Most cases of gastric cancer in East Asia are attributable to Helicobacter pylori infection. H. pylori-negative gastric cancers accounted for only 2.3% of the 1,833 Korean gastric cancer patients and 2.5% of the 666 Japanese gastric cancer patients.^1,2 During the past two decades, a continuous increase in H. pylori eradication rates in Korea led to a gradual decrease in H. pylori seropositivity from 66.9% (1998) to 41.5% (2017).³ In the younger population aged between 10 and 29 years, seropositive rates for anti-cytotoxin associated gene A immunoglobulin G decreased from 60.0%–85.0% (1994) to 12.5%–28.9% (2015).⁴ Nevertheless, the national guidelines still recommend biennial gastroscopy for all Koreans aged between 40 and 75 years.⁵ Hence, 8,462,570 participants (63.1%) underwent gastric cancer screening among the target population of 13,404,927 individuals in 2021.⁶

In H. pylori-seroprevalent populations, diagnostic criteria for naive status should be strict and based on histology, endoscopy, and serum pepsinogen (PG) assay findings.⁷ Naive condition should be diagnosed only when both noninvasive and invasive H. pylori tests show negative findings.⁸ Furthermore, there should be no intestinal metaplasia or atrophy on endoscopy, histology, and PG assay findings in H. pylori-naive participants.

This study aimed to determine whether current gastric cancer screening guidelines could be changed according to H. pylori infection status and year of birth. Furthermore, we aimed to identify types of gastric tumors according to the H. pylori infection status and year of birth. We hypothesized that if H. pylori-naive individuals are increasing in young participants for gastric cancer screening, biennial gastroscopy could be waived after the confirmation of naive status. We assumed that the year of birth is more important than the age itself because the proportion of H. pylori-naive individuals among Koreans born before 1983 (age of ≥40 in 2023) differs from that of Koreans born before 1959 (age of ≥40 in 1999 when biennial gastric cancer screening first began in Korea). Hence, these two groups should not be regarded as having similar gastric cancer risk, even in the same age group of ≥40.

1. Participants of the study

This was a single-center, long-term follow-up study on gastric cancer screening test findings that continued after our previous study registered at ClinicalTrials.gov (NCT01824953).⁹ Koreans who underwent gastric cancer screening using endoscopy and serum assays between 2010 and 2016 at our center were consecutively included. The participants were included if they underwent follow-up screening for ≥3 biennial gastric cancer screening tests up to May 2023. The participants were excluded if other H. pylori tests, including invasive tests and urea breath tests, were not performed on the day of the H. pylori serology test. Foreigners, renal failure patients, and those who previously underwent gastrectomy were also excluded.

This study was approved by the Institutional Review Board of Konkuk University Medical Center (IRB number: KUH1010478). Participants provided informed consent before the gastric cancer screening test, and all procedures were performed in accordance with the Helsinki Declaration.

2. Gastric cancer screening tests

Gastroscopy, H. pylori serology, and serum PG assays were performed after ≥12 hours of fasting. Using the latex-enhanced turbidimetric immunoassay (HBI Co., Anyang, Korea), serologically detected atrophy was diagnosed when the PG I/II ratio was ≤3.0 and the PG I level was ≤70 ng/mL as previously described.¹⁰ H. pylori serology was tested using two serum assays that are acceptable in Koreans.¹¹ Vidas H. pylori IgG (BioMérieux, Marcy-l’Etoile, France) was used until 2012. Thereafter, Chorus H. pylori IgG (DIESSE Diagnostica Senese, Siena, Italy) was used.

3. Confirmation of H. pylori infection status

Infection status was classified as H. pylori-naive, current, and past infection. Infection was diagnosed when at least one positive test finding was found among the ¹³C-urea breath tests and H. pylori invasive tests.¹² Invasive tests were done using the rapid urease test or histology with Giemsa staining as described in our previous study.¹³ With the aid of ¹³C-urea breath tests (POCone^Ⓡ; Otsuka Electronics Co., Ltd., Hirakata, Japan), a cutoff value of ≥2.5% was used for the diagnosis of H. pylori infection.¹⁴ If none of the test findings were positive, H. pylori eradication history, serologically detected atrophy (PG I ≤70 ng/mL and PG I/II ≤3.0), and intestinal metaplasia/atrophy were checked for the diagnosis of past infection. If all of these were absent, the H. pylori-naive status was confirmed.

