to be applicable in rural settlements. In the International Classification of Disease 11th Revision by World Health Organization, H. pylori infection is formally recognized as an infectious disease and it is recommended that a treatment strategy is implemented for the general population in high-risk communities.^21,22 The risk of gastric cancer development is 5 to 10 times higher in the case of certain gastric mucosal alterations (atrophic gastritis and intestinal metaplasia).^23-25 Therefore, predicting the gastric mucosal status prior to cancer may encourage patients to undergo upper gastrointestinal endoscopy, which could be crucial to decrease the prevalence of gastric cancer in Mongolia.

In this study, we evaluated both biopsy samples and serum samples from the same individuals to understand the H. pylori infection status and the gastric mucosal alteration status via the anti-H. pylori antibody data derived from the LA assay.

1. Study population and biopsy and serum collection

In total, 580 volunteers with dyspeptic symptoms from Khuvsgul province (in the north), Uvs province (in the west) and Ulaanbaatar (capital city, in the central region) in Mongolia participated in the study, during the period from November 2014 to August 2016. The study excluded the individuals who used acid inhibitors within 1 month.²⁶ Blood samples were collected on the same day that upper gastrointestinal endoscopy was performed and were centrifuged within 3 hours of collection. For the rapid detection of H. pylori infection, one biopsy specimen for the rapid urease test (Mon-Hp, developed at the Mongolian National University of Medical Sciences, Ulaanbaatar, Mongolia) was obtained from the antrum and checked in terms of infection status at the bedside. In addition, one biopsy specimen was taken from the antrum for H. pylori culture. The specimen was placed in transfer medium and then immediately frozen. Three biopsy samples for histology were also taken from the antrum, corpus and angles in each stomach, and were kept in 10% formaldehyde neutral buffer solution (Nacalai Tesque, Kyoto, Japan). Samples for culture and histology were transferred to Ulaanbaatar and serum and biopsy samples were immediately stored at –80℃. Then, all samples were transferred internationally under normal (histology samples) or frozen (stomach biopsy samples and serum samples) conditions for H. pylori culture and further analysis at the Department of Environmental and Preventive Medicine, Oita University Faculty of Medicine, Japan.

2. Serologic E-plate, PG I, PG II, and LA assays

Samples were sent to a commercial clinical laboratory center (Special Reference Laboratories, Tokyo, Japan) that performed ELISA to detect anti-H. pylori IgG (E-plate), pepsinogen (PG) I and PG II (Eiken Chemical Co., Ltd., Tokyo, Japan), and samples were returned then stored at –80℃ until the day of the LA assay. For the LA assay, conducted in 2017 to 2018, serum samples of more than 500 µL in volume were centrifuged at 9,000 ×g for 5 minutes at room temperature. The supernatants (400 µL) were then transferred to a specific sample cup. and brought to the Clinical Laboratory of Oita University Hospital, then automatically analyzed using an automated latex aggregation assay (H. pylori-LATEX “SEIKEN”, Denka, Japan) on a clinical chemical automatic analyzer (JCA-BM8020G, BioMajesty, Nippon Denshi, Japan).

3. H. pylori status and disease status as determined from gastric biopsy samples

A modified Giemsa staining method was used for the histological assessment of biopsy samples of the antrum and corpus.^27,28 One experienced pathologist (T.U.) evaluated all of the biopsy materials and determined the existence of H. pylori in the histological samples and immunohistochemistry (IHC) of H. pylori. Antral biopsy specimens for culture were homogenized and cultured on commercial Helicobacter plates (Nissui Pharmaceutical Co., Ltd., Tokyo, Japan) as described previously.²⁰

The disease status of individuals was described histologically using Lauren’s classification and the Japanese classification of gastric carcinoma.^8,29 Because our gastric samples were collected at multiple survey centers in Mongolia, we used disease information from the histology samples read by one expert pathologist. For non-malignant mucosal tissues, acute inflammation (polymorphonuclear neutrophil inflammation), chronic inflammation (mononuclear cell infiltration), atrophy and intestinal metaplasia were classified. Bacterial density and histological assignments were classified into four grades according to the updated Sydney system: 0 as normal, 1 as mild, 2 as moderate, and 3 as marked.³⁰ Biopsy samples that showed signs in one or more locations were considered positive in terms of their gastritis status based on the histopathology of gastric biopsies through the updated Sydney system described in Table 1.³⁰

Table 1 . Definition of Disease from Histology Samples According to the Updated Sydney System.

