- Email: [email protected]
The optimal serum pepsinogen cut-off value for predicting histologically confirmed atrophic gastritis
Digestive and Liver Disease 47 (2015) 663–668
Contents lists available at ScienceDirect
Digestive and Liver Disease journal homepage: www.elsevier.com/locate/dld
Alimentary Tract
The optimal serum pepsinogen cut-off value for predicting histologically confirmed atrophic gastritis Eun Hye Kim a , Huapyong Kang a , Chan Hyuk Park b , Hee Seung Choi c , Da Hyun Jung a , Hyunsoo Chung a , Jun Chul Park a , Sung Kwan Shin a , Sang Kil Lee a , Yong Chan Lee a,∗ a Division of Gastroenterology, Department of Internal Medicine, Severance Hospital, Institute of Gastroenterology, Yonsei University College of Medicine, Seoul, Republic of Korea b Department of Internal Medicine, Guri Hospital, Hanyang University College of Medicine, Guri, Republic of Korea c Department of Pathology, Yonsei University College of Medicine, Seoul, Republic of Korea
a r t i c l e
i n f o
Article history: Received 24 January 2015 Accepted 15 May 2015 Available online 23 May 2015 Keywords: Atrophic gastritis Cut-off value Helicobacter pylori Pepsinogen
a b s t r a c t Background: Although serum pepsinogen tests are useful for predicting the presence of atrophic gastritis, the optimal cut-off values have not been fully evaluated. Aim: To determine the optimal serum pepsinogen cut-off value for predicting atrophic gastritis. Methods: Patients scheduled for upper endoscopy at Severance Hospital, Korea, between August 2012 and October 2013, were recruited prospectively. Endoscopic biopsies for atrophic gastritis were obtained and histologically graded, based on the updated Sydney system. Results: Ninety-five patients were enrolled in the study. The mean age was 57.7 ± 12.1 years, and 44.2% of the patients were male. Serum pepsinogen I/II ratios were lower in patients with atrophic gastritis than in those without it (antrum, 4.2 ± 1.7 vs. 5.2 ± 2.1, P = 0.040; corpus, 3.3 ± 1.9 vs. 5.4 ± 1.9, P < 0.001). Serum pepsinogen I/II ratios were significantly correlated with histologic atrophic gastritis (antrum, P = 0.030; corpus, P < 0.001). Using a cut-off value of 4.9, the sensitivity and specificity of the serum pepsinogen I/II ratio for predicting atrophic gastritis in the antrum were 68.2% and 60.3%, respectively. Conclusion: The optimal serum pepsinogen I/II ratio cut-off values for atrophic gastritis of the antrum and for the corpus were 4.9 and 3.5, respectively. Serum pepsinogen I/II ratios, with these cut-off values, are useful for screening patients for the presence of atrophic gastritis. © 2015 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.
1. Introduction Human pepsinogen (PG) I and II are pepsin (an endoproteinase present in the gastric juice. PG I is secreted mainly from the chief cells in fundic mucosa [1]) proenzymes, whereas PG II is secreted from the cardiac, fundic, antral mucosa of the stomach and proximal duodenal mucosa [2,3]. Serum PG I and II levels and the PG I/II ratio may be related to the histologic and functional status of the gastric mucosa [4]. Previous studies have shown that serum PG I levels and PG I/II ratios reflect the morphological and functional status of the gastric mucosa [5–8]. Based on these characteristics, PG tests can be used to predict the presence of atrophic gastritis.
∗ Corresponding author at: Division of Gastroenterology, Department of Internal Medicine, Severance Hospital, Institute of Gastroenterology, Yonsei University College of Medicine, 50-1 Yonsei-ro, Seodaemun-gu, Seoul 120-752, Republic of Korea. Tel.: +82 2 2228 1960; fax: +82 2 393 6884. E-mail address: [email protected] (Y.C. Lee).
Predicting the presence of atrophic gastritis using serum PG levels is important for the following two reasons. First, blood sampling to determine PG levels is much simpler and less invasive than endoscopic biopsy, the gold standard for diagnosing atrophic gastritis. Second, atrophic gastritis is a well-known risk factor for gastric cancer [9–12]. Therefore, PG tests can be used for mass screening of individuals for the presence of heightened gastric cancer risks. In Japan, gastric cancer screening is based on serum immunoglobulin G (IgG) anti-Helicobacter pylori antibody and serum PG levels (the ABCD method), and is used for gastric cancer risk stratification [13,14]. In previous studies, serum PG status was usually defined as “atrophic” when the serum PG I level was ≤70 ng/mL, and the PG I/II ratio was simultaneously ≤3.0, as proposed by Miki et al. [8,14]. However, Miki et al. determined serum PG cut-off values based on endoscopically diagnosed atrophic gastritis rather than histologically confirmed atrophic gastritis [8]. Therefore, the best serum PG cut-off value for predicting histologic atrophic gastritis may differ from the value proposed by Miki et al. A previous study suggested
http://dx.doi.org/10.1016/j.dld.2015.05.014 1590-8658/© 2015 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.
664
E.H. Kim et al. / Digestive and Liver Disease 47 (2015) 663–668
that the best cut-off values for fundic atrophy were serum PG I levels <56 ng/mL and a serum PG I/II ratio <5; those for the antrum and the corpus were not investigated [15]. Considering that both atrophic gastritis and gastric cancer usually develop at the antrum or corpus, identification of the optimal cut-off values for atrophic gastritis in the antrum or corpus may be more important. In addition, further investigation of the relationship between serum PG levels and intestinal metaplasia may be meaningful because intestinal metaplasia is also a well-known risk factor for gastric cancer [16]. Therefore, we evaluated whether serum PG tests can be used to detect atrophic gastritis, based on the anatomic location of the lesions (antrum or corpus). Then, we proposed optimal serum PG cut-off values for predicting atrophic gastritis. We also investigated the usefulness of PG tests for predicting intestinal metaplasia. 2. Materials and methods 2.1. Study design Patients scheduled for upper endoscopy at Severance Hospital, Seoul, Korea, between August 2012 and October 2013, were prospectively recruited for this study. Patients with a history of gastric surgery or prior H. pylori eradication therapy were excluded. Endoscopic biopsies were performed based on the updated Sydney system [17]. The following data were included in this study: age; sex; serum levels of PG I, PG II, and gastrin; and the PG I/II ratio. This study was approved by the Institutional Review Board of Severance Hospital (IRB 4-2013-0722) and written informed consent was obtained from each participant. This study was registered at http://clinicaltrials.gov (registration number NCT02184910). 2.2. Histologic evaluation Biopsy specimens were fixed in formalin, and assessed for the presence of H. pylori (using modified Giemsa staining), and for degree of inflammatory cell infiltration, atrophy, and intestinal metaplasia (all determined following haematoxylin and eosin staining). Additional specimens from the lesser curvature of the antrum were subjected to the rapid urease test (CLO test, Delta West, Bentley, WA, Australia). The histologic features of the gastric mucosa were recorded using the updated Sydney system scores: 0 = none, 1 = mild, 2 = moderate and 3 = severe. All histologic examinations were performed by one experienced pathologist, unaware of patient details. H. pylori infection was assessed using pathology or the rapid urease test. 2.3. Serological tests Fasting serum samples were collected from the participants upon entry into the study. The PG I and PG II levels were measured using latex agglutination turbidimetric immunoassay kits (HiSens Pepsinogen-I LTIA and HiSens Pepsinogen-II LTIA, HBI, Anyang, Korea). Serum gastrin levels were measured using a radioimmunoassay method (Siemens, Los Angeles, CA, USA). 2.4. Statistical analysis Continuous variables are presented as means with standard deviations, and were compared using the t-test. Categorical variables are presented as sample numbers with percentages, and were compared using the 2 or Fisher’s exact tests. Correlations between histologic feature severity and serologic tests were determined using the Spearman rank correlation test. To determine the serum PG cut-off values, receiver operating characteristic (ROC) curves and Youden’s index were used. Finally, analyses of variance were
Table 1 Baseline patient characteristics. Variable
Value
Patient, n Mean age ± SD (years) Male, n (%) Positive of rapid urease test, n (%) PG I, ng/ml, mean ± SD PG II, ng/ml, mean ± SD PG I/PG II, mean ± SD Gastrin, pg/dl, mean ± SD Atrophic gastritis of antrum, n (%) None Mild Moderate Atrophic gastritis of corpus, n (%) None Mild Moderate Intestinal metaplasia of antrum, n (%) None Mild Moderate Severe Intestinal metaplasia of corpus, n (%) None Mild Moderate Severe
