Pepsinogen Assay Findings in Individuals With Gastric Atrophy and Intestinal Metaplasia
Article information
Abstract
Serum pepsinogen (PG) testing is widely used as a non-invasive biomarker for gastric atrophy; however, its clinical interpretation varies depending on the severity and distribution of gastritis as well as on the assay methods. Autoimmune gastritis is characterized by markedly decreased PG I levels and PG I/II ratios, along with increased gastrin levels, reflecting profound corpus atrophy. In contrast, Helicobacter pylori–associated atrophy typically shows moderate reductions in PG I levels and PG I/II ratios, with variability according to the extent of antral and corpus involvement. Accumulating evidence indicates that PG testing is more effective at identifying severe atrophy and higher Operative Link on Gastritis Assessment stages than detecting mild atrophy. Notably, cutoff values differ substantially between assay systems, with higher PG I and PG I/II cutoff values observed in GastroPanel tests than in East Asian kits interpreted with the ABC method. Overall, PG testing is useful for detecting advanced corpus atrophy and stratifying gastric cancer risk when combined with gastrin measurement, but has limited sensitivity for early-stage disease. Monitoring low PG I/II ratios and high gastrin levels may improve risk assessment, particularly in patients with corpus atrophy or persistent intestinal metaplasia after H. pylori eradication.
INTRODUCTION
Pepsinogen (PG) is a precursor of pepsin, a gastric protease, that is converted into pepsin to aid in food digestion [1]. Pepsinogen I (PG I) is made by the chief cells and mucous neck cells in the gastric corpus, and pepsinogen II (PG II) is made by the gastric oxyntic glands and the duodenal Brunner’s gland [2]. Low PG I levels correlate with the severity of atrophy and intestinal metaplasia (IM) in the gastric corpus, whereas high PG II levels correlate with active inflammation induced by Helicobacter pylori [3]. Therefore, gastroscopy is recommended when PG I levels or PG I/II ratios are low, especially in the absence of H. pylori infection [4]. In addition to low PG I levels and low PG I/II ratios, high gastrin levels help discriminate patients with gastric neuroendocrine tumors (NETs), adenomas, and adenocarcinomas [5].
Despite the widespread use of serum PG testing, its clinical interpretation remains challenging because of the variability in cutoff values across different assay systems and patient populations [6]. This inconsistency may lead to misclassification of high-risk individuals in clinical practice. Moreover, discrepancies between serologic findings and endoscopic or histologic atrophy may occur, especially in individuals with advanced or heterogeneous gastric mucosal changes. Such discrepancies highlight the need for cautious interpretation of PG results and underscore the complementary role of endoscopy. This review aims to not only summarize PG I levels, PG I/II ratios, and gastrin levels according to the severity and location of gastric atrophy and IM but also critically evaluate differences in diagnostic performance across assay methods and clinical conditions. Furthermore, practical considerations for interpreting PG test results in real-world settings are discussed, with their implications for gastric cancer risk stratification.
CORPUS-DOMINANT ATROPHY (TYPE A GASTRITIS)
Type A gastritis consists of autoimmune gastritis (AIG) that manifests as atrophy or IM in the corpus without involving the antrum. In H. pylori-endemic areas, AIG is diagnosed when serum anti-parietal cell antibodies (APCAs) or anti-intrinsic factor antibodies are positive in patients with corpus-dominant atrophy detected by endoscopy or microscopy [7]. During gastroscopy, transparent submucosal vessels are observed in the corpus without gastric folds and atrophic borders. Typical microscopic findings include loss of parietal cells, chief cells, and gastric glands in the corpus replaced by IM, pseudopyloric metaplasia, and enterochromaffin-like (ECL) cells [7].
