Construction and verification of risk prediction model for the occurrence of calf muscle vein thrombosis in patients with spontaneous basal ganglia hemorrhage after surgery
10.3969/j.issn.1008-9691.2025.03.007
- VernacularTitle:自发性脑基底节出血术后患者小腿肌间静脉血栓发生的风险预测模型构建及验证
- Author:
Zhenhai FEI
1
;
Yong CAI
1
;
Hua GU
1
;
Xingming ZHONG
1
Author Information
1. 湖州市第一人民医院(湖州师范学院附属第一医院)神经外科,浙江 湖州 313000
- Publication Type:Journal Article
- Keywords:
Spontaneous basal ganglia hemorrhage;
Calf muscular vein thrombosis;
Predictive model
- From:
Chinese Journal of Integrated Traditional and Western Medicine in Intensive and Critical Care
2025;32(3):291-297
- CountryChina
- Language:Chinese
-
Abstract:
Objective To construct a risk prediction model for the occurrence of calf muscle vein thrombosis(CMVT)in patients with spontaneous basal ganglia hemorrhage(sBGH)after surgery,and to verify the predictive effect of this model.Methods The surgical patients with sBGH who underwent neurosurgical procedures at the First People's Hospital of Huzhou(First Affiliated Hospital of Huzhou University)from January 2020 to December 2024 were enrolled as study subjects.A total of 345 patients were randomly divided into a modeling group(n=243)and a validation group(n=102).In the modeling group,patients were categorized into CMVT(n=72)and non-CMVT groups(n=171).Collected data included demographics[gender,age,body mass index(BMI)],laboratory indicators[D-dimer,activated partial thromboplastin time(APTT),fibrinogen(Fib),prothrombin time(PT),thrombin time(TT),hypersensitive C-reactive protein(hs-CRP)],and clinical data[hypertension,diabetes,smoking history,preoperative Glasgow coma scale(GCS),preoperative Caprini score,ventricular rupture,hemorrhage volume,surgical approach].Variables with statistical significance in univariate analysis were included in multivariate Logistic regression to identify CMVT risk factors and construct a nomogram model.Model fit was evaluated using the Hosmer-Lemeshow test,and predictive performance was assessed via receiver operator characteristic curve(ROC curve)analysis.Results The analysis of single factors showed that compared with the non-CMVT group,the proportion of male patients in CMVT group was significantly lower[63.9%(46/72)vs.77.8%(133/171)],while age,D-dimer,hs-CRP,Caprini score levels and bleeding volume were significantly higher[age(years):69.00(60.25,75.75)vs.55.00(67.00,74.00),D-dimer(mg/L):1.94(1.02,4.87)vs.0.89(0.48,1.74),hs-CRP(mg/L):21.94(7.50,55.87)vs.10.56(4.62,30.87),Caprini score:5(3,8)vs.3(2,5)].PT and TT were significantly prolonged[PT(s):13.10(12.10,14.13)vs.12.70(12.10,13.50),TT(s):15.25(14.23,16.06)vs.14.70(14.00,15.70)],the proportion of patients with hypertension,diabetes,ventricular rupture,and craniotomy was markedly increased[hypertension:70.8%(51/72)vs.43.9%(75/171),diabetes:25.0%(18/72)vs.7.0%(12/171),ventricular penetration:16.7%(12/72)vs.5.3%(9/171),craniotomy:59.7%(43/72)vs.35.7%(61/171)],GCS score significantly decreased[13(9,14)vs.15(13,15)].All differences were statistically significant(all P<0.05).The multivariate Logistic regression analysis showed that age[odds ratio(OR)=1.053,95%confidence interval(95%CI)was 1.018-1.090],D-dimer(OR=1.194,95%CI was 1.067-1.335),hypertension(OR=2.284,95%CI was 1.085-4.806),diabetes(OR=5.357,95%CI was 1.917-14.973),elevated preoperative Caprini score(OR=1.216,95%CI was 1.066-1.387),and craniotomy(OR=3.359,95%CI was 1.087-10.378)were all independent risk factors affecting the occurrence of CMVT in patients after sBGH(all P<0.05).The linear model was constructed using the aforementioned variables.The Hosmer-Lemeshow test results in the modeling group showedχ2=2.409,P=0.966,indicating good model fit where predicted values showed no significant difference from actual values.ROC curve analysis revealed an area under the curve(AUC)=0.846,95%CI was 0.792-0.899 for the validated model.When the optimal cut-off value was set at 0.538,the model achieved a sensitivity of 70.8%and specificity of 83.0%.In the validation set,the Hosmer-Lemeshow test showedχ2=11.300,P=0.185,indicating good model fit where predicted values showed no significant difference from actual values.ROC curve analysis demonstrated an AUC=0.893,95%CI was 0.832-0.955 for the validated model.At the optimal cut-off value of 0.674,the model achieved a sensitivity of 87.1%,specificity of 80.3%,and an accuracy of 82.4%.Conclusion The risk prediction model has good predictive performance and can provide reference for clinical evaluation of the risk of CMVT in patients with sBGH after surgery.
