内镜下食管胃静脉曲张LDRf分型联合临床特征对急性上消化道出血再出血的预测价值OA

Objective:Acute upper gastrointestinal hemorrhage(AUGIB)is a common clinical emergency,and rebleeding is a key factor leading to worsened condition,prolonged hospitalization,and increased mortality.Esophagogastric varices are a major cause of AUGIB.Endoscopic classification based on location,diameter,and risk factors(LDRf)can visually reflect variceal location,size,and bleeding risk factors,but its predictive efficacy for rebleeding alone is limited.This study aims to evaluate the predictive value of endoscopic LDRf classification combined with key clinical indicators for rebleeding within 7 days of hospitalization in AUGIB patients and to construct an efficient combined predictive model to guide early risk stratification,individualized treatment,and monitoring. Methods:A retrospective cohort study was conducted,including 130 AUGIB patients admitted to the Modern Hospital of Sichuan from October 2022 to October 2025.Patients were divided into a rebleeding group(n=34)and a non-rebleeding group(n=96)according to the occurrence of rebleeding within 7 d of hospitalization.Clinical features collected included hemoglobin(Hb),albumin(ALB),fibrinogen(FIB),D-dimer(D-D),and other relevant variables.All patients underwent emergency endoscopy and were classified using LDRf according to bleeding risk:grade 0,low-risk,moderate-risk,and high-risk.Differences between groups were analyzed using t-tests,chi-square test,and rank-sum test.Variables with P<0.05 in univariate analysis were included in multivariate logistic regression to identify independent predictors of rebleeding.A combined predictive model was constructed,and receiver operating characteristic(ROC)curves were plotted to assess the predictive performance of LDRf classification combined with clinical indicators,with calculation of area under the curve(AUC),sensitivity,specificity,and Youden index. Results:Univariate analysis showed that the rebleeding group had significantly higher proportions of cirrhosis,variceal rupture bleeding,bleeding volume,and D-D levels,and significantly lower Hb,ALB,and FIB levels than the non-rebleeding group(all P<0.05).LDRf classification in the rebleeding group was predominantly high-risk,with significantly higher grades than the non-rebleeding group(P<0.05).Multivariate logistic regression indicated that cirrhosis(OR=1.547,95％CI 1.234 to 1.943),variceal rupture bleeding(OR=2.090,95％CI 1.432 to 2.708),bleeding volume(OR=1.850,95％CI 1.145 to 2.134),D-D(OR=1.777,95％CI 1.241 to 2.162),and high-risk LDRf(OR=10.098,95％CI 3.652 to 27.629)were independent risk factors for AUGIB rebleeding,while Hb(OR=0.554,95％CI 0.412 to 0.981)and FIB(OR=0.620,95％CI 0.497 to 0.873)were independent protective factors(all P<0.05).ROC curve analysis showed that LDRf classification alone had an AUC of 0.782,whereas the combined model incorporating LDRf classification,cirrhosis,etiology,bleeding volume,Hb,FIB,and D-D yielded an AUC of 0.912,with a sensitivity of 85.29％and specificity of 83.33％,significantly higher than the predictive performance of a single indicator. Conclusion:High-risk LDRf classification,cirrhosis,variceal rupture bleeding,larger bleeding volume,elevated D-D,and decreased Hb and FIB are closely associated with AUGIB rebleeding.LDRf classification combined with key clinical indicators significantly improves prediction of rebleeding,facilitating rapid identification of high-risk patients,early implementation of intensified hemostasis,close monitoring,and individualized interventions to reduce rebleeding risk and improve outcomes.