PG assay findings in patients with AIG
Patients with AIG show lowest PG I levels and PG I/II ratios among all types of gastritis because of loss of chief cells (Table 1). In a Chinese study, most patients with AIG showed low PG I/II ratios of ≤1 and low PG I levels of <10 ng/mL [8]. In a Japanese study, PG I/II ratios <1.8, PG I levels <20.1 ng/mL, and gastrin levels >355 pg/mL were useful for discriminating patients with AIG [9]. A study from Türkiye reported that PG I/II ≤1.9 and PG I ≤13.5 ng/mL detected AIG with high sensitivity and specificity [10]. Compared with Japanese kits, GastroPanel tests (Biohit Oyj) yield higher cutoff values for PG I/II ratios (≤3.1 vs. ≤2.33) [11]. The only exception is a study from Russia showing a low PG I/II cutoff ratio of ≤2 for detecting AIG even with the use of GastroPanel tests [12]. In that study, 80.2% of the participants had H. pylori infection; hence, PG I/II ratios were further decreased because of increased PG II levels.
GastroPanel tests generally yield twice higher PG I levels than East Asian kits interpreted with the ABC method, leading to different thresholds for defining atrophy [13]. A study reported that advanced corpus atrophy (PG I/II ≤2 and PG I ≤30 ng/mL) measured by a latex-enhanced turbidimetric immunoassay (LTIA) kit (HBI Co.) was more consistent with serologic atrophy determined by the GastroPanel test than corpus atrophy (PG I/II ≤3 and PG I ≤70 ng/mL) determined by the ABC method [14]. Similarly, another Korean study revealed that a criteria using PG I/II ≤3 and PG I ≤70 ng/mL measured by the L-TIA kit was consistent with a combination of PG I/II ≤5.3 and PG I ≤100 ng/mL measured by GastroPanel tests [15]. These differences may arise from different calibration methods used in PG assays [16].
Test findings in patients with AIG according to H. pylori infection status
Patients with both AIG and H. pylori infection exhibit higher PG II levels and lower PG I levels than those without H. pylori infection [17]. Compared with H. pylori–naïve patients with AIG, H. pylori–exposed patients with AIG show lower gastrin levels because of atrophy in the antrum [18]. Notably, H. pylori infection status is often confused in patients with AIG because non–H. pylori, urease-producing bacteria are increased because of high gastric pH, which leads to false-positive test findings in the rapid urease tests and urea breath tests [19]. Furthermore, spontaneous regression of H. pylori may occur in AIG because of severe atrophy and IM.
Diagnosis of AIG in H. pylori–infected stomachs is challenging owing to full-thickness inflammation that conceals AIG (Fig. 1). Early-stage AIG may start with autoimmune activity following H. pylori infection, and APCA plays an important role in progression of corpus atrophy [20]. Conversely, some autoimmune activities are suppressed because H. pylori inhibits CD4+ T helper 1 (Th1) cells [21]. In that study, suppression of Th1-mediated AIG development was terminated after H. pylori eradication, allowing autoreactive Th1 cells to produce autoantibodies that attack parietal cells. As a result, some cases of AIG were detected after H. pylori eradication.
Autoimmune gastritis masked by H. pylori–associated inflammation. A: Diffuse redness in the fundus and upper body indicates H. pylori infection. B: Thick gastric folds and diffuse redness are observed in the greater curvature side of the body. C: The antrum has irregular mucosal nodularity, indicating metaplastic gastritis. D: One year after successful H. pylori eradication, diffuse redness is replaced by atrophy. Serologic atrophy (PG I: 15.5 ng/mL, PG II: 17.4 ng/mL, and PG I/II: 0.9) was reported. E: Atrophy is observed without gastric folds. APCA revealed positive test finding. F: There are no notable findings in the antrum. APCA, anti-parietal cell antibodies; PG, pepsinogen.
