Study On The Relationship Between PGⅠ、PGⅡ And Gastric Precancerous Lesions

Background serum-pepsinogen testing as an initial screening method for atrophicgastritis and early gastric cancer has been implemented in countries such as Japan,Finland. In China, PG research is in its early period, especially about precancerouslesions research including atrophy, intestinal metaplasia and dysplasia.Objective1. To compare PG measurements of314serum samples between chemiluminescentmicroparticle immunoassay and enzyme-linked immunosorbent assay.2. To explore influential factors related with pepsinogen in Liyang City, Jiangsuprovince.To study the distribution of pepsinogen in both case and control, Hp+/-, anddifferent pathological diagnosis population, and to further estimate relationshipbetween pepsinogen concentration and gastric precancerous lesions risk.Method1. Case-control study was carried out in Liyang City between March1,2008andFebruary28,2010. Cases were diagnosed by endoscopy and pathology methods andregistered from two resources: the first part was the newly diagnosed patients in thegastroenterology outpatient of Hospital of Chinese Medicine of Liyang City because ofthe illness of stomach; the second was the patients who had been diagnosed with gastricprecancerous lesions in the census on upper digestive tract disease in2004in Liyangcity and re-diagnosed as the disease during our study period. Finally there were317eligible in all subjects. Using matching frequency method,398non-precancerouslesions controls were included by same district, similar occupation, and similarage(±3years old). All cases and controls were interviewed by standard structuredquestionnaire to acquire the information on demographic characteristics, smoking,alcohol drinking, tea drinking, diet history and family history of upper digestive tract,etc. 2. Four pieces of gastric mucosa were collected from the diseased region ofsubjects, two of these were used to carry out the pathological diagnosis and Hp infectiontest, respectively, and the other two were used to identify the methylation status of p16,E-Cadherin, MGMT genes in both case and control population. The serum PG levelswere measured both using the ARCHITECT PG Ⅰ&PG Ⅱassay, a chemiluminescentmicroparticle immunoassay(Abbott Laboratories, USA) in211cases and133controlsand in205cases and125controls by enzyme-linked immunosorbent assay.3. Spearman rank correlation, linear regression and logarithmic curve fitting wereused to analyze the consistency of the two detection methods. Wilcoxon rank sum testwere to contrast the distribution of sex, age, smoking, alcohol, Hp infection anddifferent pathological diagnosis population. Unconditional logistic regression was usedto estimate the association strength of the relationship between pepsinogenconcentration and gastric precancerous lesions by using OR and95%CI. SAS9.2software were conducted in all the analysis. A two-sided test P-value<0.05wasconsidered to be statistically significant.Results1. The results of spearman rank correlation analysis between CMIA and ELISAshowed a significant correlation (PG Ⅰ: rs=0.3583, P<0.0001; PG ⅠI: rs=0.3521,P<0.0001). The linear regression equations and logarithmic curve equations of PG Ⅰ(thevalues by ELISA as the dependent variable y and CMIA as the independent variable x)were=93.48+0.30x and=-34.20+85.50lgx; The linear regression equations andlogarithmic curve equations of PG Ⅱwere=10.53+0.90x and=-4.16+25.46lgx,respectively.2. It was found that serum PG Ⅰand PG Ⅱvalues were higher in case than incontrol(P<0.05), but PGR levels between them were not statistically significant. PG Ⅰvalues were not significantly associated with sex, age, smoking and alcoholconsumption in the case group; but in the control group it was higher in smokers than innon-smokers(P=0.0389). PG Ⅱvalues were not significantly associated with sex, age,smoking and alcohol consumption in both groups. All subjects in cases were dividedinto four groups by age(≤49,~56,~60,>60), It was found that significant differenceswere shown between the four groups (P=0.0057), and PGR levels decreased withincreasing age. The median PGR values of≤48,~56,~61,>61were5.38,4.24, 4.72,3.94, respectively.3. The subjects were devided into normal, superficial gastritis, chronic atrophicgastritis (CAG), intestinal metaplasia(IM) and dysplasia(DYS) according to degree ofseverity of pathological diagnosis. It was found that significant differences were shownbetween the PG Ⅰlevels of five groups(χ2=12.45，P=0.0143), their median PG Ⅰvalueswere38.35ng/ml,50.50ng/ml,59.70ng/ml,55.70ng/ml and52.30ng/ml; PG Ⅱvalueswere significantly associated with severity of pathological diagnosis(χ2=15.55，P=0.0037), the median PG Ⅱvalues were6.90ng/ml,10.00ng/ml,11.20ng/ml,13.20ng/ml and12.10ng/ml; also there was statistically significant in five PGRlevels(χ2=13.34, P=0.0097), the median PGR values were5.56,5.18,5.21,4.17and4.49, respectively.4. After adjustment for age, sex, occupation, education, income, smoking andalcohol consumption and taking the PG Ⅰconcentrations≤37.45ng/ml as a reference, theORs and95%CIs of gastric precancerous lesions for PG Ⅰconcentrations in37.45ng/ml~51.10ng/ml,51.10ng/ml~66.15ng/ml,>66.15ng/ml were1.36(0.27-2.55),1.55(0.83-2.90),1.86(0.97-3.57), respectively, the risk of gastric precancerous lesionswas linked with the increased PG Ⅰlevels, but it did not reach statistical level (P=0.2833for trend test). The risk of gastric precancerous lesions was associated with theincreased PG Ⅱlevels, trend test was close statistical significance(P=0.0575). With thePG Ⅱconcentrations≤7.35ng/ml as a reference, the ORs and95%CIs of gastricprecancerous lesions for PG Ⅱconcentrations in7.35ng/ml~11.20ng/ml,11.20ng/ml~15.40ng/ml,>15.40ng/ml were1.47(0.78-2.78),1.62(0.86-3.06),2.48(1.29-4.76), respectively. With the increase of PGR levels, the risk of gastricprecancerous lesions was decreased(P=0.0069for trend test). Taking the PGR≤3.54asa reference, the ORs and95%CIs of gastric precancerous lesions for PGR in3.54~4.77,4.77~6.39,>6.39were0.80(0.41-1.54),0.34(0.17-0.65),0.71(0.36-1.38), respectively.Conclusion1. Our study showed that serum PG Ⅰ, PG Ⅱlevels were significantly higher inpeople of gastric precancerous lesions than that of normal or superficial gastritis inLiyang City, Jiangsu Province. But there was no difference in PGR between the twogroups. The risk of gastric precancerous lesions was linked with the increased PG Ⅰ, PG Ⅱlevels.2. PG Ⅰ, PG Ⅱand PGR were associated with types of pathological and endoscopicdiagnosis. Gender, age, occupation, marital status and smoking may be the influencingfactors.3. It had some relevance between CMIA and ELISA. Because of easy operationand application to batch specimens, it will bring new opportunities for gastric cancerscreening.4. The results inconsistent with the previous study suggested that medicationhistory could affect the PG levels. Therefore, PG test used for screening gastricprecancerous lesions was feasible, but also with limitations. In the future, more careful,thorough studies are needed to support our point.