Disease	Definition
Healthy	Scores of neutrophils, monocytes, and intestinal metaplasia were zero in all biopsy locations in the stomach and there was no gastric cancer
Chronic gastritis	Score of neutrophils and/or monocytes infiltration was 1 or more, but no gastric cancer, intestinal metaplasia, or atrophy in any biopsy location in the stomach
Atrophic gastritis	Score of atrophy was 1 or more, but no gastric cancer or intestinal metaplasia in any biopsy location in the stomach
Intestinal metaplasia	Score of intestinal metaplasia was 1 or more, but no gastric cancer in any biopsy location in the stomach
Gastric cancer	Gastric cancer was observed

4. Statistical analysis

SPSS version 25.0 (SPSS Inc., Chicago, IL, USA) was used for all analyses. The median of continuous variables, the odds ratio (OR), sensitivity, specificity, the Mann-Whitney U test, and the univariate and multiple regressions of the OR were calculated with a 95% confidence interval and p-value. For univariate logistic analysis, all factors with a p-value less than 0.05 were subject to multivariate analysis. Receiver operating characteristic (ROC) curves were drawn for serum tests and markers, and the area under the curve (AUC) and cutoff values were calculated.

5. Ethics

All participants provided written consent and the research was approved by the Ethics Committees of the Ministry of Health (No. 03), the Ethics Committee at Mongolian National University of Medical Sciences (No. 13-02/1A) and the Ethics Committee at Oita University, Faculty of Medicine (Yufu, Oita, Japan) (P12-10 and No. 1660).

In total, 580 individuals were enrolled in the study and underwent upper gastrointestinal endoscopic examination for gastric biopsy collection and serum sample collection for LA, E-plate test, PG I, and PG II analyses. Histology, IHC, and H. pylori culture were used as gold standards to determine the H. pylori infection status. Specimens were categorized as Hp+ if one of the histology, IHC, or culture results were H. pylori-positive, and Hp– if all of the histology, IHC, and culture results were H. pylori-negative. Out of 580 individuals, serum samples from 96 participants were excluded from the analysis for various reasons, such as an insufficient volume of serum, missing histology data, missing serum test data, and unavailable data (Fig. 1, Supplementary Table 1). The rapid urease test was not used to categorize the H. pylori status in this study, but 18 cases that underwent rapid urease test were H. pylori-positive, but all cases for which the histology, IHC, and culture results were H. pylori-negative were removed from the main analysis, and 18 cases reporting past infection on questionnaires (Hp– and successful eradication) were also removed from the main analysis and showed separately (Supplementary Table 1). Overall, 484 individuals were enrolled in this study.

Figure 1. Summary of the screening examination. In the initial study, individuals (n=580) were excluded who used acid inhibitors within 1 month. After the exclusion of samples by previously eradicated history etc., enrolled individuals in this study were selected (n=484). Hp+ and Hp– groups were determined by criteria showing a dashed line. Specimens were categorized as Hp+ if one of the histology, immunohistochemistry (IHC) or culture results were H. pylori-positive, and Hp– if all of the histology, IHC and culture results were H. pylori-negative. Atrophic gastritis (AG)+ group includes individuals of AG, AG– group includes individuals of healthy, chronic gastritis, intestinal metaplasia, and gastric cancer. LA, latex agglutination turbidity; E-plate, anti-H. pylori IgG enzyme-linked immunosorbent assay test; PG, pepsinogen.

Regarding the characteristics of the participating individuals, the mean age was 44 years (range, 16 to 79 years) and 63.2% (306/484) were female. There were 128 individuals categorized as Hp– and 356 as Hp+. There were statistically significant differences between the Hp– and Hp+ groups based on the data from the LA, IgG, PG I, PG II, and PG I/II analyses (p<0.0001 or all) (Table 2).

Table 2 . Characteristics of the Individuals Enrolled in the Study.