4. Study outcomes and measurements

The primary outcome was measured by the infection status of H. pylori according to the year of birth. The proportions of naive, current, and past-infected participants were also measured among the participants born before 1950, between 1950 and 1960, between 1960 and 1970, and after 1970. The secondary outcome was measured by gastric tumor (cancer or adenoma) detection during the initial and follow-up screening tests. The incidence and types of gastric tumors were measured according to the year of birth and infection status.

5. Statistical analysis

Differences between the H. pylori-naive, infected, and past-infected groups were analyzed using analysis of variance with Bonferroni correction for continuous variables. Data are presented as mean and standard deviation. For continuous variables with asymmetrical distribution, data are presented as medians and ranges using the Kruskal-Wallis test. For categorical variables, the chi-square test with Bonferroni correction was used, and data are presented as the number of participants with proportion (%). For categorical variables with asymmetrical distribution, the Fisher exact test was used. Furthermore, correlation analysis was done to verify the link between the year of birth and H. pylori infection status, and data are presented as Pearson correlation coefficient (r). An r-value over 0.7 indicates a strong correlation, and r-values between 0.3 and 0.7 indicate moderate correlation. To verify independent risk factors for gastric tumor, logistic regression analysis was done. Data are presented as odds ratio (OR) with 95% confidence intervals (CIs) and standard error. A p-value less than 0.05 was considered statistically significant. Statistical analyses were performed using PASW statistics (version 28.0; IBM Corp., Armonk, NY, USA).

1. H. pylori infection status of the participants

Two thousand and two Koreans underwent screening tests for 95.5±28.4 months. One-thousand and three participants had current H. pylori infection, 743 had past infection, and 256 were H. pylori-naive (Fig. 1). The mean age and proportions of heavy alcohol drinkers, diabetes, hypertension, or coronary heart disease patients were significantly lower in the naive group than in the other two groups (Table 1). Correlation analysis revealed that H. pylori-naive status was correlated with high PG I/II ratios (r=0.360, p<0.001). Conversely, current infection status was correlated with low PG I/II ratios (r=–0.643, p<0.001) and PG II levels (r=0.704, p<0.001).

Figure 1. Study flow. A total of 8,652 participants underwent gastric cancer screening tests between 2010 and 2016 at our center. Foreigners, postgastrectomy patients, and patients without other Helicobacter pylori tests on the day of serology were excluded.

Table 1 . Baseline Characteristics and Initial Serum Assay Findings of the Study Participants.

Variable	All participants (n=2,002)	Helicobacter pylori-naive (n=256)	Past infection (n=743)	Current infection (n=1,003)
Age, yr	52.3±10.2	42.5±7.2*,†	54.4±9.7	53.2±9.8
Male sex	1,156 (58.6)	135 (52.7)	436 (58.7)	585 (58.3)
Year of birth	1959±10	1969±7*,†	1957±10	1958±10
≤1949	382 (19.1)	0	178 (24.0)	204 (20.3)
1950–1959	624 (31.2)	21 (8.2)	257 (34.6)	346 (34.5)
1960–1969	631 (31.5)	105 (41.0)	223 (30.0)	303 (30.2)
≥1970	365 (18.2)	130 (50.8)	85 (11.4)	150 (15.0)
Body mass index, kg/m2	24.2±3.0	23.3±3.3	24.4±3.0	24.3±3.0
Current cigarette smoking	344 (17.2)	53 (20.7)	127 (17.1)	164 (16.4)
Heavy alcohol drinking	126 (6.3)	10 (3.9)*,†	45 (6.1)	71 (7.1)
Comorbidities under medication
Diabetes	155 (7.7)	7 (2.7)*,†	56 (7.5)	92 (9.2)
Hypertension	430 (21.3)	23 (9.0)*,†	186 (25.0)	221 (22.0)
Coronary heart disease	85 (4.2)	4 (1.6)*,†	42 (5.7)	39 (3.9)
Cerebrovascular disease	21 (1.0)	1 (0.4)	9 (1.2)	11 (1.1)
Serum pepsinogen I level, ng/mL	60.8±27.5	51.0±15.6†	51.7±22.8†	67.0±29.9
Serum pepsinogen II level, ng/mL	15.5±9.7	8.4±3.0*,†	10.0±4.6†	21.5±9.8
Serum pepsinogen I/II ratio	4.6±1.8	6.2±1.3*,†	5.4±1.6†	3.5±1.3
Anti-H. pylori IgG titer using Vidas kit	1.60 (0.05–4.00)	0.34 (0.05–1.10)†	0.54 (0.10–1.99)†	3.79 (0.23–4.00)
Anti-H. pylori IgG titer using Chorus kit	20.1 (5.0–200.0)	5.1 (5.0–15.7)†	5.1 (5.0–65.8)†	199.6 (6.7–200.0)
Type of initial serology kit (Vidas:Chorus)	760:1,242	56:200	295:448	409:594

