| Background and purposeGastric xanthoma(GX)is composed of foam cells rich in lipid particles in the lamina propria of gastric mucosa,which is easy to diagnose under endoscopy and does not require further biopsy in most cases.In recent years,studies have shown that GX is a simple and reliable endoscopic marker for early Gastric cancer(GC),metachronous and synchronous GC,and may even become a new precancerous lesion.However,the etiology and pathogenesis of GX remain unclear.Studies have found that GX is related to lipid metabolism disorders,diabetes,intestinal metaplasia,bile reflux,etc.,but the effects of these factors on GX need further research.At present,most studies at home and abroad focus on the correlation between GX and GC,while there are few and controversial studies on its etiology and pathogenesis.The purpose of this study was to analyze the current incidence of GX,further explore the risk factors in the course of GX,and further analyze the predictive value of related indicators in the diagnosis of GX.MethodsA retrospective study was conducted to collect the data of patients undergoing gastroscopy in our hospital from October 2020 to October 2022(including inpatients in the Department of Gastroenterology and patients in the Physical Examination Center).In strict accordance with the inclusion and exclusion criteria,a total of 130 patients with GX,1683 patients without GX(excluding GX but containing other lesions in the stomach),and 130 patients in the healthy control group(healthy people with no abnormal results after gastroscopy at the same period)were collected.The general and endoscopic data of GX group and non-Gx group were analyzed by univariate and multivariate analysis.In order to avoid the influence of other diseases on serological indexes,serological data of patients in GX group and healthy control group were analyzed by univariate and multivariate analysis,and the risk factors affecting the incidence of GX were obtained.Subsequently,according to the results of multi-factor analysis,the ROC curve model of related serological indicators and GX was constructed and the test statistics were calculated to analyze the predictive value of related serological indicators for the diagnosis of GX.In this study,SPSS26.0 software was used for statistical analysis of the data,and P<0.05 indicated that the difference was statistically significant.Results1.The general data and endoscopic data of GX group and non-GX group were compared.Single factor analysis showed that the two groups of patients in age(X~2=12.005~a,P=0.001),diabetes(X~2=4.510~a,P=0.034),dyslipidemia(X~2=14.489,P<0.001),smoking history(X~2=25.718~a,P<0.001),alcohol(X~2=32.950~a,P<0.001),family history of gastric cancer(P=0.01),shrinkage(X~2=245.758~a,P<0.001),intestinal metaplasia(X~2=169.770~a,P<0.001),hyperplasia(X~2=6.981,P=0.008),gastric hyperplastic polyp(X~2=22.096~a,P<0.001),and bile reflux(X~2=51.444~a,P<0.001),the difference is statistically significant;However,there were no significant differences in gender,obesity,hypertension,coronary heart disease and HP infection between the two groups(P>0.05).Multivariate binary Logistic regression analysis showed that dyslipidemia(OR=9.121,95%CI 3.123-26.641,P<0.001),alcohol consumption(OR=3.547,95%CI1.373-9.159,P=0.009),atrophy(OR=10.841,95%CI 6.442-18.242,P<0.001),intestinal metaplasia(OR=6.597,95%CI 4.036-10.782,P<0.001),gastric hyperplastic polyp(OR=2.438,95%CI 1.550-3.836,P<0.001)and bile reflux(OR=6.159,95%CI 3.470-11.060,P<0.001)were independently associated with GX.2.Serological data of GX group and healthy control group were compared.Univariate analysis showed that there was no statistical significance in age and gender between the two groups(P>0.05),indicating that the two groups were comparable.In both groups,neutrophils(Z=-2.516,P=0.012),NLR(Z=-3.031,P=0.002),CRP(Z=-7.847,P<0.001),TC(Z=-2.185,P=0.029),HDL-C(Z=-2.878,P=0.004),TG(Z=-3.403,P=0.001),FBG(Z=-4.835,P<0.001),Hb A1c(Z=-4.569,P<0.001)and CEA(Z=-4.342,P<0.001)were statistically significant.However,there were no significant differences in the levels of white blood cell count,platelet,hemoglobin,LDL-C,liver function(TP,ALB,ALT,AST,ALP,GGT,TBIL),uric acid,CA199 and CA125 between 2 groups(P>0.05).Multivariate binary Logistic regression analysis showed that NLR(OR=1.900,95%CI 1.102-3.277,P=0.021),CRP(OR=2.612,95%CI 1.800-3.791,P<0.001),TG(OR=3.090,95%CI 1.448-6.593,P=0.004),Hb A1c(OR=2.253,95%CI 1.016-4.997,P=0.046),CEA(OR=1.660,95%CI 1.198-2.299,P=0.002)were independently correlated with GX.3.ROC curves of NLR,CRP,TG,Hb A1c and CEA and GX were plotted and analyzed.AUC of NLR,CRP,TG,Hb A1c and CEA were 0.609,0.781,0.622,0.663 and 0.656,respectively.The Youden index was0.223,0.477,0.269,0.300 and 0.369,respectively.The optimal diagnostic threshold of NLR was 2.5854,the sensitivity was 27.7%and the specificity was 94.6%.The optimal diagnostic threshold of CRP was2.185,the sensitivity was 70.8%,and the specificity was 76.9%.The optimal diagnostic threshold of TG was 0.269,the sensitivity was 30.8%,and the specificity was 96.2%.The optimal diagnostic threshold of Hb A1c was 5.605,the sensitivity was 34.6%,and the specificity was 95.4%.The optimal diagnostic threshold,sensitivity and specificity of CEA were 1.785,57.7%and 79.2%.Conclusions1.History of dyslipidemia,alcohol consumption,atrophy,intestinal metaplasia,gastric hyperplastic polyps,bile reflux and the increase of NLR,CRP,TG,Hb A1c and CEA were independent risk factors for GX.2.The rise of NLR,CRP,TG,Hb A1c and CEA all had a certain predictive value for the diagnosis of GX,while the rise of CRP had a higher predictive value. |