Variable	Helicobacter pylori infection group*		p-value	Total
	Hp–	Hp+
Total	128	356		484
Age, mean±SD (range), yr	47±14 (18–79)	42±14 (16–78)	NS	44±14 (16–79)
Sex, No. (%)			NS
Male	52 (40.6)	122 (34.3)		172 (35.5)
Female	75 (58.6)	231 (64.9)		306 (63.2)
Unknown	1 (0.8)	3 (0.8)		4 (0.8)
Anti-H. pylori LA, median (IQR), U/mL	11.7 (4.8–28.6)	30.4 (17.5–52.4)	0.0001	26.4 (12.2–46.5)
Anti-H. pylori IgG, E-plate, median (IQR), U/mL	5.0 (3.0–10.0)	15.0 (9.0–24.0)	0.0001	12.0 (6.0–21.0)
PG I, median (IQR), U/mL	22.2 (15.4–32.7)	39.6 (28.7–51.5)	0.0001	34.5 (23.4–48.0)
PG II, median (IQR), U/mL	6.6 (5.0–9.0)	13.7 (10.5–19.3)	0.0001	12.1 (8.1–17.3)
PG I/II ratio, median (IQR)	3.5 (2.3–4.5)	2.9 (2.1–3.5)	0.0001	3.0 (2.1–3.8)

LA, latex agglutination turbidity; IQR, interquartile range; IgG, immunoglobulin G; E-plate, anti-H. pylori IgG enzyme-linked immunosorbent assay test; PG, pepsinogen; NS, not significant..

*Specimens were categorized as Hp+ if one of the histology, immunohistochemistry (IHC), or culture results were H. pylori-positive, and Hp– if all of the histology, IHC, and culture results were H. pylori-negative..

Using the test value of the Hp+ and Hp– groups, ROC analysis was performed for the LA, E-plate, PG I, PG II, and PG I/II analyses and age (Table 3). The AUC of the LA assay was 0.736, and its cutoff value was 18.35 U/mL. The sensitivity and specificity of the LA assay were 74.2% and 65.6%, respectively.

Table 3 . Receiver Operating Characteristic Analysis between the Hp+ and Hp– Groups.

Test	Diagnostic accuracy of the serology for predicting the Helicobacter pylori status
	AUC (95% CI)	Cutoff value	Sensitivity, %	Specificity, %
LA	0.736 (0.682–0.790)	18.35	74.2	65.6
E-plate	0.786 (0.734–0.838)	8.50	75.6	80.3
PG I	0.759 (0.710–0.809)	32.75	66.6	75.8
PG II	0.829 (0.781–0.877)	9.35	83.4	76.6
PG I/II ratio	0.381 (0.317–0.445)	2.65	57.6	32.8
Age	0.400 (0.343–0.458)	47.50	38.6	47.2

Specimens were categorized as Hp+ if one of the histology, immunohistochemistry (IHC) or culture results were H. pylori-positive, and Hp– if all of the histology, IHC and culture results were H. pylori-negative..

AUC, area under the curve; CI, confidence interval; LA, latex agglutination turbidity; E-plate, anti-H. pylori IgG enzyme-linked immunosorbent assay test; PG, pepsinogen..

To understand the characteristics of the serum test values, we added disease status information, as defined for the five groups shown in Table 1. Of the 484 cases, the most abundant disease status was atrophic gastritis (221 cases; 45.6%), followed by intestinal metaplasia (94 cases; 19.4%), chronic gastritis (86 cases; 17.7%) and gastric cancer (37 cases; 7.6%) (Fig. 2). There were 46 healthy individuals (9.5%). Importantly, the sex ratio (female/male) in the gastric cancer group was significantly lower (0.42) than in all groups (1.79) and the healthy group (1.70).

Figure 2. Distribution of case numbers with respect to disease and sex. The ratio of females/males for each disease group is shown at the top of each bar, and the value for all samples was 1.79. CG, chronic gastritis; AG, atrophic gastritis, IM, intestinal metaplasia; GC, gastric cancer.

We also analyzed the LA values between disease groups. The median values for healthy cases, chronic gastritis, atrophic gastritis, intestinal metaplasia and gastric cancer were 11.7, 26.6, 35.7, 24.3, and 17.6 U/mL, respectively (Fig. 3A, Supplementary Table 2). The atrophic gastritis group showed the highest median value for the LA test compared with the other tests, and this difference was statistically significant (Mann-Whitney test, p<0.0001 to p<0.006). In addition, the Hp+ groups showed higher LA values than the Hp– group for all disease types, although the differences were not statistically significant (Fig. 3B). Moreover, the Hp+ group of atrophic gastritis showed a significantly higher LA value than all other Hp+ and Hp– groups.

Figure 3. Comparison of the latex agglutination turbidity (LA) value distribution across the five disease groups (A) and those with Hp status (B). Statistical analysis was performed by a Mann-Whitney U test. Specimens were categorized as Hp+ if one of the histology, immunohistochemistry (IHC), or culture results were H. pylori-positive, and Hp– if all of the histology, IHC, and culture results were H. pylori-negative. CG, chronic gastritis; AG, atrophic gastritis, IM, intestinal metaplasia; GC, gastric cancer. *p≤0.01, ^†p≤0.001, ^‡p≤0.0001.