Gastrin levels and gastric cancer risk in patients with AIG
H. pylori gastritis is not rare entity in patients with AIG who develop gastric cancer [22]. In a recent study, gastric adenocarcinoma was 15.6 times more common in patients with gastric NETs with ≥3 risk factors, such as severe corpus atrophy, micronodular ECL hyperplasia, positive H. pylori serology findings, low hemoglobin levels, low vitamin B12 levels, low ferritin levels, and high gastrin levels [23]. Patients with AIG and H. pylori gastritis with high gastrin levels require gastroscopic surveillance because parietal cell loss leads to overproduction of gastrin, which binds to the cholecystokinin-2 receptor to stimulate parietal cells and proliferate ECL cells [24]. Avoiding long-term use of potassium-competitive acid blockers (PCABs) and proton pump inhibitors (PPIs) is important in patients with gastric atrophy because these drugs further stimulate gastrin production [25].
ANTRUM-DOMINANT ATROPHY (TYPE B GASTRITIS)
Type B gastritis consists of atrophy and IM initiated by H. pylori infection, which starts from the opposite side of AIG [26]. When atrophy is limited in the antrum, PG I levels are stable, but PG I/II ratios decrease because of high PG II levels [27]. In addition, gastrin-17 (G-17) levels are low because of the loss of G cells in the antrum.
PG levels for detecting the severity of atrophy based on the updated Sydney system
Studies that reported cutoff values for detecting moderate-to-severe atrophy are summarized in Table 2. For detecting more severe atrophy in the corpus, a combination of PG I/II ≤2 and PG I ≤30 ng/mL showed good diagnostic utility with the aid of Japanese assay kits [28]. Similar to the findings in patients with AIG, East Asian kits showed lower PG I/II cutoff ratios than did GastroPanel tests (≤2.0 vs. <3) [29].
Studies on cutoff values for detecting the severity of gastric atrophy based on the updated Sydney system or the operative link on the gastritis assessment stage
In two large-scale studies, patients with severe atrophy had lower PG I/II ratios and PG I levels than those with mild-to-moderate atrophy [30,31]. Sensitivities and specificities for detecting mild-to-moderate atrophy were unsatisfactory in both studies; therefore, PG assay is better suited for detecting patients with severe atrophy than those with mild or moderate atrophy.
PG levels for detecting the severity of atrophy based on the Operative Link on Gastritis Assessment stage
Similar to the updated Sydney system, PG assays perform better in detecting higher Operative Link on Gastritis Assessment (OLGA) stages with severe atrophy (Table 2). A Russian study reported that either PG I/II ≤2 or PG I <22.5 μg/L is useful for detecting OLGA stages III–IV in patients with AIG [12]. A Chinese study reported that PG I/II <2.08 is useful for detecting OLGA stages II–IV [32]. Studies from Nepal and Myanmar showed lower PG I/II cutoff values for detecting OLGA stages ≥II than for detecting OLGA stage ≥I [33,34]. Given the lower sensitivities and specificities for detecting OLGA stage I reported in those studies, PG assays should be used for detecting OLGA stages II–IV.
PG levels for detecting the severity of atrophy based on the Kimura–Takemoto classification
During gastroscopy, the extent of gastric atrophy can be categorized from closed-type 0 (C-0, no atrophy) to Op (opentype pan-atrophy) according to the Kimura–Takemoto classification [35]. A study from Vietnam revealed that a combination of PG I/II ≤4.6 and PG I ≤69.0 ng/mL helped identify individuals with ≥C-3 atrophy [36]. In their study, the sensitivities and specificities were higher when the Kimura–Takemoto classification was used instead of the OLGA stage. In Korea, PG I/II <3.2 could be used to detect closed-type atrophy; however, the sensitivity and specificity were too low to apply in daily practice [37]. In a Japanese study, a lower PG I/II cutoff ratio was needed for detecting open-type atrophy compared with that for closed-type atrophy (<2.9 vs. <3.9) [38]. Taken together, PG assay showed higher accuracy for detecting open-type atrophy than closed-type atrophy.
