Article Text

Download PDFPDF
Serum pepsinogen reference intervals in apparently healthy Chinese population with latex enhanced turbidimetric immunoassay
  1. Meng Huang,
  2. Ai-Guo Tang,
  3. Sa Mu,
  4. Jing-Jing Yang,
  5. Zhong-Yuan Xiang,
  6. Bo Liu,
  7. Jia-Jin Yang
  1. Department of Clinical Laboratory, Second Xiangya Hospital of Central South University, Changsha, Hunan, PR China
  1. Corresponding to Jia-Jin Yang, Department of Clinical Laboratory, Second Xiangya Hospital, Central South University, 139 Renmin Zhong Road, Changsha, Hunan 410011, PR China; Yangjiajin1988{at}


Aim Serum pepsinogen (sPG) has been used to help in diagnosing atrophic corpus gastritis and in screening for gastric cancer non-invasively. There are as yet no reports on sPG reference intervals (RIs) with latex enhanced turbidimetric immunoassay (LIA). In this study, we established the RIs for sPG in a healthy Chinese population using LIA.

Methods Serum PGI and PGII levels in a healthy population (aged 17–80 years) were measured simultaneously using LIA. RIs were determined following Clinical Laboratory and Standards Institute C28-A3 guidelines using a non-parametric method.

Results 95% RIs in men (ng/mL) were: ≤40 years old, 25.53–100.76 for PGI and ≤24.42 for PGII; 41–50 years old, 26.62–124.74 for PGI and ≤26.81 for PGII; and 51–80 years old, 30.40–153.25 for PGI and ≤32.62 for PGII. Corresponding RIs for women (ng/mL) were: ≤40 years old, 21.20–87.44 for PGI and ≤25.53 for PGII; and 41–80 years, 26.40–127.46 for PGI and ≤30.18 for PGII. 95% RI for PGI/PGII in both men and women at any age was ≥2.51.

Conclusions We established the RIs for sPG using LIA in a healthy Chinese population, which can provide a reference for clinical and laboratory studies.

  • diagnostic screening

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Pepsinogen (PG) is an inactive precursor of pepsin in gastric juice which can be classified into two immunochemically distinct groups: PG isozymes PGI and PGII.1 Serum PG (sPG) levels reflect the morphological and functional state of the gastric mucosa.2 They are efficacious biomarkers used to help in diagnosing atrophic corpus gastritis and in screening for gastric cancer non-invasively.3–5 ELISA, radioimmunoassay (RIA) and time resolved fluoroimmunoassay (TRFI) have been developed to measure sPG.2 ,6 ,7 Despite the fact that these assays have high sensitivity and high specificity, they do have some drawbacks, in terms of their dependence on a proficient technician, they are time consuming, unstable and cause radioactive pollution. Recently, the latex enhanced turbidimetric immunoassay (LIA) has become a prevalent method for detecting sPG. LIA is a precise, rapid analysis, is easy to operate and is suitable for mass screening.8–10

The reference intervals (RIs) in a healthy population are of great importance in diagnosis and therapy. Lack of appropriate RIs will affect clinicians when making decisions about applicable medical interventions. The Clinical Laboratory and Standards Institute (CLSI) recommended that reference values have to be re-established or adjusted when the biochemical method improves or changes.11 Although various reference values were proposed using ELISA, RIA and TRFI,10 ,12–16 no available RIs for LIA in the healthy population have yet been reported.

In the present study, we recruited a sufficient number of healthy individuals after exclusion, and established age and gender specific RIs for sPG by LIA in the apparently healthy Chinese population, according to the recommendations of the CLSI C28-A3 document.11

Subjects and methods


The study was approved by the ethics committee of the Second Xiangya Hospital of Central South University. A total of 3622 participants (aged 17–80 years old) were enrolled, covering citizens and villagers. All participants gave signed informed consent and completed a questionnaire. To select apparently healthy individuals, we performed detailed physical examinations and clinical laboratory tests. Exclusion criteria were as follows:16–18

  1. A diagnosis of acute or chronic infections, gastritis, gastric ulcer, or malignancy, vascular, heart, liver, lung, kidney or endocrine disease, diastolic pressure ≥ 90 mm Hg or systolic pressure ≥140 mm Hg.

