Reference Ranges and Comparison of Pepsinogen by Chemiluminescence Immunoassay and Enzyme-Linked Immunosorbent Assay in Chinese Population.

Abstract

Serum pepsinogen (PG) is a good indicator of atrophic changes in the gastric mucosa. Gastric mucosal atrophy is a high-risk factor for gastric cancer. Serological testing for PG combined with endoscopy can help to improve gastric cancer screening. In this study, we established the reference ranges of serum PG-I, PG-II, and the PG-I/II ratio (PGR) in the Chinese population by chemiluminescence immunoassay (CLIA) and enzyme-linked immunosorbent assay (ELISA). Besides, in the real world, doctors are often confused by the results of different testing platforms. Thus, a comparison of methods CLIA and ELISA was performed. 2904 individuals were enrolled from six regions in China as part of the Chinese Adult Digestive Diseases Surveillance (2016) program. The individuals completed questionnaires and volunteered to undergo examinations, including gastroscopy, urea breath test, abdominal ultrasound examination and routine serologic tests. Serum was collected to measure PGs (including PG-I, PG-II and PGR) by CLIA and ELISA. Participants who were found obvious abnormalities or absent from the examinations were excluded. Ultimately, 747 healthy individuals were enrolled in this study. The Kolmogorov-Smirnov test was used to assess the distribution of variables. The Kruskal-Wallis H or Mann-Whitney U-tests were used to compare different sex, age, and geographical groups. The 95% reference ranges of PGs obtained by the two methods were established according to document CLSI-EP28-A3, with covariates of sex, age, and region. Spearman correlation analysis, linear regression analysis and allowable total error (ATE) zone analysis were utilized for comparing the two methods. On overall, the 95% reference ranges of PG-I, PG-II, and PGR measured by CLIA were 23.00-110.64 ng/mL, 2.50-19.13 ng/mL, and 3.87-13.30, respectively. Meanwhile, the reference ranges of PG-I, PG-II, and PGR measured by ELISA were 36.93-205.06 ng/mL, 1.65-17.96 ng/mL, and 7.50-33.60, respectively. Both PG-I and PG-II levels measured by the two platforms were found to be influenced by sex and age. PGR measured by CLIA was influenced by age but not by sex, while PGR measured by ELISA was not affected by either age or sex. Regional factors did not significantly impact the PG results, except for PG-I detected by ELISA. Ultimately, reference ranges for PGs were established based on age and sex stratification. Additionally, the Spearman correlation analysis revealed that the correlation coefficients for PG-I, PG-II, and PGR detected by the two methods were 0.899, 0.887, and 0.777, respectively, indicating a strong correlation between the two methods. The regression equation for the PG levels detected by two methods was obtained through linear regression analysis. The ATE analysis provided a visual depiction of the consistency between the two methods, clearly indicating the poor agreement between them. This study established the reference ranges of PGs by strict and intact enrollment standard. In addition, the results indicated a strong linear relationship between the two methods, yet with a clear bias, which was valuable for laboratory interpretation.