PG levels after successful H. pylori eradication
Although PG assay is not a confirmatory test, a decrease in PG II levels and an increase in PG I/II ratios suggest successful H. pylori eradication [39]. Reportedly, an increase in the PG I/II ratio was more profound in patients with severe corpus atrophy [40]. Patients with H. pylori infection and PG I/II ratios <3.0 showed ≥40% increase in PG I/II ratio after eradication, whereas those with PG I/II ratio between ≥3.0 and <5.0 showed ≥25% increase. Moreover, patients with PG I/II ratios ≥5.0 showed only ≥10% increase after successful eradication. These changes were mostly observed within 1–2 years after eradication.
PG levels after reinfection or new infection
Seroconversion of anti-H. pylori IgG titer may occur in seronegative individuals; however, only 9.8% have true H. pylori infection [41]. Compared with patients without true infection, those with new infection or re-infection had higher PG II levels and lower PG I/II ratios in that study. In Koreans, PG I/II ratios <4.35 and PG II levels >12.95 ng/mL suggest H. pylori infection [42]. In a previous study, compared with 256 H. pylori–naïve individuals, 1003 patients with infection and 743 with past infections had lower PG I/II ratios and higher incidence of gastric cancer [43]. In that study, H. pylori–naïve individuals showed mean PG I level of 51.0±15.6 ng/mL, PG II level of 8.4±3.0 ng/mL, and PG I/II ratio of 6.2±1.3.
PAN-ATROPHY UNRELATED TO AIG
Antrum-dominant atrophy may progress to pan-atrophy if H. pylori is not eradicated at an early stage [44]. In patients with pan-atrophy and no clear history of H. pylori eradication, spontaneous regression or unintended eradication at a later stage can be suspected. Despite the severe atrophy observed on gastroscopy, patients with pan-atrophy may have normal PG test findings (Fig. 2). Different from AIG, such discrepancies between the endoscopic and serologic test findings are not rare in H. pylori–induced pan-atrophy. Therefore, H. pylori–related pan-atrophy should be considered if PG assay findings are within the normal range in patients with severe atrophy on gastroscopy (Fig. 3).
Pan-atrophy in H. pylori–exposed stomach unrelated to autoimmune gastritis. A: Atrophy is observed without a gastric fold. B: A flat and raised subepithelial lesion is seen on the greater curvature side of the upper body. Primary low-grade, MALT lymphoma was diagnosed using the biopsied specimens. The Giemsa staining result was positive, and H. pylori was eradicated. C: Atrophy is observed in the antrum. D: Four years later, the atrophic mucosa looks whitish and pale compared with that in the initial test findings. Despite severe endoscopic atrophy, there was no serologic atrophy (PG I: 34.2 ng/mL, PG II: 9.3 ng/mL, and PG I/II: 3.7). E: After MALT lymphoma regression, the gastric body is replaced by atrophy. Both APCA and AIFA testing returned negative results. F: The antrum exhibits atrophy and IM, with small erosions. MALT, mucosa-associated lymphoid tissue; IM, intestinal metaplasia; APCA, anti-parietal cell antibody; AIFA, anti-intrinsic factor antibody; PG, pepsinogen.
Brief interpretation of serum pepsinogen and gastrin levels. PG, pepsinogen; APCA, anti-parietal cell antibody; AIFA, anti-intrinsic factor antibody; AIG, autoimmune gastritis; NET, neuroendocrine tumor; IM, intestinal metaplasia; PPI, proton pump inhibitor; PCAB, potassium-competitive acid blocker.
PG levels in patients with atrophy in the corpus
Several studies have reported cutoff values for detecting corpus atrophy without testing autoimmune antibodies for AIG (Table 3). High gastrin levels, low PG I levels, and low PG I/II ratios were useful for detecting corpus atrophy in those studies. A study from Iran revealed that either G-17 <2.6 pmol/L or >40 pmol/L was useful for detecting atrophy in the fundus [45]. As shown in a previous study, G-17 levels can decrease even in the presence of corpus atrophy if there is an advanced atrophy in the antrum [5].