  2. A history of neurological disease, inherited diseases or suffering from anorexia, dyspepsia or gastrointestinal bleeding.

  3. Pregnant or lactating women, miscarriage in previous 3 months, menstrual disturbance (eg, irregular cycle or heavy flow).

  4. Excessive drinking (more than 30 g per day) and smoking (more than 20 cigarettes per day).

  5. Taken drugs within 4 weeks, blood transfusion or blood donation within 4 months, undergone surgery in the past 6 months and postpartum less than a year.

  6. Heavy exercise and manual work (eg, athletes, construction site workers).

  7. Body mass index (BMI) ≥28 kg/m2 or ≤18.5 kg/m2.

  8. Laboratory tests for exclusion: triglyceride ≥2.26 mmol/L, total cholesterol ≥6.22 mmol/L, fasting blood glucose >7.0 mmol/L, hepatitis B surface antigen, anti-hepatitis C virus or anti-HIV positive, Helicobacter pylori seropositive and abnormal urinalysis.

Based on the above exclusion criteria, 2034 individuals were excluded and 1588 healthy individuals (853 men and 735 women, aged 17–80 years) were included in our study.

Laboratory measurements

Serum samples were obtained after subjects had fasted overnight. Participants did not engage in strenuous exercise for 3 days before sample collection and were encouraged to maintain their normal life habits. Participants sat for 30 min before blood was collected. All samples were collected in 5 mL gel separator tubes (BD Biosciences, New Jersey, USA) by skilled nurses or technicians and allowed to clot at room temperature for 30 min. Samples were then centrifuged for 10 min at 1200×g. All samples were analysed within 2 h.

PGI and PGII levels were measured by LIA on an Architect C8000 automatic biochemical analyser (Abbott Laboratories, Illinois, USA) with a measuring range of 2.5–200 ng/mL for PGI and 2–100 ng/mL for PGII. The reagents for analysis of PGI and PGII, including calibration reagents and controls, were provided by Japan Kanto Chemical Co Inc. To minimise errors, tests were performed according to standard operation procedures recommended in the CLSI documents. Calibrators were used to calibrate the analytical system before measurement of PG. For quality control, two control concentrations of PG were tested. Precision and accuracy were demonstrated according to the document EP 15A recommended by CLSI.19 Precision was expressed as total coefficient of variation (CV), and accuracy was expressed as recovery. The estimated total CV (<10%) and recovery (range 90–105%) met the requirement of the manufacturer.

Statistical analyses

SPSS V.17.0 software (SPSS Inc, Chicago, Illinois, USA) was used for data analysis. The distribution of the data was detected using the Kolmogorov–Smirnov test. Outliers were determined using the Dixon test. After excluding outliers, the medians of variables were compared between men and women using the Mann–Whitney U test. The Kruskal–Wallis H test was applied to compare variables among different age groups. Pearson's correlation analysis was used to test for the correlation between BMI, age, gender and sPG levels. Differences at p<0.05 were considered statistically significant. The 95% RIs were estimated according to the recommendations of the CLSI C28-A3 document.11


The results of the Kolmogorov–Smirnov test showed that levels of PGI, PGII and PGI/PGII in the study showed a non-Gaussian distribution (p<0.05). Therefore, a non-parametric method was used to calculate the 95% RIs. The lower and upper reference limits were defined as the 2.5th and 97.5th percentiles of PGI levels, respectively. For PGII, the upper reference limit is of medical importance and hence it was determined as the 95th percentile. For PGI/PGII, the lower reference limit is of medical importance, and this was determined as the 5th percentile.