95 57.7 ± 12.1 42 (44.2) 23 (24.2) 62.4 ± 41.7 14.3 ± 10.5 5.0 ± 2.1 86.0 ± 179.2 73 (76.8) 18 (18.9) 4 (4.2) 76 (80.0) 16 (16.8) 3 (3.2) 67 (70.5) 20 (21.1) 5 (5.3) 3 (3.2) 86 (90.5) 4 (4.2) 3 (3.2) 2 (2.1)
PG, pepsinogen; SD, standard deviation
used to compare the mean levels of the serologic tests according to the H. pylori infection status and the presence of atrophic gastritis. A P-value <0.05 was considered significant for group comparisons. All statistical analyses were performed using the statistical software programme SPSS for Windows (version 18.0; SPSS, Chicago, IL, USA). 3. Results 3.1. Baseline characteristics The baseline characteristics of the 95 enrolled patients are shown in Table 1. The mean age was 57.7 ± 12.1 years, and 44.2% of the patients were male. The proportion of positive rapid urease tests was 24.2%. The mean serum levels of PG I and PG II, and the PG I/II ratio were 62.4 ng/mL, 14.3 ng/mL and 5.0, respectively. The mean serum gastrin level was 86.0 pg/dL. Atrophic gastritis of the antrum and the corpus was identified in 22 (23.2%) and 19 (20.0%) patients, respectively. In addition, intestinal metaplasia of the antrum and the corpus was shown in 28 (29.5%) and 9 (9.5%) patients, respectively. 3.2. Relationships between serum levels of PG or gastrin and histologic feature severities The correlations between the serum PG or gastrin levels and the histologic feature severity, based on the semi-quantitative scores in the antrum and the corpus, are summarized in Table 2. Although serum PG I levels were correlated with the neutrophilic infiltration severity in the antrum (P = 0.014), such a correlation was not observed in the corpus (P = 0.733). Additionally, a significant correlation was not observed between the serum PG I levels and the severity of lymphoplasmacytic infiltration in either the antrum or the corpus (P = 0.216 and P = 0.574, respectively). In contrast, serum PG II levels showed significant correlation with the severity of both the lymphoplasmacytic infiltration and the neutrophilic infiltration, regardless of location (lymphoplasmacytic infiltration, antrum and corpus, both P < 0.001; neutrophilic infiltration, antrum and
E.H. Kim et al. / Digestive and Liver Disease 47 (2015) 663–668
665
Table 2 Correlation between the serum levels of pepsinogen or gastrin and the severity of histologic features. Variable
PG I
PG II
PG I/II ratio
Gastrin
Correlation efficiency
P-value
Correlation efficiency
P-value
Correlation efficiency
P-value
Correlation efficiency
Antrum Lymphoplasmacytic infiltration Neutrophilic infiltration Atrophic gastritis Intestinal metaplasia
0.128 0.250 0.128 −0.193
0.216 0.014 0.218 0.060
0.445 0.537 0.244 −0.060
<0.001 <0.001 0.017 0.562
−0.539 −0.527 −0.223 −0.168
<0.001 <0.001 0.030 0.104
0.216 0.154 −0.064 −0.020
0.036 0.139 0.541 0.851
Corpus Lymphoplasmacytic infiltration Neutrophilic infiltration Atrophic gastritis Intestinal metaplasia
−0.058 −0.035 −0.180 −0.225
0.574 0.733 0.081 0.028
0.413 0.412 0.158 −0.206
<0.001 <0.001 0.127 0.045
−0.645 −0.620 −0.387 −0.050
<0.001 <0.001 <0.001 0.627
0.418 0.361 0.221 0.079
<0.001 <0.001 0.032 0.452
corpus, both P < 0.001). Additionally, the serum PG I/II ratio showed an inverse correlation with both the severity of lymphoplasmacytic infiltration and the severity of neutrophilic infiltration, regardless of location (lymphoplasmacytic infiltration, antrum and corpus, both P < 0.001; neutrophilic infiltration, antrum and corpus, both P < 0.001). In patients with severe atrophic gastritis, serum PG I levels were not significant factors in either the antrum or the corpus (P = 0.218 and P = 0.081, respectively). Serum PG II levels correlated well with atrophic gastritis severity in the antrum (P = 0.017), but not in the corpus (P = 0.127). However, the serum PG I/II ratio showed a significant, inverse correlation with atrophic gastritis severity in both the antrum and the corpus (P = 0.030 and P < 0.001, respectively). The scatter plots show the distribution of serum PG I/II ratios, according to histologic feature severity (Fig. 1). However, the severity of intestinal metaplasia in the corpus was inversely correlated with both serum PG I and serum PG II levels (P = 0.028 and P = 0.045, respectively), but not with the serum PG I/II ratio (P = 0.627). 3.3. Serum levels of PG and gastrin according to the atrophic gastritis or intestinal metaplasia To evaluate whether the PG or gastrin levels can be used to predict the presence of atrophic gastritis or intestinal metaplasia, the mean serum levels were compared, based on the confirmed presence of atrophic gastritis or intestinal metaplasia (Table 3). Serum PG I levels did not differ between patients with and without atrophic gastritis, regardless of the anatomic location of the lesions. Although PG II serum levels were higher in patients with atrophic gastritis of the antrum than in those without it (20.1 ± 14.9 vs. 12.5 ± 8.0 ng/mL, P = 0.030), there were no differences in serum PG II levels, relative to the presence of atrophic gastritis in the corpus. The serum PG I/II ratio, however, was significantly lower in patients with atrophic gastritis than in those without, regardless of anatomic location (atrophic gastritis vs. no atrophic gastritis; antrum, 4.2 ± 1.7 vs. 5.2 ± 2.1 ng/mL, P = 0.040; corpus, 3.3 ± 1.9 vs. 5.4 ± 1.9 ng/mL, P < 0.001; diffuse, 3.4 ± 1.6 vs. 5.2 ± 2.1 ng/mL, P = 0.009). In patients with intestinal metaplasia, no significant differences in serum PG or gastrin levels were observed, according to the presence of intestinal metaplasia, except among the serum PG II levels in patients with corpus intestinal metaplasia or diffuse intestinal metaplasia (corpus, 7.9 ± 2.6 vs. 14.9 ± 10.8 ng/mL, P < 0.001; diffuse, 7.5 ± 2.8 vs. 14.8 ± 10.7 ng/mL, P < 0.001). 3.4. Receiver operating characteristics curve of PG and gastrin for predicting atrophic gastritis or intestinal metaplasia As shown in Fig. 2, the serum PG I/II ratio was useful for predicting the existence of atrophic gastritis in the antrum,
P-value
corpus, and both of them (area under the receiver operating characteristics [AUROC] (95% confidence interval [CI]): antrum, 0.650 [0.524–0.775]; corpus, 0.781 [0.658–0.903]; diffuse, 0.761 [0.613–0.910]). The optimal serum PG I/II ratio cut-off values, derived from the Youden’s index, were 4.9, 3.5, and 3.5 for atrophic gastritis in the antrum, corpus, and both, respectively. Using a 4.9 cut-off value, the sensitivity and specificity of the serum PG I/II ratio for predicting the presence of atrophic gastritis in the antrum were 68.2% and 60.3%, respectively. With a 3.5 cut-off value, the sensitivity and specificity of serum PG I/II ratio for predicting the presence of atrophic gastritis of the corpus were 68.4% and 81.6%, respectively. Additionally, with a cut-off value of 3.5, the sensitivity and specificity of the serum PG I/II ratio for predicting the presence of diffuse atrophic gastritis were 70.0% and 80.0%, respectively. When Miki et al.’s criteria for atrophic gastritis (PG I, ≤70 ng/mL; and PG I/II ratio, ≤3.0) were applied in this study population, the sensitivity and specificity for predicting the presence of atrophic gastritis in the antrum were 22.7% and 87.7%, respectively; and those for predicting the presence of atrophic gastritis in the corpus were 42.1% and 92.1%, respectively. In contrast to predicting atrophic gastritis, the serum PG I, PG II, and gastrin levels, and the PG I/II ratio did not predict intestinal metaplasia, except that serum PG I levels were predictive of the presence of intestinal metaplasia in the corpus (AUROC [95% CI], 0.721 [0.555–0.887]). With a cut-off value of 44.2 ng/mL, the associated serum PG I level sensitivity and specificity were 77.8% and 65.1%, respectively.
3.5. Effect of H. pylori infection and atrophic gastritis on serum levels of PG and gastrin To determine the effect of H. pylori infection and atrophic gastritis on serum PG and gastrin levels, the mean serum levels were compared, based on the patients H. pylori infection and atrophic gastritis status (Table 4). Patients were classified into four groups, similar to those used in the ABCD screening method, as follows: (1) no H. pylori infection and no atrophic gastritis (group A), (2) H. pylori infection and no atrophic gastritis (group B), (3) H. pylori infection and atrophic gastritis (group C), (4) no H. pylori infection and atrophic gastritis (group D). Although the serum PG I levels differed among the four groups (P = 0.016), a trend was not identified across the groups (P = 0.238). In patients with serum PG II levels and PG I/II ratios, significant differences were identified among the four groups (PG II, P < 0.001; PG I/II ratio, P < 0.001). Although the PG II levels and PG I/II ratios in group D individuals were out of linearity, trends were identified across the four groups (PG II, P for trend = 0.014; PG I/II ratio, P for trend <0.001). Serum gastrin levels did not differ according to H. pylori infection status or the presence of atrophic gastritis.
