Risk Factor Analysis And Establishment Of Nomogram For Patients With Chronic Atrophic Gastritis

ObjectiveThe aim of the research was to identify the risk factor and to construct the predictive model for patients with chronic atrophic gastritis(CAG)based on their basic characteristics,living habits,clinical symptoms and traditional Chinese medicine(TCM)syndrome differentiation.MethodsFrom August 15,2020 to February 09,2021,patients were selected who were diagnosed as CAG in Guangdong Provincial Hospital of TCM.Clinical information,including basic characteristics,living habits,clinical symptoms and traditional Chinese medicine(TCM)syndrome differentiation was collected through structured questionnaire which was composed of twenty-four relative factors.The independent risk factors of CAG were screened by univariate and multivariate logistic regression analysis.Based on logistic regression coefficient,nomogram was constructed.The predictive accuracy of the model was evaluated by C-index,the area under the ROC curve(AUC)and the calibration curve.P value less than 0.05 was considered as statistically significant in this paper.ResultsA total of 244 eligible patients were included in this study,including 132 CAG patients and 112 in control group.1.Baseline comparison of the two groups(1)Basic characteristics:there were significant differences in age,helicobacter pylori(Hp)infection between the two groups(P<0.05),no significant difference was found in the aspect of gender,body mass index(BMI),education,marital status,family history of gastric cancer,history of taking non steroidal anti-inflammatory drugs(NSAIDs)and combined diseases(diabetes and hypertension)(P>0.05).(2)Living habits:there was significant difference in drinking between the two groups(P<0.05),no significant difference was found in the aspect of smoking,tea drinking and diet(P>0.05).(3)Clinical symptoms:there were significant differences in appetite and quality of sleep between the two groups(P<0.05).There was no significant difference in terms of other symptoms,including distension,stomachache,belching,acid reflux,nausea,dry and bitter in mouth,and stool(P>0.05).(4)TCM syndrome differentiation:A total of six syndromes were included in this research,and among them,77 cases(31.6%)were classified as Qizhi,followed by Shire with 52 cases(21.3%).Other syndromes were Yinxu(16.8%),Qixu(10.7%),Xueyu(10.2%)and Shizu(9.4%)in the sequence of frequency.There was significant difference in syndrome differentiation between the two groups(f<0.05).2.Screening of the independent risk factorsThrough the logistic regression analysis,five predictive factors were finally screened,including age(OR 1.048,95%CI 1.016-1.083)、Hp infection(OR 2.759,95%CI 1.469-5.292)、long-term drinking(OR 2.629,95%CI 1.320-5.433)、sleep with poor quality(OR 2.052,95%CI 1.103-3.880)and syndrome of Xueyu(OR 4.046,95%CI 1.361-13.982).3.Construction of the regression equation and prediction modelLogit(P)=-3.422+0.047X1+1.015X2+0.967X3+0.719X4+1.398X5(X1=age(Year),X2=Hp infection,X3=long-term drink,X4=poor sleep,X5=Xueyu syndrome)4.Evaluation of the nomogram for CAGThe C-index and AUC value are both 0.761(95%CI 0.701,0.822),which indicated a good discrimination.The calibration curve had a high coincidence degree with the ideal standard curve,which also indicated a good consistency in predicting CAG.Conclusion1.Age,HP infection,long-term drinking,sleep with poor quality and syndrome of Xueyu are independent risk factors for CAG.2.The prediction model is as follows:Logit(P)=-3.422+ 0.047X1+1.015X2+0.967X3+0.719X4+1.398X5(X1=age(Year),X2=Hp infection,X3=long-term drink,X4=poor sleep,X5=Xueyu syndrome)3.The CAG prediction model has good prediction accuracy,which still needs to be further verified by studies with large sample size.