Baseline characteristics of the apparently healthy individuals

After screening the original subjects and eliminating any outliers, the current study included a total of 1580 apparently healthy individuals (849 men and 731 women). The baseline characteristics of the individuals are presented in table 1. Both PGI and PGII levels had positive correlations with age (PGI, r=0.231, p<0.001; PGII, r=0.187, p<0.001) and sex (PGI, r=0.158, p<0.001; PGII, r=0.084, p=0.001), but PGI/PGII appeared to have no correlation with age (r=0.004, p=0.873) or sex (r=0.041, p=0.102) (table 2). No correlation was found between sPG levels and BMI (table 2).

Table 1

Baseline characteristics of the apparently healthy individuals

Table 2

Associations between the main variables and serum pepsinogen levels

RIs of sPG in the apparently healthy Chinese population

Median sPG levels in the healthy population were as follows: 46.30 ng/mL for PGI, 11.21 ng/mL for PGII and 4.17 for PGI/PGII. The different percentiles of sPG levels in the healthy Chinese population are shown in table 3.

Table 3

Reference intervals for serum pepsinogen levels (ng/mL) in an apparently healthy Chinese population

RIs of sPG in specific gender and age groups

Selected participants were grouped by gender and age. Participants were divided into five groups (17–30, 31–40, 41–50, 51–60 and 61–80 years old) according to age, for each gender. PGI and PG II levels were significantly higher in men than women (sPGI, 50.31 vs 42.36 ng/mL, p<0.001; sPGII, 12.15 vs 10.08 ng/mL, p<0.001, respectively) while PGI/PGII levels did not show a statistically significant difference (p=0.431) between the sexes (table 4). sPG levels for some age groups in each gender group exhibited some consistency (p>0.05). These age groups were labelled with the same marker (data shown in table 4).

Table 4

Reference intervals for serum pepsinogen (ng/mL) in gender and age specific groups in an apparently healthy Chinese population

RIs of sPG in the apparently healthy Chinese population

PGI and PGII levels did not show statistically significant differences (p>0.05) in the aforementioned groups (data shown in table 4). Therefore, we combined the age groups which had consistent sPG levels into one group (17–40, 41–50, 51–80 for PGI and PGII in men; 17–50, 51–80 for PGI and PGII in women; no partitioning for PGI/PGII). Separate RIs were established for these age groups using a non-parametric approach (table 5). Furthermore, 90% CI for the lower and upper 95% reference limits for PGI were defined (table 6).

Table 5

Gender and age specific reference intervals for serum pepsinogen (ng/mL) in an apparently healthy Chinese population

Table 6

90% CIs of the upper and lower reference limits for pepsinogen I (ng/mL)


sPG is still an effective serological biomarker in defining atrophic gastritis. With the advantages of high sensitivity, non-invasive nature and low cost, it has been applied to screening for gastric cancer worldwide.4 ,20 To ensure an accurate interpretation of the results, appropriate RIs should be given. However, the usefulness of sPG reference values for the identification of diseased gastric mucosa is also affected by variations in methodology, sex, age, race and diet.20–22 There are few reliable reference values for LIA assayed sPG levels categorised within a small age interval in the healthy Chinese population. Hence it is important to establish RIs for sPG by LIA with the necessity of partitioning to understand the extent of improving the health of our population.

Unlike the so-called ‘healthy’ subjects in most previous studies who had normal gastric mucosa, we enrolled 1588 apparently healthy individuals, based on comprehensive exclusion criteria. It enabled us to eliminate the identified effects of smoking, drinking and H pylori infection,16 ,20 ,23 as well as the potential influences of other factors, such as inflammation, infection and other stomach unrelated diseases. In this way, the relationship between sPG and physiological indexes under normal conditions can be truly reflected. BMI was found to be associated with altered levels of some biochemical indicators, such as alanine aminotransferase, gamma glutamyltransferase, uric acid and high density lipoprotein cholesterol,17 but no correlation was observed between sPG and BMI in our study. Furthermore, we found that PGI and PGII levels increased with age and were significantly higher in men than women, which had been proved in previous studies.16 ,24 But PGI/PGII levels did not change with gender or age, which was partially different from these other studies. Participants were not all healthy in these studies, which may explain the discrepancies. We speculate that PGI is always in balance with PGII under normal physiological conditions, and that this balance is not influenced by gender or age.