666
E.H. Kim et al. / Digestive and Liver Disease 47 (2015) 663–668
Fig. 1. The scatter plots for serum pepsinogen I/II ratios, according to the severity of histologic features, in the antrum (A–D) and the corpus (E–H). Panel A, Lymphoplasmacytic infiltration of the antrum; Panel B, Neutrophilic infiltration of the antrum; Panel C, Atrophic gastritis of the antrum; Panel D, Intestinal metaplasia of the antrum; Panel E, Lymphoplasmacytic infiltration of the corpus; Panel F, Neutrophilic infiltration of the corpus; Panel G, Atrophic gastritis of the corpus; Panel H, Intestinal metaplasia of the corpus. P-values represented the results of the correlation tests between serum pepsinogen I/II ratio and the severity of each histologic feature. PG, pepsinogen. Table 3 Serum levels of pepsinogen and gastrin according to the atrophic gastritis or intestinal metaplasia. Variable
No atrophic gastritis
Antrum PG I, ng/ml PG II, ng/ml PG I/II ratio Gastrin, pg/dl
59.3 12.5 5.2 95.6
± ± ± ±
72.6 8.0 2.1 201.0
Corpus PG I, ng/ml PG II, ng/ml PG I/II ratio Gastrin, pg/dl
64.3 13.2 5.4 65.2
± ± ± ±
38.6 9.1 1.9 92.6
Diffuse PG I, ng/ml PG II, ng/ml PG I/II ratio Gastrin, pg/dl
61.0 13.3 5.2 88.1
± ± ± ±
38.2 9.1 2.1 188.5
Atrophic gastritis
P-value
No intestinal metaplasia
72.6 20.1 4.2 52.7
± ± ± ±
50.5 14.9 1.7 48.9
0.192 0.030 0.040 0.337
64.4 14.3 5.2 92.3
± ± ± ±
38.1 9.9 2.0 204.9
54.6 18.5 3.3 168.4
± ± ± ±
52.9 14.2 1.9 349.0
0.369 0.139 <0.001 0.217
65.1 14.9 5.0 71.4
± ± ± ±
42.3 10.8 2.0 98.2
73.7 22.9 3.4 68.5
± ± ± ±
65.9 16.7 1.6 63.1
0.365 0.105 0.009 0.746
64.4 14.8 5.0 87.8
± ± ± ±
42.3 10.7 2.1 185.8
Intestinal metaplasia
P-value
57.5 14.3 4.5 70.5
± ± ± ±
49.6 11.9 2.2 89.3
35.8 7.9 4.6 224.0
± ± ± ±
22.9 2.6 2.7 498.6
0.044 <0.001 0.593 0.386
37.2 7.5 5.0 64.7
± ± ± ±
21.6 2.8 2.5 49.5
0.097 <0.001 0.950 0.745
0.460 0.983 0.172 0.597
Table 4 Serum levels of pepsinogen and gastrin according to the Helicobacter pylori infection and atrophic gastritis. Variable
No H. pylori infection & No atrophic gastritis (n = 47)
PG I, ng/ml PG II, ng/ml PG I/II ratio Gastrin, pg/dl
66.0 10.5 6.3 66.5
± ± ± ±
43.2 6.9 1.5 103.1
H. pylori infection & No atrophic gastritis (n = 17) 54. 17.5 3.3 79.0
Values were presented as mean with standard deviation. PG, pepsinogen.
± ± ± ±
23.4 8.4 1.3 89.0
H. pylori infection & Atrophic gastritis (n = 12) 89.4 30.4 2.9 86.6
± ± ± ±
57.8 13.6 1.0 65.7
No H. pylori infection & Atrophic gastritis (n = 19) 43.4 10.5 4.5 140.5
± ± ± ±
28.4 6.2 2.2 353.9
P-value
P for trend
0.016 <0.001 <0.001 0.512
0.238 0.014 <0.001 0.203
E.H. Kim et al. / Digestive and Liver Disease 47 (2015) 663–668
667
Fig. 2. Receiver operating characteristics curve for serum pepsinogen and gastrin predicting atrophic gastritis of the antrum (A), corpus (B), and both of them (C). PG, pepsinogen; AUROC, area under the receiver operating characteristics; CI, confidence interval.
4. Discussion The serum levels of PG I, PG II, and gastrin are known to be markers of the gastric mucosa functional status [3,6,18–21]. That is, serum PG I levels and PG I/II ratios decrease in patients with chronic atrophic gastritis, which is believed to be an important risk factor for the development of gastric carcinoma, particularly the intestinal type. In contrast, serum gastrin, an agonist of gastric acid secretion, increases in patients with corpus atrophic gastritis [22]. Serum PG I levels decrease with the progression of corpus gastric atrophy because of the loss of chief cells in the fundic glands [23]. As serum levels of PG II may remain fairly constant, the serum PG I/II ratios decrease. Theoretically, both low serum PG I levels and low serum PG I/II ratios can be utilized to predict the presence of atrophic gastritis. In our study, however, only the serum PG I/II ratio, rather than serum PG I level, was useful for predicting atrophic gastritis in either the antrum or the corpus. Contrary to common belief, many previous studies have also shown the absence of a significant correlation between serum PG I levels and atrophic gastritis [24–27]. For example, Haj-Sheykholeslami et al. reported that Iranian patients with atrophic gastritis had higher serum PG II levels and lower serum PG I/II ratios than patients without atrophic gastritis [26]; neither serum PG I or gastrin levels differed between the two groups. Hosseini et al. also reported an association between atrophic gastritis and serum PG II levels and PG I/II ratios; however, the serum PG I levels were not associated with atrophic gastritis [27]. In a previous study of Korean population, the prevalences of atrophic gastritis in the antrum and corpus were 42.5% and 20.1%, and those of intestinal metaplasia were 28.6% and 21.2%, respectively [28]. Our data also showed similar proportions of the atrophic gastritis and intestinal metaplasia. In addition, we demonstrated
that the presence of atrophic gastritis could be predicted using only the serum PG I/II ratio. The observed serum PG I/II ratio sensitivity and specificity values (approximately 60–80%) were comparable to those associated with Miki et al.’s criteria. In our study population, the PG criteria for atrophic gastritis, proposed by Miki et al., had acceptable specificity (87.7% for the antrum and 92.1% for the corpus); however, the sensitivities were very low (22.7% for the antrum and 42.1% for the corpus). Considering that serum PG tests are usually used for screening for the presence of atrophic gastritis or gastric cancer, the low sensitivity may be an obvious limitation of the Miki et al. criteria. We recommend the use of the serum PG I/II ratio, alone, rather than in combination with serum PG I levels for predicting atrophic gastritis. One of the other valuable findings from our study was the higher serum PG I/II ratio cut-off value for atrophic gastritis of the antrum compared to that of the corpus, or for both the antrum and the corpus. Considering that atrophic gastritis usually develops in the antrum and progresses to the corpus, a relatively high cut-off value in the antrum is reasonable [29]. Atrophic gastritis or gastric cancer screening, therefore, should involve a serum PG I/II cut-off value of 4.9 (the serum PG I/II ratio cut-off value for antrum atrophic gastritis) as the gastric cancer screening criterion based on serum PG I/II ratios. We also investigated the relationships between serum PG and gastrin levels and histopathologic gastric features. Higher grades of chronic inflammation and atrophy were strongly associated with low PG I/II ratios due to a disproportionate increase in serum PG II levels. These findings suggest that gastric mucosal inflammation, induced by H. pylori infection or parietal cell masses, influence PG levels [30]. Kikuchi et al. observed the impact of H. pylori infection and the extent of infection on PG levels in 2584 subjects over 5 years, and found that PG I/II ratios decreased significantly in the H. pylori-positive group [31]. Wu et al. reported lower PG I/II ratios in
668
E.H. Kim et al. / Digestive and Liver Disease 47 (2015) 663–668
H. pylori-positive groups than in negative ones, similar the present study [32]. A polypeptide, secreted by H. pylori early in an infection, has been hypothesized to directly stimulate chief cells and promote the synthesis and secretion of PG, mainly PG II, through intracellular mechanisms involving increased Ca2+ , cAMP and phosphoinositide [33]. In this study, serum PG II levels and PG I/II ratios were significantly different among the patients classified according to their H. pylori infection and atrophic gastritis status. Additionally, serum PG II levels increased and PG I/II ratios decreased from group A to group C patients. These findings, therefore, support the ABCD screening method. However, more investigation into group D patients may be required because specific trends were not evident in this group. Although this was the first study to suggest optimal cut-off values for the PG test for atrophic gastritis in specific anatomic locations, it has several limitations. First, even though this was the prospective study, we could not calculate adequate sample size before initiation of the study because proper preceding research for sample size calculation lacked. Therefore our results may be considered as preliminary data and should be supported by a