Data are presented as mean±SD, number (%), or median (range). Based on the National Institute for Alcohol Abuse and Alcoholism guideline, heavy alcohol drinking was defined as ≥15 drinks/week for men and ≥8 drinks/week for women. One drink was equal to 350 mL of beer (5% alcohol), 150 mL of wine (12% alcohol), and 40 mL of whiskey (40% alcohol)..

*Significantly different from the previously-infected group (p<0.05); ^†Significantly different from the H. pylori-infected group (p<0.05)..

2. Link between year of birth and H. pylori infection status

H. pylori-naive status was correlated with the recent year of birth (r=0.368, p<0.001). The mean year of birth in the H. pylori-naive group (1969±7) differed from those of the past-infected (1957±10, p<0.001) and infected (1958±10, p<0.001) groups. The naive condition was not found among the participants born before 1950 (Fig. 2).

Figure 2. Helicobacter pylori infection status and birth year. Proportions of H. pylori-naive status increased with a more recent birth year. Among the 382 participants born before 1950, there were no H. pylori-naive participants. There were only 21 H. pylori-naive participants (3.4%) among the 624 participants born between 1950 and 1959. The prevalence of participants-naive status was increased in participants born after 1960.

3. Gastric tumors detected in the initial screening and follow-up screening tests

Twenty-three gastric adenomas, 18 adenocarcinomas, and two neuroendocrine tumors were detected in 1.9% (14/743) of past-infected and 2.5% (25/1,003) of infected participants, including four infected participants with metachronous tumors (Table 2). Conversely, there was no gastric tumor observed among the 256 H. pylori-naive participants (p=0.007). Numbers of gastric tumors and patients in the H. pylori-naive, past-infected, and infected groups are summarized in Fig. 3. During the follow-up tests, 18 (1.8%) showed new gastric tumors among the 1,003 infected participants, including two advanced gastric cancer patients. Among the 743 past-infected participants, 11 (1.6%) showed new gastric tumors, including two patients with reinfection. All four metachronous tumors and two advanced gastric cancers were found in infected participants born before 1960.

Figure 3. Types of gastric tumors and birth years of all gastric tumor patients. A total of 43 gastric tumors were detected in 39 participants, including 4 with metachronous tumors. Among the participants with past infection, 3 individuals had 3 tumors detected during the initial tests, and another 11 individuals had 11 tumors detected during follow-up tests. Among the participants with H. pylori infection, 11 participants had 11 tumors detected during the initial tests. Thereafter, another 14 individuals with H. pylori infection had 14 newly detected tumors during follow-up tests. Furthermore, among the 11 individuals with H. pylori infection who were diagnosed with gastric tumors during the initial tests, 4 individuals had 4 newly detected metachronous tumors. EGC, early gastric cancer; NET, neuroendocrine tumor; AGC, advanced gastric cancer. *The four patients who developed metachronous gastric tumors are underlined.

Table 2 . Gastric Cancers and Adenomas Detected during the Initial or Follow-up Screening Tests.

Variable	All participants (n=2,002)	Helicobacter pylori-naive (n=256)	Past infection (n=743)	Current infection (n=1,003)
Follow-up, mo	95.5±28.4	96.4±29.2	95.8±27.6	95.0±28.4
Tumors detected in the initial tests	14	0*,†	3†	11
New tumors detected in the follow-up tests	29 Including 4 metachronous tumors	0*,†	11	18 Including 4 metachronous tumors
Total number of gastric tumor patients	39 (1.9)	0*,†	14 (1.9)	25 (2.5)
Gastric adenoma patients	23 (1.1)	0*,†	8 (1.1)	15 (1.5)
Early gastric cancer patients	16 (0.8)	0*,†	5 (0.7)	11 (1.1)
Advanced gastric cancer patients	2 (0.1)	0	0	2 (0.2)
Gastric neuroendocrine tumor patients	2 (0.1)	0	1 (0.1)	1 (0.1)

Data are presented as mean±SD or number (%)..

