Study on the application effect of logistic regression prediction model for stress ulcer after cerebral hemorrhage based on serum indicators
10.3969/j.issn.1671-8348.2025.06.006
- VernacularTitle:基于血清指标的脑出血后应激性溃疡logistic回归预测模型的应用效果研究
- Author:
Lixiang WANG
1
;
Yue ZHENG
;
Chanchan FANG
;
Sisi YE
;
Yang ZHANG
;
Yun LI
Author Information
1. 温州医科大学附属第一医院神经内科护理单元,浙江 温州 325000
- Keywords:
cerebral hemorrhage;
stress ulcer;
logistic model;
heat shock protein 70;
heat shock protein 90;
prediction
- From:
Chongqing Medicine
2025;54(6):1312-1318
- CountryChina
- Language:Chinese
-
Abstract:
Objective To construct a logistic regression prediction model for stress ulcer(SU)after cerebral hemorrhage.Methods A total of 230 patients with cerebral hemorrhage admitted to our hospital from January 2020 to January 2023 were prospectively selected as the study subjects.They were randomly di-vided into a training group and a validation group using a random number table method,with 115 patients in each group.The incidence of postoperative SU was statistically compared between the two groups.The least absolute shrinkage and selection operator(Lasso)and logistic regression were used to analyze the influencing factors of SU after cerebral hemorrhage,and a logistic regression prediction model was established and valida-ted.Results The incidence of SU was 19.13%in the training group and 20.00%in the validation group.In-crement of age,blood loss≥30 mL,higher levels of neutrophil-to-lymphocyte ratio(NLR),heat shock protein 70(HSP70)and HSP90 were identified as independent risk factors for SU after cerebral hemorrhage(P<0.05),while lower levels of Glasgow Coma Scale(GCS)score and albumin(Alb)were protective factors(P<0.05).The prediction model was logit(P)=0.409×age+1.288×blood loss-1.335×GCS score-1.126×Alb+0.452×NLR+1.483×HSP70+1.593×HSP90-10.325.The areas under the receiver operat-ing characteristic(ROC)curve(AUC)for the training group and the validation group were 0.845(95%CI:0.765-0.906)and 0.855(95%CI:0.777-0.913),respectively.The sensitivities were 81.82%and 90.91%,and the specificities were 76.34%and 70.97%,respectively.Conclusion A logistic regression prediction model was successfully constructed,which has certain predictive value for SU after cerebral hemorrhage.