largenumber prospective study. Second, the numbers of patients with atrophic gastritis or intestinal metaplasia were relatively small. Thus, low statistical power increases the difficulty in proving the usefulness of serum PG I and gastrin levels for predicting the presence of these diseases. Third, the results of the IgG anti-H. pylori antibody tests were unavailable for this study. Therefore, we could not determine the optimal PG I/II ratio cut-off values based on IgG anti-H. pylori antibody status. Because both the PG tests and the IgG anti-H. pylori antibody tests are performed via blood sampling, and this combination of serum tests is used for gastric cancer risk stratification in Japan, evaluation of the optimal PG I/II ratio cut-off values, according to IgG anti-H. pylori antibody status, may have significant implications [14]. Fourth, the generalizability of our results is limited because the study was performed in regions where H. pylori infections are prevalent. In western countries, the optimal serum PG level cut-off value may be different from that in Asian countries. More validation studies are needed to reach a definitive conclusion. Despite these limitations, our data form the basis of a system for determining the optimal PG cut-off values for use in predicting atrophic gastritis, based on the anatomic location of the lesions. The optimal serum PG I/II ratio cut-off values for atrophic gastritis of the antrum and corpus were 4.9 and 3.5, respectively. Serum PG I/II ratios employing these cut-off values would be useful for atrophic gastritis screening. Conflict of interest None declared. References [1] Samloff IM. Cellular localization of group I pepsinogens in human gastric mucosa by immunofluorescence. Gastroenterology 1971;61:185–8. [2] Samloff IM, Liebman WM. Cellular localization of the group II pepsinogens in human stomach and duodenum by immunofluorescence. Gastroenterology 1973;65:36–42. [3] Miki K, Ichinose M, Kawamura N, et al. The significance of low serum pepsinogen levels to detect stomach cancer associated with extensive chronic gastritis in Japanese subjects. Japanese Journal of Cancer Research 1989;80:111–4. [4] Varis K, Sipponen P, Laxen F, et al. Implications of serum pepsinogen I in early endoscopic diagnosis of gastric cancer and dysplasia. Helsinki Gastritis Study Group. Scandinavian Journal of Gastroenterology 2000;35:950–6. [5] Borch K, Axelsson CK, Halgreen H, et al. The ratio of pepsinogen A to pepsinogen C: a sensitive test for atrophic gastritis. Scandinavian Journal of Gastroenterology 1989;24:870–6. [6] Samloff IM, Varis K, Ihamaki T, et al. Relationships among serum pepsinogen I, serum pepsinogen II, and gastric mucosal histology. A study in relatives of patients with pernicious anemia. Gastroenterology 1982;83:204–9.
[7] Nakanome C, Akai H, Goto Y. Serum group I pepsinogen levels in patients with peptic ulcer and normal subjects. Tohoku Journal of Experimental Medicine 1983;139:151–8. [8] Miki K, Ichinose M, Shimizu A, et al. Serum pepsinogens as a screening test of extensive chronic gastritis. Gastroenterologia Japonica 1987;22:133–41. [9] Li P, He C, Sun L, et al. Pepsinogen I and II expressions in situ and their correlations with serum pesignogen levels in gastric cancer and its precancerous disease. BMC Clinical Pathology 2013;13:22. [10] McNicholl AG, Forne M, Barrio J, et al. Accuracy of GastroPanel for the diagnosis of atrophic gastritis. European Journal of Gastroenterology and Hepatology 2014;26:941–8. [11] Nejadi-Kelarijani F, Roshandel G, Semnani S, et al. Diagnostic values of serum levels of pepsinogens and gastrin-17 for screening gastritis and gastric cancer in a high risk area in northern Iran. Asian Pacific Journal of Cancer Prevention 2014;15:7433–6. [12] Zhang XM, Li JX, Zhang GY, et al. The value of serum pepsinogen levels for the diagnosis of gastric diseases in Chinese Han people in midsouth China. BMC Gastroenterology 2014;14:3. [13] Ohata H, Kitauchi S, Yoshimura N, et al. Progression of chronic atrophic gastritis associated with Helicobacter pylori infection increases risk of gastric cancer. International Journal of Cancer 2004;109:138–43. [14] Miki K. Gastric cancer screening by combined assay for serum anti-Helicobacter pylori IgG antibody and serum pepsinogen levels – ABC method. Proceedings of the Japan Academy – Series B: Physical & Biological Sciences 2011;87: 405–14. [15] Nasrollahzadeh D, Aghcheli K, Sotoudeh M, et al. Accuracy and cut-off values of pepsinogens I, II and gastrin 17 for diagnosis of gastric fundic atrophy: influence of gastritis. PLoS ONE 2011;6:e26957. [16] Gomez JM, Wang AY. Gastric intestinal metaplasia and early gastric cancer in the west: a changing paradigm. Gastroenterology & Hepatology 2014;10:369–78. [17] Dixon MF, Genta RM, Yardley JH, et al. Classification and grading of gastritis. The updated Sydney System. International Workshop on the Histopathology of Gastritis, Houston 1994. American Journal of Surgical Pathology 1996;20:1161–81. [18] Vaananen H, Vauhkonen M, Helske T, et al. Non-endoscopic diagnosis of atrophic gastritis with a blood test. Correlation between gastric histology and serum levels of gastrin-17 and pepsinogen I: a multicentre study. European Journal of Gastroenterology and Hepatology 2003;15:885–91. [19] Oksanen A, Sipponen P, Miettinen A, et al. Evaluation of blood tests to predict normal gastric mucosa. Scandinavian Journal of Gastroenterology 2000;35:791–5. [20] Schlemper RJ, van der Werf SD, Biemond I, et al. Seroepidemiology of gastritis in Japanese and Dutch male employees with and without ulcer disease. European Journal of Gastroenterology and Hepatology 1996;8:33–9. [21] Sipponen P, Ranta P, Helske T, et al. Serum levels of amidated gastrin-17 and pepsinogen I in atrophic gastritis: an observational case–control study. Scandinavian Journal of Gastroenterology 2002;37:785–91. [22] Westerveld BD, Pals G, Lamers CB, et al. Clinical significance of pepsinogen A isozymogens, serum pepsinogen A and C levels, and serum gastrin levels. Cancer 1987;59:952–8. [23] Song HJ, Jang SJ, Yun SC, et al. Low levels of pepsinogen I and pepsinogen I/II ratio are valuable serologic markers for predicting extensive gastric corpus atrophy in patients undergoing endoscopic mucosectomy. Gut Liver 2010;4:475–80. [24] Broutet N, Plebani M, Sakarovitch C, et al. Pepsinogen A, pepsinogen C, and gastrin as markers of atrophic chronic gastritis in European dyspeptics. British Journal of Cancer 2003;88:1239–47. [25] Bodger K, Wyatt JI, Heatley RV. Variation in serum pepsinogens with severity and topography of Helicobacter pylori-associated chronic gastritis in dyspeptic patients referred for endoscopy. Helicobacter 2001;6:216–24. [26] Haj-Sheykholeslami A, Rakhshani N, Amirzargar A, et al. Serum pepsinogen I, pepsinogen II, and gastrin 17 in relatives of gastric cancer patients: comparative study with type and severity of gastritis. Clinical Gastroenterology and Hepatology 2008;6:174–9. [27] Hosseini M, Amoueian S, Attaranzadeh A, et al. Serum gastrin 17, pepsinogen I and pepsinogen II in atrophic gastritis patients living in North-East of Iran. Journal of Research in Medical Sciences 2013;18:225–9. [28] Kim N, Park YS, Cho SI, et al. Prevalence and risk factors of atrophic gastritis and intestinal metaplasia in a Korean population without significant gastroduodenal disease. Helicobacter 2008;13:245–55. [29] Quach DT, Le HM, Hiyama T, et al. Relationship between endoscopic and histologic gastric atrophy and intestinal metaplasia. Helicobacter 2013;18:151–7. [30] Wagner S, Haruma K, Gladziwa U, et al. Helicobacter pylori infection and serum pepsinogen A, pepsinogen C, and gastrin in gastritis and peptic ulcer: significance of inflammation and effect of bacterial eradication. American Journal of Gastroenterology 1994;89:1211–8. [31] Kikuchi S, Kurosawa M, Sakiyama T, et al. Long-term effect of Helicobacter pylori infection on serum pepsinogens. Japanese Journal of Cancer Research 2000;91:471–6. [32] Wu KC, Li HT, Qiao TD, et al. Diagnosis of atrophic body gastritis in Chinese patients by measuring serum pepsinogen. Chinese Journal of Digestive Diseases 2004;5:22–7. [33] Lorente S, Doiz O, Trinidad Serrano M, et al. Helicobacter pylori stimulates pepsinogen secretion from isolated human peptic cells. Gut 2002;50:13–8.