Studies using Biohit’s enzyme-linked immunosorbent assay (ELISA) kits provided a larger range of PG I/II cutoff ratios (from ≤2.2 to ≤10.0) than did Eiken’s ELISA kits and other assays (Fig. 4). A study from Chile showed a low PG I/II cutoff value of ≤2.2 even with the use of GastroPanel tests [46]. The majority of participants in that study (79.7%) had H. pylori infection; as a result, PG I/II ratios were further reduced due to elevated PG II levels in patients with infection.
PG levels in patients with IM
Few studies have reported cutoff values for detecting IM (Table 4). Two studies revealed that PG I/II levels <3.1 and 3.25 are useful for detecting IM when Japanese kits are used [47,48]. In the later study, the PG assay showed better performance for detecting IM than for detecting mild and moderate atrophy. Conversely, GastroPanel tests revealed higher cutoff values of PG I/II ≤10.0 and PG I/II <6.0 for detecting IM with low accuracy [49,50].
A study from Iran reported that a combination of PG I/II ratios ≤3 and G-17 levels >17 pmol/mL is useful for detecting IM in the corpus [51]. Despite its high specificity (97.8%), the sensitivity was too low (14.8%) for application in clinical practice. Considering the high H. pylori infection rate (78.7%) in their study, the PG assay seems to be less accurate for detecting IM, particularly in the presence of H. pylori infection.
Gastrin levels and gastric cancer risk after eradication
High gastrin levels and severe atrophy are common in patients who develop metachronous gastric cancer [52]. In a previous study, PPI-independent elevation of gastrin levels over 99 pg/mL (sensitivity 61.6%, and specificity 80.0%) correlated with the occurrence of metachronous gastric cancers. Other two studies also recommended to avoid using PCABs or PPIs after eradication because these drugs increased the risk of metachronous gastric cancers [53,54]. Taken together, H. pylori–exposed individuals with corpus atrophy or persistent IM should be monitored for serum gastrin levels when using these drugs [55]. Furthermore, serum PG testing should not be used as a standalone diagnostic tool but as a triage biomarker in gastric cancer screening, as summarized in Table 5.
CONCLUSIONS
Serum PG testing provides valuable information on gastric mucosal status, particularly for identifying AIG and advanced corpus atrophy (OLGA stages II–IV) in the absence of H. pylori infection. However, its diagnostic performance is limited in patients with mild atrophy, OLGA stage I, and closed-type atrophy. AIG is characterized by the most profound reductions in PG I/II ratios and PG I levels, whereas H. pylori–associated atrophy shows more variable patterns depending on disease distribution. The limitations of PG testing are particularly evident in patients with pan-atrophy, in whom serologic findings may remain within normal ranges despite severe endoscopic atrophy. Thus, careful interpretation in the appropriate clinical context is essential because H. pylori infection, recent medications, impaired renal function, dietary factors, age, sex, and genetic backgrounds may further influence PG levels. Furthermore, variability in cutoff values across different assay systems should be considered when using GastroPanel tests instead of East Asian kits interpreted with the ABC method. Future studies should focus on standardizing assay-specific cutoff values and developing integrated diagnostic algorithms that combine serologic, endoscopic, and clinical parameters to improve risk prediction and guide surveillance strategies. In summary, monitoring low PG I/II ratios and high gastrin levels is recommended during gastric cancer surveillance, particularly in patients with corpus atrophy or persistent IM even after H. pylori eradication.
Notes
Availability of Data and Material
All data generated or analyzed during the study are included in this published article.
Conflicts of Interest
The author has no financial conflicts of interest.
Funding Statement
This work was supported by Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Education (NR017128).
Acknowledgements
An artificial intelligence (AI) tool (ChatGPT5.3, OpenAI, San Francisco, CA, USA) was used for editing. There is no AI-generated data or figures. All scientific contents and interpretations were created and validated by the author. Furthermore, the document was edited for proper English language, grammar, punctuation, spelling, and overall style by the highly qualified English-speaking editors at Springer Nature Author Services.