This result indicated that patients with atrophic gastritis expressed the highest serum anti-H. pylori values throughout gastric disease progression from early infection toward gastric cancer. It may therefore be possible to detect atrophic gastritis, via this characteristic high level of serum anti-H. pylori antibody, before the onset of intestinal metaplasia or gastric cancer. Next, we divided individuals into two groups, the atrophic gastritis-positive group (AG+ group) and the atrophic gastritis-negative group (AG– group), which included healthy cases and those with chronic gastritis, intestinal metaplasia and gastric cancer), then compared these two AG+ and AG– groups by ROC analysis. We compared the AUC and the diagnostic cutoff values between AG+ cases and age groups (Table 4). Among the tests to predict atrophic gastritis patients, the LA assay resulted in the highest AUC (0.673) and its best cutoff value was 31.95 U/mL. Using this cutoff value, the LA assay showed the highest specificity (73%) among the five tests. On the other hand, PG I and PG II showed higher sensitivity (68% and 71%, respectively). We determined the AG positive and negative status, performed logistic regression analysis and calculated the OR in univariate and multivariate logistic regression analysis (Table 5). We included the serum tests that showed significant univariate regression (p<0.05) and performed multivariate analysis. For multivariate analysis, four tests, namely the LA assay, E-plate, PG I, and PG II tests, were calculated as giving statistically significant results. The PG I/II ratio and age were not found to be risk factors for atrophic gastritis at the individual level. The OR, as determined by multiple regression, for the LA assay was 2.51, which was the highest among the tests (p<0.0001). Basically, individuals who had a test LA value of more than 32.0 had a 2.5 times higher risk of gastric atrophic gastritis than those with a test LA value of less than 32.0 (p<0.0001).

Table 4 . ROC Analysis of Atrophic Gastritis and Other Diseases Based on LA, E-plate, PG I, PG II, and PG I/II Ratio Tests and Age.

Test	Diagnostic accuracy of serology for predicting atrophic gastritis
	AUC (95% CI)	Cutoff	Sensitivity, %	Specificity, %
LA	0.673 (0.625–0.721)	31.95	57	73
E-plate	0.658 (0.610–0.707)	12.50	62	63
PG I	0.632 (0.583–0.681)	32.95	68	57
PG II	0.640 (0.591–0.689)	10.95	71	53
PG I/II*	0.471 (0.419–0.522)	2.95	51	48
Age*	0.367 (0.317–0.417)	47.50	33	50

ROC, receiver operating characteristic; LA, latex agglutination turbidity; E-plate, anti-Helicobacter pylori IgG enzyme-linked immunosorbent assay test; PG, pepsinogen; AUC, area under the curve; CI, confidence interval..

*PG I/II ratio and age were excluded from further analysis because they are not effective in determining the AG status (sensitivity and specificity both less than 50%)..

Table 5 . Univariate and Multiple Logistic Regression Analysis for Atrophic Gastritis.

Test	AG status by best cutoff	AG status by histology (gold standard)*		Univariate regression		Multivariate regression
		AG+	AG–	Odds ratio (95% CI)	p-value	Odds ratio (95% CI)	p-value
Latex, U/mL	Positive	125	71	3.52 (2.41–5.12)	0.0001‡	2.51 (1.51–4.18)	0.0001‡
	Negative	96	192
E-plate, U/mL	Positive	137	98	2.74 (1.89–3.97)	0.0001‡	1.17 (0.70–1.96)	0.538
	Negative	84	165
PG I, U/mL	Positive	150	112	2.84 (1.96–4.13)	0.0001‡	1.90 (1.22–2.95)	0.004†
	Negative	71	151
PG II, U/mL	Positive	157	124	2.75 (1.88–4.01)	0.0001‡	1.50 (0.94–2.37)	0.082
	Negative	64	139

AG, atrophic gastritis; CI, confidence interval; E-plate, anti-Helicobacter pylori IgG enzyme-linked immunosorbent assay test; PG, pepsinogen..

*AG+ or AG– in each test was determined using the best cutoff of each test in Table 4. AG+ or AG– was determined by histology; ^†p≤0.01; ^‡p≤0.0001..