Contents lists available at ScienceDirect
Digestive and Liver Disease journal homepage: www.elsevier.com/locate/dld
Alimentary Tract
The optimal serum pepsinogen cut-off value for predicting histologically confirmed atrophic gastritis Eun Hye Kim a , Huapyong Kang a , Chan Hyuk Park b , Hee Seung Choi c , Da Hyun Jung a , Hyunsoo Chung a , Jun Chul Park a , Sung Kwan Shin a , Sang Kil Lee a , Yong Chan Lee a,∗ a Division of Gastroenterology, Department of Internal Medicine, Severance Hospital, Institute of Gastroenterology, Yonsei University College of Medicine, Seoul, Republic of Korea b Department of Internal Medicine, Guri Hospital, Hanyang University College of Medicine, Guri, Republic of Korea c Department of Pathology, Yonsei University College of Medicine, Seoul, Republic of Korea
a r t i c l e
i n f o
Article history: Received 24 January 2015 Accepted 15 May 2015 Available online 23 May 2015 Keywords: Atrophic gastritis Cut-off value Helicobacter pylori Pepsinogen
a b s t r a c t Background: Although serum pepsinogen tests are useful for predicting the presence of atrophic gastritis, the optimal cut-off values have not been fully evaluated. Aim: To determine the optimal serum pepsinogen cut-off value for predicting atrophic gastritis. Methods: Patients scheduled for upper endoscopy at Severance Hospital, Korea, between August 2012 and October 2013, were recruited prospectively. Endoscopic biopsies for atrophic gastritis were obtained and histologically graded, based on the updated Sydney system. Results: Ninety-five patients were enrolled in the study. The mean age was 57.7 ± 12.1 years, and 44.2% of the patients were male. Serum pepsinogen I/II ratios were lower in patients with atrophic gastritis than in those without it (antrum, 4.2 ± 1.7 vs. 5.2 ± 2.1, P = 0.040; corpus, 3.3 ± 1.9 vs. 5.4 ± 1.9, P < 0.001). Serum pepsinogen I/II ratios were significantly correlated with histologic atrophic gastritis (antrum, P = 0.030; corpus, P < 0.001). Using a cut-off value of 4.9, the sensitivity and specificity of the serum pepsinogen I/II ratio for predicting atrophic gastritis in the antrum were 68.2% and 60.3%, respectively. Conclusion: The optimal serum pepsinogen I/II ratio cut-off values for atrophic gastritis of the antrum and for the corpus were 4.9 and 3.5, respectively. Serum pepsinogen I/II ratios, with these cut-off values, are useful for screening patients for the presence of atrophic gastritis. © 2015 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.
1. Introduction Human pepsinogen (PG) I and II are pepsin (an endoproteinase present in the gastric juice. PG I is secreted mainly from the chief cells in fundic mucosa [1]) proenzymes, whereas PG II is secreted from the cardiac, fundic, antral mucosa of the stomach and proximal duodenal mucosa [2,3]. Serum PG I and II levels and the PG I/II ratio may be related to the histologic and functional status of the gastric mucosa [4]. Previous studies have shown that serum PG I levels and PG I/II ratios reflect the morphological and functional status of the gastric mucosa [5–8]. Based on these characteristics, PG tests can be used to predict the presence of atrophic gastritis.
∗ Corresponding author at: Division of Gastroenterology, Department of Internal Medicine, Severance Hospital, Institute of Gastroenterology, Yonsei University College of Medicine, 50-1 Yonsei-ro, Seodaemun-gu, Seoul 120-752, Republic of Korea. Tel.: +82 2 2228 1960; fax: +82 2 393 6884. E-mail address: [email protected] (Y.C. Lee).
Predicting the presence of atrophic gastritis using serum PG levels is important for the following two reasons. First, blood sampling to determine PG levels is much simpler and less invasive than endoscopic biopsy, the gold standard for diagnosing atrophic gastritis. Second, atrophic gastritis is a well-known risk factor for gastric cancer [9–12]. Therefore, PG tests can be used for mass screening of individuals for the presence of heightened gastric cancer risks. In Japan, gastric cancer screening is based on serum immunoglobulin G (IgG) anti-Helicobacter pylori antibody and serum PG levels (the ABCD method), and is used for gastric cancer risk stratification [13,14]. In previous studies, serum PG status was usually defined as “atrophic” when the serum PG I level was ≤70 ng/mL, and the PG I/II ratio was simultaneously ≤3.0, as proposed by Miki et al. [8,14]. However, Miki et al. determined serum PG cut-off values based on endoscopically diagnosed atrophic gastritis rather than histologically confirmed atrophic gastritis [8]. Therefore, the best serum PG cut-off value for predicting histologic atrophic gastritis may differ from the value proposed by Miki et al. A previous study suggested
http://dx.doi.org/10.1016/j.dld.2015.05.014 1590-8658/© 2015 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.
664
E.H. Kim et al. / Digestive and Liver Disease 47 (2015) 663–668
that the best cut-off values for fundic atrophy were serum PG I levels <56 ng/mL and a serum PG I/II ratio <5; those for the antrum and the corpus were not investigated [15]. Considering that both atrophic gastritis and gastric cancer usually develop at the antrum or corpus, identification of the optimal cut-off values for atrophic gastritis in the antrum or corpus may be more important. In addition, further investigation of the relationship between serum PG levels and intestinal metaplasia may be meaningful because intestinal metaplasia is also a well-known risk factor for gastric cancer [16]. Therefore, we evaluated whether serum PG tests can be used to detect atrophic gastritis, based on the anatomic location of the lesions (antrum or corpus). Then, we proposed optimal serum PG cut-off values for predicting atrophic gastritis. We also investigated the usefulness of PG tests for predicting intestinal metaplasia. 2. Materials and methods 2.1. Study design Patients scheduled for upper endoscopy at Severance Hospital, Seoul, Korea, between August 2012 and October 2013, were prospectively recruited for this study. Patients with a history of gastric surgery or prior H. pylori eradication therapy were excluded. Endoscopic biopsies were performed based on the updated Sydney system [17]. The following data were included in this study: age; sex; serum levels of PG I, PG II, and gastrin; and the PG I/II ratio. This study was approved by the Institutional Review Board of Severance Hospital (IRB 4-2013-0722) and written informed consent was obtained from each participant. This study was registered at http://clinicaltrials.gov (registration number NCT02184910). 2.2. Histologic evaluation Biopsy specimens were fixed in formalin, and assessed for the presence of H. pylori (using modified Giemsa staining), and for degree of inflammatory cell infiltration, atrophy, and intestinal metaplasia (all determined following haematoxylin and eosin staining). Additional specimens from the lesser curvature of the antrum were subjected to the rapid urease test (CLO test, Delta West, Bentley, WA, Australia). The histologic features of the gastric mucosa were recorded using the updated Sydney system scores: 0 = none, 1 = mild, 2 = moderate and 3 = severe. All histologic examinations were performed by one experienced pathologist, unaware of patient details. H. pylori infection was assessed using pathology or the rapid urease test. 2.3. Serological tests Fasting serum samples were collected from the participants upon entry into the study. The PG I and PG II levels were measured using latex agglutination turbidimetric immunoassay kits (HiSens Pepsinogen-I LTIA and HiSens Pepsinogen-II LTIA, HBI, Anyang, Korea). Serum gastrin levels were measured using a radioimmunoassay method (Siemens, Los Angeles, CA, USA). 2.4. Statistical analysis Continuous variables are presented as means with standard deviations, and were compared using the t-test. Categorical variables are presented as sample numbers with percentages, and were compared using the 2 or Fisher’s exact tests. Correlations between histologic feature severity and serologic tests were determined using the Spearman rank correlation test. To determine the serum PG cut-off values, receiver operating characteristic (ROC) curves and Youden’s index were used. Finally, analyses of variance were
Table 1 Baseline patient characteristics. Variable
Value
Patient, n Mean age ± SD (years) Male, n (%) Positive of rapid urease test, n (%) PG I, ng/ml, mean ± SD PG II, ng/ml, mean ± SD PG I/PG II, mean ± SD Gastrin, pg/dl, mean ± SD Atrophic gastritis of antrum, n (%) None Mild Moderate Atrophic gastritis of corpus, n (%) None Mild Moderate Intestinal metaplasia of antrum, n (%) None Mild Moderate Severe Intestinal metaplasia of corpus, n (%) None Mild Moderate Severe
95 57.7 ± 12.1 42 (44.2) 23 (24.2) 62.4 ± 41.7 14.3 ± 10.5 5.0 ± 2.1 86.0 ± 179.2 73 (76.8) 18 (18.9) 4 (4.2) 76 (80.0) 16 (16.8) 3 (3.2) 67 (70.5) 20 (21.1) 5 (5.3) 3 (3.2) 86 (90.5) 4 (4.2) 3 (3.2) 2 (2.1)