The morbidity of gastric cancer between men and women, and between ages ≤40 and >40 years is great,20 and hence sex and age should be taken into account when screening for gastric cancer using sPG. Moreover, although decreased PGI/PGII levels have better performance than decreased levels of PGI in indicating atrophic gastric mucosa, PGI levels can increase in non-atrophic gastritis, H pylori infection and peptic ulcer.15 Hence it will be helpful in discriminating these diseases from atrophic gastritis with adequate RIs for PGI. In our study, the relative differences in PGI levels between men and women aged 51–60 years and 61–80 years reached 21.4% and 30.5%, respectively (table 4), and the relative difference in PGI levels between those aged ≤40 years and 51–80 years for men was 24.6% (table 5). We believe our efforts to establish gender and age specific RIs for sPG is worthwhile because the appropriate RIs can improve laboratory service quality and have a conclusive impact on patient management and mass screening.

Huang et al15 reported that the RIs for sPG levels measured by TRFI in a healthy population from Southern China were 58.2≤PGI≤266.6 ng/mL, PGII≤25.3 ng/mL and PGI/PGII≥6.0. RIs for sPG levels in the aforementioned study exhibited some discrepancies. Considering the diversity between the geographical environment and lifestyle of participants, the methodological variation may account for the differences.

The RIs for sPG levels reported from other Asian nations also showed some discrepancies. In Japan, the common reference values used to indicate atrophic gastritis were obtained by RIA, which refer to PGI≤70 ng/mL and PGI/PGII≤3.0.7 ,12 In Korea, Kim’s study with the LIA included 521 healthy H pylori negative participants; they showed lower PGI and PGII levels than ours.16 There are several possible explanations for these regional differences, including the definition of a normal healthy population as well as variations in methodology, race and dietary habits. Nevertheless, a definitive explanation for these changes still requires further validation. Although the applicability of sPG when measured by LIA has been verified, there are no credible RIs with LIA worldwide due to the shortage of systematic studies in an apparently healthy population.

In conclusion, this is the first systematic study with LIA to establish RIs for sPG in an apparently healthy Chinese population, according to the recommendation of the CLSI C28-A3 document. Our results can provide a useful and reliable reference for clinical and laboratory studies.

Take home messages

  • It is essential that the appropriate serum pepsinogen (sPG) reference intervals (RIs) are used for accurate performance of gastric cancer screening. The usefulness of RIs are affected by many factors, such as variations in methodology, sex, age, race and diet.

  • Although various reference values have been proposed for ELISA, radioimmunoassay and time resolved fluoroimmunoassay, no credible RIs for latex enhanced turbidimetric immunoassay (LIA) have yet been reported.

  • Establishing RIs by LIA with a systematic healthy population study and specific partitioning can improve patient management and mass screening.


The authors wish to thank their colleagues at the Department of Clinical Laboratory and Health Centre at the Second Xiangya Hospital, Central South University, for their assistance.



  • Contributors Study design: MH and Jin-JY. Participants selection: SM, MH and Jin-JY. Specimen analysis: Z-YX and BL. Data analysis: A-GT, Jia-JY and MH. Manuscript editing and preparation: Jin-JY, Jia-JY, Z-YX and BL. Manuscript review: A-GT and MH.

  • Funding The work described in this paper was supported by grants from the Natural Science Foundation of the China Hunan Provincial of Science and Technology Department (grant No 2013FJ4087).

  • Competing interests None.

  • Ethics approval The ethics committee of the Second Xiangya Hospital of Central South University approved the study.

  • Provenance and peer review Not commissioned; externally peer reviewed.