*Significantly different from the previously-infected group (p<0.05); ^†Significantly different from the H. pylori-infected group (p<0.05)..

4. Factors correlated with the occurrence of gastric tumors

Comparisons were made between the 39 participants with gastric tumor detection and 1,963 participants without gastric tumors. Gastric tumors were more common in participants with a remote year of birth, old age, male sex, low PG I levels, low PG I/II ratios, diabetes, hypertension, and coronary heart disease than in their counterparts (Table 3). Among the significant variables, remote year of birth, old age, male sex, and low PG I/II ratios predicted the risks of developing gastric tumors (Table 4).

Table 3 . Variables Related to Gastric Tumor Detection.

Variable	With gastric tumor (n=39)	Without gastric tumor (n=1,963)	p-value
Year of birth	1951±11	1959±10	<0.001
Age, yr	60.9±10.6	52.1±10.1	<0.001
Male sex	31 (79.5)	1,125 (57.3)	0.005
Helicobacter pylori-naive	0	256 (13.0)	0.007*
Past H. pylori infection	14 (35.9)	729 (37.1)	0.874
Current H. pylori infection	25 (64.1)	978 (49.8)	0.077
Pepsinogen I level, ng/mL	51.7±28.2	61.0±27.4	0.036
Pepsinogen II level, ng/mL	16.9±11.3	15.5±9.6	0.390
Pepsinogen I/II ratio	3.9±2.0	4.6±1.8	<0.001
Current cigarette smoking	6 (15.4)	733 (37.3)	0.466
Heavy alcohol drinking	10 (25.6)	440 (22.4)	0.716
Body mass index, kg/m2	24.4±2.6	24.2±3.0	0.608
Diabetes	8 (20.5)	147 (7.5)	0.008*
Hypertension	15 (38.5)	415 (21.1)	0.009
Coronary heart disease	5 (12.8)	80 (4.1)	0.023*

Data are presented as mean±SD or number (%)..

*Fisher exact test..

Table 4 . Logistic Regression Analysis for Significant Variables Related to Gastric Tumor Detection.

Significant variable	Univariate analysis			Multivariate analysis including birth year instead of age			Multivariate analysis including age instead of birth year
	OR (95% CI)	p-value		OR (95% CI)	p-value		OR (95% CI)	p-value
Year of birth	0.920 (0.891–0.950)	<0.001		0.936 (0.904–0.968)	<0.001		-
Age	1.091 (1.056–1.128)	<0.001		-			1.074 (1.036–1.112)	<0.001
Male sex	2.886 (1.320–6.312)	0.008		3.071 (1.387–6.802)	0.006		3.047 (1.376–6.744)	0.006
Helicobacter pylori-naive	0	0.994		-			-
Pepsinogen I/II ratio	0.689 (0.560–0.848)	<0.001		0.733 (0.594–0.903)	0.004		0.734 (0.596–0.903)	0.003
Diabetes	0.314 (0.142–0.695)	0.004		0.547 (0.238–1.261)	0.157		0.549 (0.238–1.266)	0.160
Hypertension	0.429 (0.223–0.825)	0.011		0.827 (0.408–1.675)	0.598		0.826 (0.408–1.672)	0.595
Coronary heart disease	0.289 (0.110–0.758)	0.012		0.565 (0.201–1.587)	0.278		0.560 (0.200–1.570)	0.270

OR, odds ratio; CI, confidence interval..

Among the significant variables listed in Table 3, the serum pepsinogen I level was omitted because it overlapped with the pepsinogen I/II ratio..

5. Differences between the patients with gastric adenoma, early gastric cancer, advanced gastric cancer, and gastric neuroendocrine tumor

Gastric adenoma patients showed more remote year of birth, older age, male predominance, lower PG I levels, lower PG I/II ratios, and more comorbidities (diabetes, hypertension, and coronary heart disease) than those without gastric adenoma (Supplementary Table 1). Among these variables, remote year of birth, male sex, and low PG I/II ratios were independent risk factors for gastric adenoma (Supplementary Table 2). Male sex (OR, 0.083; 95% CI, 0.011 to 0.637) and low PG I/II ratios (OR, 0.657; 95% CI, 0.472 to 0.917) were independent risk factors for early gastric cancer, whereas high PG II levels (OR, 1.115; 95% CI, 1.009 to 1.233) were risk factors for advanced gastric cancer.