PG, pepsinogen; SD, standard deviation
used to compare the mean levels of the serologic tests according to the H. pylori infection status and the presence of atrophic gastritis. A P-value <0.05 was considered significant for group comparisons. All statistical analyses were performed using the statistical software programme SPSS for Windows (version 18.0; SPSS, Chicago, IL, USA). 3. Results 3.1. Baseline characteristics The baseline characteristics of the 95 enrolled patients are shown in Table 1. The mean age was 57.7 ± 12.1 years, and 44.2% of the patients were male. The proportion of positive rapid urease tests was 24.2%. The mean serum levels of PG I and PG II, and the PG I/II ratio were 62.4 ng/mL, 14.3 ng/mL and 5.0, respectively. The mean serum gastrin level was 86.0 pg/dL. Atrophic gastritis of the antrum and the corpus was identified in 22 (23.2%) and 19 (20.0%) patients, respectively. In addition, intestinal metaplasia of the antrum and the corpus was shown in 28 (29.5%) and 9 (9.5%) patients, respectively. 3.2. Relationships between serum levels of PG or gastrin and histologic feature severities The correlations between the serum PG or gastrin levels and the histologic feature severity, based on the semi-quantitative scores in the antrum and the corpus, are summarized in Table 2. Although serum PG I levels were correlated with the neutrophilic infiltration severity in the antrum (P = 0.014), such a correlation was not observed in the corpus (P = 0.733). Additionally, a significant correlation was not observed between the serum PG I levels and the severity of lymphoplasmacytic infiltration in either the antrum or the corpus (P = 0.216 and P = 0.574, respectively). In contrast, serum PG II levels showed significant correlation with the severity of both the lymphoplasmacytic infiltration and the neutrophilic infiltration, regardless of location (lymphoplasmacytic infiltration, antrum and corpus, both P < 0.001; neutrophilic infiltration, antrum and
E.H. Kim et al. / Digestive and Liver Disease 47 (2015) 663–668
665
Table 2 Correlation between the serum levels of pepsinogen or gastrin and the severity of histologic features. Variable
PG I
PG II
PG I/II ratio
Gastrin
Correlation efficiency
P-value
Correlation efficiency
P-value
Correlation efficiency
P-value
Correlation efficiency
Antrum Lymphoplasmacytic infiltration Neutrophilic infiltration Atrophic gastritis Intestinal metaplasia
0.128 0.250 0.128 −0.193
0.216 0.014 0.218 0.060
0.445 0.537 0.244 −0.060
<0.001 <0.001 0.017 0.562
−0.539 −0.527 −0.223 −0.168
<0.001 <0.001 0.030 0.104
0.216 0.154 −0.064 −0.020
0.036 0.139 0.541 0.851
Corpus Lymphoplasmacytic infiltration Neutrophilic infiltration Atrophic gastritis Intestinal metaplasia
−0.058 −0.035 −0.180 −0.225
0.574 0.733 0.081 0.028
0.413 0.412 0.158 −0.206
<0.001 <0.001 0.127 0.045
−0.645 −0.620 −0.387 −0.050
<0.001 <0.001 <0.001 0.627
0.418 0.361 0.221 0.079
<0.001 <0.001 0.032 0.452
corpus, both P < 0.001). Additionally, the serum PG I/II ratio showed an inverse correlation with both the severity of lymphoplasmacytic infiltration and the severity of neutrophilic infiltration, regardless of location (lymphoplasmacytic infiltration, antrum and corpus, both P < 0.001; neutrophilic infiltration, antrum and corpus, both P < 0.001). In patients with severe atrophic gastritis, serum PG I levels were not significant factors in either the antrum or the corpus (P = 0.218 and P = 0.081, respectively). Serum PG II levels correlated well with atrophic gastritis severity in the antrum (P = 0.017), but not in the corpus (P = 0.127). However, the serum PG I/II ratio showed a significant, inverse correlation with atrophic gastritis severity in both the antrum and the corpus (P = 0.030 and P < 0.001, respectively). The scatter plots show the distribution of serum PG I/II ratios, according to histologic feature severity (Fig. 1). However, the severity of intestinal metaplasia in the corpus was inversely correlated with both serum PG I and serum PG II levels (P = 0.028 and P = 0.045, respectively), but not with the serum PG I/II ratio (P = 0.627). 3.3. Serum levels of PG and gastrin according to the atrophic gastritis or intestinal metaplasia To evaluate whether the PG or gastrin levels can be used to predict the presence of atrophic gastritis or intestinal metaplasia, the mean serum levels were compared, based on the confirmed presence of atrophic gastritis or intestinal metaplasia (Table 3). Serum PG I levels did not differ between patients with and without atrophic gastritis, regardless of the anatomic location of the lesions. Although PG II serum levels were higher in patients with atrophic gastritis of the antrum than in those without it (20.1 ± 14.9 vs. 12.5 ± 8.0 ng/mL, P = 0.030), there were no differences in serum PG II levels, relative to the presence of atrophic gastritis in the corpus. The serum PG I/II ratio, however, was significantly lower in patients with atrophic gastritis than in those without, regardless of anatomic location (atrophic gastritis vs. no atrophic gastritis; antrum, 4.2 ± 1.7 vs. 5.2 ± 2.1 ng/mL, P = 0.040; corpus, 3.3 ± 1.9 vs. 5.4 ± 1.9 ng/mL, P < 0.001; diffuse, 3.4 ± 1.6 vs. 5.2 ± 2.1 ng/mL, P = 0.009). In patients with intestinal metaplasia, no significant differences in serum PG or gastrin levels were observed, according to the presence of intestinal metaplasia, except among the serum PG II levels in patients with corpus intestinal metaplasia or diffuse intestinal metaplasia (corpus, 7.9 ± 2.6 vs. 14.9 ± 10.8 ng/mL, P < 0.001; diffuse, 7.5 ± 2.8 vs. 14.8 ± 10.7 ng/mL, P < 0.001). 3.4. Receiver operating characteristics curve of PG and gastrin for predicting atrophic gastritis or intestinal metaplasia As shown in Fig. 2, the serum PG I/II ratio was useful for predicting the existence of atrophic gastritis in the antrum,
P-value
corpus, and both of them (area under the receiver operating characteristics [AUROC] (95% confidence interval [CI]): antrum, 0.650 [0.524–0.775]; corpus, 0.781 [0.658–0.903]; diffuse, 0.761 [0.613–0.910]). The optimal serum PG I/II ratio cut-off values, derived from the Youden’s index, were 4.9, 3.5, and 3.5 for atrophic gastritis in the antrum, corpus, and both, respectively. Using a 4.9 cut-off value, the sensitivity and specificity of the serum PG I/II ratio for predicting the presence of atrophic gastritis in the antrum were 68.2% and 60.3%, respectively. With a 3.5 cut-off value, the sensitivity and specificity of serum PG I/II ratio for predicting the presence of atrophic gastritis of the corpus were 68.4% and 81.6%, respectively. Additionally, with a cut-off value of 3.5, the sensitivity and specificity of the serum PG I/II ratio for predicting the presence of diffuse atrophic gastritis were 70.0% and 80.0%, respectively. When Miki et al.’s criteria for atrophic gastritis (PG I, ≤70 ng/mL; and PG I/II ratio, ≤3.0) were applied in this study population, the sensitivity and specificity for predicting the presence of atrophic gastritis in the antrum were 22.7% and 87.7%, respectively; and those for predicting the presence of atrophic gastritis in the corpus were 42.1% and 92.1%, respectively. In contrast to predicting atrophic gastritis, the serum PG I, PG II, and gastrin levels, and the PG I/II ratio did not predict intestinal metaplasia, except that serum PG I levels were predictive of the presence of intestinal metaplasia in the corpus (AUROC [95% CI], 0.721 [0.555–0.887]). With a cut-off value of 44.2 ng/mL, the associated serum PG I level sensitivity and specificity were 77.8% and 65.1%, respectively.
3.5. Effect of H. pylori infection and atrophic gastritis on serum levels of PG and gastrin To determine the effect of H. pylori infection and atrophic gastritis on serum PG and gastrin levels, the mean serum levels were compared, based on the patients H. pylori infection and atrophic gastritis status (Table 4). Patients were classified into four groups, similar to those used in the ABCD screening method, as follows: (1) no H. pylori infection and no atrophic gastritis (group A), (2) H. pylori infection and no atrophic gastritis (group B), (3) H. pylori infection and atrophic gastritis (group C), (4) no H. pylori infection and atrophic gastritis (group D). Although the serum PG I levels differed among the four groups (P = 0.016), a trend was not identified across the groups (P = 0.238). In patients with serum PG II levels and PG I/II ratios, significant differences were identified among the four groups (PG II, P < 0.001; PG I/II ratio, P < 0.001). Although the PG II levels and PG I/II ratios in group D individuals were out of linearity, trends were identified across the four groups (PG II, P for trend = 0.014; PG I/II ratio, P for trend <0.001). Serum gastrin levels did not differ according to H. pylori infection status or the presence of atrophic gastritis.