In this study, we analyzed anti-H. pylori antibodies in serum samples from 484 Mongolian individuals that tested H. pylori-negative or -positive by an LA assay using an automatic clinical chemical analyzer. From the ROC analysis of Hp+ and Hp– Mongolian specimens, the best cutoff value for the LA assay was 18.35 U/mL, which was higher than the data from Japan and other Asian countries (6.55 to 12.55 U/mL), with the exception of Bhutan (23.95 U/mL).¹⁶ Regarding the sensitivity and specificity for H. pylori infection, the LA assay did not perform highly and the PG II test was considered the most sensitive and specific among the five tests analyzed in this study (Table 3). However, we found that the LA assay values for individuals with atrophic gastritis were higher than healthy and other disease groups among tested (Fig. 3, Table 4, and Supplementary Table 2). According to the results of this study, there was a 2.5 times higher atrophy risk in the individuals with latex values of more than 32 U/mL by multiple logistic regression analysis (Table 5). This result is consistent to previous reports on correlation between gastric mucosal alteration and serological test values.^31-33 We propose the usage of a noninvasive LA assay to predict atrophy in the Mongolian population.

Here, individuals with an LA value greater than the 32.0 U/mL cutoff (AG+ group) are recommended for upper gastrointestinal endoscopy and considered for H. pylori eradication therapy to halt the progression of gastric mucosal alterations into intestinal metaplasia and ultimately gastric cancer. In our study population, the AG+ group identified from the LA assay (221 cases in total) included 56.5% (125 cases) of atrophic gastritis, 27 cases of intestinal metaplasia and six cases of gastric cancer (group 1 in Fig. 4).

Figure 4. Grouping of individuals based on the detection of those requiring upper gastrointestinal endoscopy. Group 1 (196 cases), group 2 (202 cases), group 3 (86 cases). LA, latex agglutination turbidity; PG, pepsinogen; CG, chronic gastritis; AG, atrophic gastritis; IM, intestinal metaplasia; GC, gastric cancer.

Atrophic gastritis is an intermediate disease stage, from the viewpoint of gastric disease progression, that follows initial H. pylori infection. In the case of AG– group individuals with an LA value of less than 32.0 U/mL, two distinct disease stages are observed that occur before atrophic gastritis (healthy individuals and chronic gastritis cases) and after atrophic gastritis (intestinal metaplasia and gastric cancer cases). The AG– group included 31 gastric cancer patients (Fig. 4); however, it was difficult to distinguish these patients from noninfected healthy individuals because their LA values tended to be low and 50% tested negative for H. pylori infection (Fig. 3B, Supplementary Table 2) because of the progression of gastric mucosal alterations. Combining the serum anti-H. pylori antibody test and the PG test for gastric cancer screening has been a widely employed strategy throughout the world, especially in Japan.³⁴ Based on our findings from a previous Mongolian previous study,³⁵ we changed the cutoff value for the PG I/II ratio to less than 3.1 when predicting gastric cancer cases. We found that when the PG I/II ratio was set to less than 3.8 U/mL for the AG– group in this study population, 30 gastric cancer patients (96.7%) and 55 intestinal metaplasia cases were defined within group 2 in Fig. 4. Currently, only upper gastrointestinal endoscopic observations can distinguish disease status precisely, which means that individuals with a low LA value and H. pylori-negative cases require upper gastrointestinal endoscopy for confirmation of their gastric mucosal condition. Our study recommends that individuals with an LA value higher than 32.0 U/mL (group 1, atrophic gastritis suspicious group), and those with an LA value less than 32.0 U/mL and PG I/II ratio less than 3.8 U/mL (group 2, intestinal metaplasia suspicious group), should undergo upper gastrointestinal endoscopy for gastric cancer prevention.

Finally, for individuals with an LA value less than 32.0 U/mL and a high PG I/II ratio (≥3.8 U/mL), a second noninvasive H. pylori test, such as a stool antigen test, is recommended (group 3, second H. pylori test required). Of these group 3 individuals, 45% tested positive for H. pylori in this study, and therefore needed to undergo eradication to prevent the progression of disease. Considering these study outcomes, the role of serology in this LA assay was less concerned with H. pylori infection status and more with predicting gastric mucosal alterations using a noninvasive method for the prediction of disease status.