The median PG I levels and PG I/II ratios were lower in gastric tumor patients with metachronous tumors than those without metachronous tumors (Supplementary Table 3). Moreover, male predominance was found in patients with metachronous tumors. Conversely, both gastric neuroendocrine tumors were found in female participants (p=0.031) in the initial tests (p=0.040).

H. pylori-naive participants are often overlooked in an H. pylori-seroprevalent population. Such an underestimation led to biennial gastroscopy in all Koreans after the age of 40, irrespective of H. pylori infection status. In this study, H. pylori-naive status was correlated with the most recent year of birth. Naive condition was increasing in younger participants born since 1960, and none of the naive participants showed gastric tumor during 95.5±28.4 months. Our findings are consistent with recent studies that showed the rarity of H. pylori-negative gastric cancers.^15,16

Notably, naive status was not found among the participants born before 1950 in this study. Moreover, it was rare in those born between 1950 and 1960. Most Koreans born before 1960 had current or past infection, and gastric tumors were detected in 2.5% of infected and 1.9% of past-infected participants. The findings are consistent with a recent study which confirmed that gastric tumor incidence is highest in the H. pylori-infected group, followed by those with past infection.¹⁷ In this study, advanced cancers and metachronous gastric tumors were found in infected patients during 8 years of follow-up. Furthermore, two previously-infected participants had reinfection at the time of new gastric tumor detection during the follow-up period. Altogether, eradicating H. pylori is necessary to decrease the risk of metachronous gastric tumors as well as to slow the rapid progression of gastric cancer to advanced stages.

The occurrence of gastric adenoma and early gastric cancer was related to the remote year of birth, male sex, and low PG I/II ratios in this study. Most gastric tumors in the past-infected group were slow-growing adenomas and early gastric cancers. These findings are supported by previous studies that showed irreversible damage in the corpus observed in eradicated patients with low PG I levels and PG I/II ratios.^18,19 Atrophic and metaplastic changes may reverse after successful eradication,^20,21 but not in severe cases.²² Hence, it is important to eradicate H. pylori before the development of extensive atrophy or intestinal metaplasia, which are important risk factors for developing gastric tumors.²³ On the other hand, the occurrence of advanced gastric cancer was related to high PG II levels, which indicate infected stomachs with a risk of diffuse-type gastric cancer.²⁴ With regard to gastric neuroendocrine tumors, none of the risk factors for adenomas and adenocarcinomas could predict the occurrence. Adding serum gastrin levels to serum PG assay would be useful for detecting high risk individuals for gastric neuroendocrine tumors.²⁵

There are limitations in this study. First, participants with a history of unintended eradication might have been included in the naive group, especially in those who experienced it at an earlier age. Nonetheless, our study findings are still valuable because gastric cancer is rare in past-infected participants who have recovered without sequelae on endoscopy, histology, and serum assay findings. Thus, the Maastricht VI/Florence consensus report recommends regular endoscopic surveillance only in stomachs with a high operative link on gastritis assessment stages after eradication confirmation.²⁶ In their recent study, operative link on gastritis assessment stages III-IV increased the risk of gastric epithelial neoplasia, whereas neoplastic lesions were not detected in H. pylori-naive patients.²⁷ Second, to overcome the birth cohort effect on gastric cancer incidence, participants should have been compared in the same age period. A proper comparison would be possible after 2033 when young participants born after 1970 are ≥63 years old, just like the older participants born before 1960 (who are ≥63 years old in this study).

In conclusion, the prevalence of H. pylori-naive status is increasing in younger Koreans, and gastric tumors are rare in this population. Therefore, biennial gastroscopy could be waived in young population after the confirmation of naive status. Infected participants should undergo eradication first, and biennial gastroscopy should be forced only to previously infected participants. Future gastric cancer screening guidelines should differ among naive, infected, and past-infected participants.

This work was supported by Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Education 2016R1D1A1B02008937.

S.Y.L. is an editorial board member of the journal but was not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.

Study concept and design: S.Y.L. Data acquisition: J.H.N., S.Y.L. Data analysis and interpretation: J.H.N., S.Y.L. Drafting of the manuscript: J.H.N., S.Y.L. Critical revision of the manuscript for important intellectual content: J.H.N., S.Y.L. Statistical analysis: J.H.N., S.Y.L. Obtained funding: S.Y.L. Administrative, technical, or material support; study supervision: S.Y.L., J.H.K., I.K.S., H.S.P. Approval of final manuscript: all authors.

Supplementary materials can be accessed at https://doi.org/10.5009/gnl230211.