666
E.H. Kim et al. / Digestive and Liver Disease 47 (2015) 663–668
Fig. 1. The scatter plots for serum pepsinogen I/II ratios, according to the severity of histologic features, in the antrum (A–D) and the corpus (E–H). Panel A, Lymphoplasmacytic infiltration of the antrum; Panel B, Neutrophilic infiltration of the antrum; Panel C, Atrophic gastritis of the antrum; Panel D, Intestinal metaplasia of the antrum; Panel E, Lymphoplasmacytic infiltration of the corpus; Panel F, Neutrophilic infiltration of the corpus; Panel G, Atrophic gastritis of the corpus; Panel H, Intestinal metaplasia of the corpus. P-values represented the results of the correlation tests between serum pepsinogen I/II ratio and the severity of each histologic feature. PG, pepsinogen. Table 3 Serum levels of pepsinogen and gastrin according to the atrophic gastritis or intestinal metaplasia. Variable
No atrophic gastritis
Antrum PG I, ng/ml PG II, ng/ml PG I/II ratio Gastrin, pg/dl
59.3 12.5 5.2 95.6
± ± ± ±
72.6 8.0 2.1 201.0
Corpus PG I, ng/ml PG II, ng/ml PG I/II ratio Gastrin, pg/dl
64.3 13.2 5.4 65.2
± ± ± ±
38.6 9.1 1.9 92.6
Diffuse PG I, ng/ml PG II, ng/ml PG I/II ratio Gastrin, pg/dl
61.0 13.3 5.2 88.1
± ± ± ±
38.2 9.1 2.1 188.5
Atrophic gastritis
P-value
No intestinal metaplasia
72.6 20.1 4.2 52.7
± ± ± ±
50.5 14.9 1.7 48.9
0.192 0.030 0.040 0.337
64.4 14.3 5.2 92.3
± ± ± ±
38.1 9.9 2.0 204.9
54.6 18.5 3.3 168.4
± ± ± ±
52.9 14.2 1.9 349.0
0.369 0.139 <0.001 0.217
65.1 14.9 5.0 71.4
± ± ± ±
42.3 10.8 2.0 98.2
73.7 22.9 3.4 68.5
± ± ± ±
65.9 16.7 1.6 63.1
0.365 0.105 0.009 0.746
64.4 14.8 5.0 87.8
± ± ± ±
42.3 10.7 2.1 185.8
Intestinal metaplasia
P-value
57.5 14.3 4.5 70.5
± ± ± ±
49.6 11.9 2.2 89.3
35.8 7.9 4.6 224.0
± ± ± ±
22.9 2.6 2.7 498.6
0.044 <0.001 0.593 0.386
37.2 7.5 5.0 64.7
± ± ± ±
21.6 2.8 2.5 49.5
0.097 <0.001 0.950 0.745
0.460 0.983 0.172 0.597
Table 4 Serum levels of pepsinogen and gastrin according to the Helicobacter pylori infection and atrophic gastritis. Variable
No H. pylori infection & No atrophic gastritis (n = 47)
PG I, ng/ml PG II, ng/ml PG I/II ratio Gastrin, pg/dl
66.0 10.5 6.3 66.5
± ± ± ±
43.2 6.9 1.5 103.1
H. pylori infection & No atrophic gastritis (n = 17) 54. 17.5 3.3 79.0
Values were presented as mean with standard deviation. PG, pepsinogen.
± ± ± ±
23.4 8.4 1.3 89.0
H. pylori infection & Atrophic gastritis (n = 12) 89.4 30.4 2.9 86.6
± ± ± ±
57.8 13.6 1.0 65.7
No H. pylori infection & Atrophic gastritis (n = 19) 43.4 10.5 4.5 140.5
± ± ± ±
28.4 6.2 2.2 353.9
P-value
P for trend
0.016 <0.001 <0.001 0.512
0.238 0.014 <0.001 0.203
E.H. Kim et al. / Digestive and Liver Disease 47 (2015) 663–668
667
Fig. 2. Receiver operating characteristics curve for serum pepsinogen and gastrin predicting atrophic gastritis of the antrum (A), corpus (B), and both of them (C). PG, pepsinogen; AUROC, area under the receiver operating characteristics; CI, confidence interval.
4. Discussion The serum levels of PG I, PG II, and gastrin are known to be markers of the gastric mucosa functional status [3,6,18–21]. That is, serum PG I levels and PG I/II ratios decrease in patients with chronic atrophic gastritis, which is believed to be an important risk factor for the development of gastric carcinoma, particularly the intestinal type. In contrast, serum gastrin, an agonist of gastric acid secretion, increases in patients with corpus atrophic gastritis [22]. Serum PG I levels decrease with the progression of corpus gastric atrophy because of the loss of chief cells in the fundic glands [23]. As serum levels of PG II may remain fairly constant, the serum PG I/II ratios decrease. Theoretically, both low serum PG I levels and low serum PG I/II ratios can be utilized to predict the presence of atrophic gastritis. In our study, however, only the serum PG I/II ratio, rather than serum PG I level, was useful for predicting atrophic gastritis in either the antrum or the corpus. Contrary to common belief, many previous studies have also shown the absence of a significant correlation between serum PG I levels and atrophic gastritis [24–27]. For example, Haj-Sheykholeslami et al. reported that Iranian patients with atrophic gastritis had higher serum PG II levels and lower serum PG I/II ratios than patients without atrophic gastritis [26]; neither serum PG I or gastrin levels differed between the two groups. Hosseini et al. also reported an association between atrophic gastritis and serum PG II levels and PG I/II ratios; however, the serum PG I levels were not associated with atrophic gastritis [27]. In a previous study of Korean population, the prevalences of atrophic gastritis in the antrum and corpus were 42.5% and 20.1%, and those of intestinal metaplasia were 28.6% and 21.2%, respectively [28]. Our data also showed similar proportions of the atrophic gastritis and intestinal metaplasia. In addition, we demonstrated
that the presence of atrophic gastritis could be predicted using only the serum PG I/II ratio. The observed serum PG I/II ratio sensitivity and specificity values (approximately 60–80%) were comparable to those associated with Miki et al.’s criteria. In our study population, the PG criteria for atrophic gastritis, proposed by Miki et al., had acceptable specificity (87.7% for the antrum and 92.1% for the corpus); however, the sensitivities were very low (22.7% for the antrum and 42.1% for the corpus). Considering that serum PG tests are usually used for screening for the presence of atrophic gastritis or gastric cancer, the low sensitivity may be an obvious limitation of the Miki et al. criteria. We recommend the use of the serum PG I/II ratio, alone, rather than in combination with serum PG I levels for predicting atrophic gastritis. One of the other valuable findings from our study was the higher serum PG I/II ratio cut-off value for atrophic gastritis of the antrum compared to that of the corpus, or for both the antrum and the corpus. Considering that atrophic gastritis usually develops in the antrum and progresses to the corpus, a relatively high cut-off value in the antrum is reasonable [29]. Atrophic gastritis or gastric cancer screening, therefore, should involve a serum PG I/II cut-off value of 4.9 (the serum PG I/II ratio cut-off value for antrum atrophic gastritis) as the gastric cancer screening criterion based on serum PG I/II ratios. We also investigated the relationships between serum PG and gastrin levels and histopathologic gastric features. Higher grades of chronic inflammation and atrophy were strongly associated with low PG I/II ratios due to a disproportionate increase in serum PG II levels. These findings suggest that gastric mucosal inflammation, induced by H. pylori infection or parietal cell masses, influence PG levels [30]. Kikuchi et al. observed the impact of H. pylori infection and the extent of infection on PG levels in 2584 subjects over 5 years, and found that PG I/II ratios decreased significantly in the H. pylori-positive group [31]. Wu et al. reported lower PG I/II ratios in
668
E.H. Kim et al. / Digestive and Liver Disease 47 (2015) 663–668
H. pylori-positive groups than in negative ones, similar the present study [32]. A polypeptide, secreted by H. pylori early in an infection, has been hypothesized to directly stimulate chief cells and promote the synthesis and secretion of PG, mainly PG II, through intracellular mechanisms involving increased Ca2+ , cAMP and phosphoinositide [33]. In this study, serum PG II levels and PG I/II ratios were significantly different among the patients classified according to their H. pylori infection and atrophic gastritis status. Additionally, serum PG II levels increased and PG I/II ratios decreased from group A to group C patients. These findings, therefore, support the ABCD screening method. However, more investigation into group D patients may be required because specific trends were not evident in this group. Although this was the first study to suggest optimal cut-off values for the PG test for atrophic gastritis in specific anatomic locations, it has several limitations. First, even though this was the prospective study, we could not calculate adequate sample size before initiation of the study because proper preceding research for sample size calculation lacked. Therefore our results may be considered as preliminary data and should be supported by a