Concerning the high anti-H. pylori antibody values found in atrophic gastritis patients, we interpreted its reflection from gastric mucosal inflammation.³¹ On the other hand, from the intestinal metaplasia and gastric cancer stages, H. pylori commence to extinct; therefore, serum antibody decline may come from the reduction of H. pylori amount that leads to reduced inflammation process. When we compared four tests for diagnostic accuracy to predict atrophic gastritis, we found that the LA assay showed high specificity (73%); on the flip side, PG I and PG II showed high sensitivity (68% and 71%, respectively). PG I is synthesized by the chief cells and neck cells of the gastric corpus, while PG II is also synthesized widely in the cardiac, pyloric, and Brunner gland cells in the proximal duodenum.³² Active inflammation caused by H. pylori increases the blood levels of PGs; however, active inflammation causes atrophy of the corpus, damages PG I expressing cells, then related to the decreased PG I level.^33,36,37 In the same timing or earlier timing of reduction of PG I or PG I/II ratio, induction of high anti-H. pylori antibody value might occur along gastric disease progression.

As mentioned repeatedly, one of the limitations of the LA test is discrimination between current and past infection, the same as E-test. In past infection cases, the serum level of antibodies stays high for 1 to 4 years posttreatment.^13-15 This is disadvantage of both serum antibody tests. However, we believe that this unique memory of infection might be an advantage to understand of infection history in patients, even if H. pylori is negative in stool antigen test. This new type of combining LA values with PG I/II ratio (Fig. 4), may increase the gastric cancer and pre-gastric cancer screening efficiency in Mongolia, by accurately identifying individuals requiring upper gastrointestinal endoscopic examination.

Additionally, the sensitivity and specificity of the LA assay on H. pylori infection in Japan, have been reported to be 85%–95%, which is consistent with other countries that have carried out effective mass H. pylori detection.^16,17,38 The AUC, sensitivity, and specificity values in our study were 0.736, 74.2%, and 65.6%, which were similar to those reported for Bangladesh (0.746, 71.1%, and 55.6%) and Myanmar (0.830, 79.6%, and 72.4%) in our previous study.¹⁶ Furthermore, the LA values for the Mongolian healthy group (Hp–, with histological scores of zero using the Updated Sydney System) were relatively high (median 11.7 U/mL) compared with a previous study in seven Asian countries (median 1.2 to 8.4 U/mL for Hp– cases). This might indicate that high immunity against H. pylori is a characteristic feature of the Mongolian population, or it could suggest a high nonspecific reaction to antigen proteins from Japanese H. pylori immobilized on the surface to the latex beads. The study results were from the Mongolian population, and the results are not generally applicable to other areas, which is the limitation of this study. To address this latter possibility, we propose the development of a regional LA serological kit for use with the Mongolian population.

From the Global Burden of Disease-2019 data, Mongolia was highlighted as a country with high gastric cancer incidence and mortality rates of more than 43.7 and 40.04 per 100,000 individuals.³⁹ Therefore, to develop a preventive strategy for gastric cancer, it is important to determine the current H. pylori infection status and frequency of gastric mucosal alterations among the Mongolian population. Gastric atrophic changes are associated with a higher risk of developing gastric cancer compared with non-atrophic conditions.^24,40,41 However, widespread screening based on endoscopy is not feasible for a number of reasons, including anxiety about the procedure by patients, religious beliefs, financial constraints, current disease status, and the availability of specialized equipment and personnel. Therefore, multiple serologic and/or noninvasive tests may be crucial in identifying high-risk individuals requiring further investigation. Although it is not possible to define all of the characteristics of gastric mucosal changes based on noninvasive tests and clinical symptoms, the optimization of such tests may enable the accurate prediction of patients that require further endoscopic examination.

This research was funded by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan (18KK0266, 19H03473, 21H00346 and 22H02871 to Y.Y.), (17K09353, 21K07898 to J.K.) and (18K16182, 21K08010 to T.M.) and partially supported by the Research Center for GLOBAL and LOCAL Infectious Diseases, Oita University (2021B13). A.K. and B.S. were doctoral students supported by the Japanese Government (Monbukagakusho: MEXT) Scholarship Program for 2020 and 2019.

Supplementary materials can be accessed at https://doi.org/10.5009/gnl220464

No potential conflict of interest relevant to this article was reported.

Study concept and design: A.K., J.A., Y.Y. Data acquisition: A.K., J.A., B.S., B.G. Drafting of the manuscript: A.K. Critical revision of the manuscript for important intellectual content: J.A., Y.Y. Statistical analysis: A.K., J.A. Obtained funding: J.A., T.M., Y.Y. Administrative, technical, or material support, study supervision: A.K., B.S., B.G., D.A., K.O., D.D. Approval of the final manuscript: all authors.