largenumber prospective study. Second, the numbers of patients with atrophic gastritis or intestinal metaplasia were relatively small. Thus, low statistical power increases the difficulty in proving the usefulness of serum PG I and gastrin levels for predicting the presence of these diseases. Third, the results of the IgG anti-H. pylori antibody tests were unavailable for this study. Therefore, we could not determine the optimal PG I/II ratio cut-off values based on IgG anti-H. pylori antibody status. Because both the PG tests and the IgG anti-H. pylori antibody tests are performed via blood sampling, and this combination of serum tests is used for gastric cancer risk stratification in Japan, evaluation of the optimal PG I/II ratio cut-off values, according to IgG anti-H. pylori antibody status, may have significant implications [14]. Fourth, the generalizability of our results is limited because the study was performed in regions where H. pylori infections are prevalent. In western countries, the optimal serum PG level cut-off value may be different from that in Asian countries. More validation studies are needed to reach a definitive conclusion. Despite these limitations, our data form the basis of a system for determining the optimal PG cut-off values for use in predicting atrophic gastritis, based on the anatomic location of the lesions. The optimal serum PG I/II ratio cut-off values for atrophic gastritis of the antrum and corpus were 4.9 and 3.5, respectively. Serum PG I/II ratios employing these cut-off values would be useful for atrophic gastritis screening. Conflict of interest None declared. References [1] Samloff IM. Cellular localization of group I pepsinogens in human gastric mucosa by immunofluorescence. Gastroenterology 1971;61:185–8. [2] Samloff IM, Liebman WM. Cellular localization of the group II pepsinogens in human stomach and duodenum by immunofluorescence. Gastroenterology 1973;65:36–42. [3] Miki K, Ichinose M, Kawamura N, et al. The significance of low serum pepsinogen levels to detect stomach cancer associated with extensive chronic gastritis in Japanese subjects. Japanese Journal of Cancer Research 1989;80:111–4. [4] Varis K, Sipponen P, Laxen F, et al. Implications of serum pepsinogen I in early endoscopic diagnosis of gastric cancer and dysplasia. Helsinki Gastritis Study Group. Scandinavian Journal of Gastroenterology 2000;35:950–6. [5] Borch K, Axelsson CK, Halgreen H, et al. The ratio of pepsinogen A to pepsinogen C: a sensitive test for atrophic gastritis. Scandinavian Journal of Gastroenterology 1989;24:870–6. [6] Samloff IM, Varis K, Ihamaki T, et al. Relationships among serum pepsinogen I, serum pepsinogen II, and gastric mucosal histology. A study in relatives of patients with pernicious anemia. Gastroenterology 1982;83:204–9.
[7] Nakanome C, Akai H, Goto Y. Serum group I pepsinogen levels in patients with peptic ulcer and normal subjects. Tohoku Journal of Experimental Medicine 1983;139:151–8. [8] Miki K, Ichinose M, Shimizu A, et al. Serum pepsinogens as a screening test of extensive chronic gastritis. Gastroenterologia Japonica 1987;22:133–41. [9] Li P, He C, Sun L, et al. Pepsinogen I and II expressions in situ and their correlations with serum pesignogen levels in gastric cancer and its precancerous disease. BMC Clinical Pathology 2013;13:22. [10] McNicholl AG, Forne M, Barrio J, et al. Accuracy of GastroPanel for the diagnosis of atrophic gastritis. European Journal of Gastroenterology and Hepatology 2014;26:941–8. [11] Nejadi-Kelarijani F, Roshandel G, Semnani S, et al. Diagnostic values of serum levels of pepsinogens and gastrin-17 for screening gastritis and gastric cancer in a high risk area in northern Iran. Asian Pacific Journal of Cancer Prevention 2014;15:7433–6. [12] Zhang XM, Li JX, Zhang GY, et al. The value of serum pepsinogen levels for the diagnosis of gastric diseases in Chinese Han people in midsouth China. BMC Gastroenterology 2014;14:3. [13] Ohata H, Kitauchi S, Yoshimura N, et al. Progression of chronic atrophic gastritis associated with Helicobacter pylori infection increases risk of gastric cancer. International Journal of Cancer 2004;109:138–43. [14] Miki K. Gastric cancer screening by combined assay for serum anti-Helicobacter pylori IgG antibody and serum pepsinogen levels – ABC method. Proceedings of the Japan Academy – Series B: Physical & Biological Sciences 2011;87: 405–14. [15] Nasrollahzadeh D, Aghcheli K, Sotoudeh M, et al. Accuracy and cut-off values of pepsinogens I, II and gastrin 17 for diagnosis of gastric fundic atrophy: influence of gastritis. PLoS ONE 2011;6:e26957. [16] Gomez JM, Wang AY. Gastric intestinal metaplasia and early gastric cancer in the west: a changing paradigm. Gastroenterology & Hepatology 2014;10:369–78. [17] Dixon MF, Genta RM, Yardley JH, et al. Classification and grading of gastritis. The updated Sydney System. International Workshop on the Histopathology of Gastritis, Houston 1994. American Journal of Surgical Pathology 1996;20:1161–81. [18] Vaananen H, Vauhkonen M, Helske T, et al. Non-endoscopic diagnosis of atrophic gastritis with a blood test. Correlation between gastric histology and serum levels of gastrin-17 and pepsinogen I: a multicentre study. European Journal of Gastroenterology and Hepatology 2003;15:885–91. [19] Oksanen A, Sipponen P, Miettinen A, et al. Evaluation of blood tests to predict normal gastric mucosa. Scandinavian Journal of Gastroenterology 2000;35:791–5. [20] Schlemper RJ, van der Werf SD, Biemond I, et al. Seroepidemiology of gastritis in Japanese and Dutch male employees with and without ulcer disease. European Journal of Gastroenterology and Hepatology 1996;8:33–9. [21] Sipponen P, Ranta P, Helske T, et al. Serum levels of amidated gastrin-17 and pepsinogen I in atrophic gastritis: an observational case–control study. Scandinavian Journal of Gastroenterology 2002;37:785–91. [22] Westerveld BD, Pals G, Lamers CB, et al. Clinical significance of pepsinogen A isozymogens, serum pepsinogen A and C levels, and serum gastrin levels. Cancer 1987;59:952–8. [23] Song HJ, Jang SJ, Yun SC, et al. Low levels of pepsinogen I and pepsinogen I/II ratio are valuable serologic markers for predicting extensive gastric corpus atrophy in patients undergoing endoscopic mucosectomy. Gut Liver 2010;4:475–80. [24] Broutet N, Plebani M, Sakarovitch C, et al. Pepsinogen A, pepsinogen C, and gastrin as markers of atrophic chronic gastritis in European dyspeptics. British Journal of Cancer 2003;88:1239–47. [25] Bodger K, Wyatt JI, Heatley RV. Variation in serum pepsinogens with severity and topography of Helicobacter pylori-associated chronic gastritis in dyspeptic patients referred for endoscopy. Helicobacter 2001;6:216–24. [26] Haj-Sheykholeslami A, Rakhshani N, Amirzargar A, et al. Serum pepsinogen I, pepsinogen II, and gastrin 17 in relatives of gastric cancer patients: comparative study with type and severity of gastritis. Clinical Gastroenterology and Hepatology 2008;6:174–9. [27] Hosseini M, Amoueian S, Attaranzadeh A, et al. Serum gastrin 17, pepsinogen I and pepsinogen II in atrophic gastritis patients living in North-East of Iran. Journal of Research in Medical Sciences 2013;18:225–9. [28] Kim N, Park YS, Cho SI, et al. Prevalence and risk factors of atrophic gastritis and intestinal metaplasia in a Korean population without significant gastroduodenal disease. Helicobacter 2008;13:245–55. [29] Quach DT, Le HM, Hiyama T, et al. Relationship between endoscopic and histologic gastric atrophy and intestinal metaplasia. Helicobacter 2013;18:151–7. [30] Wagner S, Haruma K, Gladziwa U, et al. Helicobacter pylori infection and serum pepsinogen A, pepsinogen C, and gastrin in gastritis and peptic ulcer: significance of inflammation and effect of bacterial eradication. American Journal of Gastroenterology 1994;89:1211–8. [31] Kikuchi S, Kurosawa M, Sakiyama T, et al. Long-term effect of Helicobacter pylori infection on serum pepsinogens. Japanese Journal of Cancer Research 2000;91:471–6. [32] Wu KC, Li HT, Qiao TD, et al. Diagnosis of atrophic body gastritis in Chinese patients by measuring serum pepsinogen. Chinese Journal of Digestive Diseases 2004;5:22–7. [33] Lorente S, Doiz O, Trinidad Serrano M, et al. Helicobacter pylori stimulates pepsinogen secretion from isolated human peptic cells. Gut 2002;50:13–8.