Study on predicting the risk of retinal vein occlusion based on nomogram model and systemic risk factors
10.3760/cma.j.cn511434-20230110-00017
- VernacularTitle:基于列线图模型及系统性发病危险因素预测视网膜静脉阻塞发生风险的研究
- Author:
Meilin SHAO
1
;
Meimei REN
;
Wenyi ZHANG
;
Zhuoyan YANG
;
Yidan WU
;
Jianming WANG
;
Lijun WANG
Author Information
1. 西安交通大学第二附属医院眼科, 西安 710004
- Keywords:
Retinal vein occlusion;
Nomogram;
Risk model
- From:
Chinese Journal of Ocular Fundus Diseases
2023;39(5):381-386
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To establish and preliminarily validate a nomogram model for predicting the risk of retinal vein occlusion (RVO).Methods:A retrospective clinical study. A total of 162 patients with RVO (RVO group) diagnosed by ophthalmology examination in The Second Affiliated Hospital of Xi'an Jiaotong University from January 2017 to April 2022 and 162 patients with age-related cataract (nRVO group) were selected as the modeling set. A total of 45 patients with branch RVO, 45 patients with central RVO and 45 patients with age-related cataract admitted to Xi'an Fourth Hospital from January 2022 to February 2023 were used as the validation set. There was no significant difference in gender composition ratio ( χ2=2.433) and age ( Z=1.006) between RVO group and nRVO group ( P=0.120, 0.320). Age, gender, blood routine (white blood cell count, hemoglobin concentration, platelet count, neutrophil count, monocyte count, lymphocyte count, erythrocyte volume, mean platelet volume, platelet volume distribution width), and four items of thrombin (prothrombin time, activated partial thrombin time, fibrinogen, and thrombin time) were collected in detail ), uric acid, blood lipids (total cholesterol, triglyceride, high-density lipoprotein, low-density lipoprotein, lipoprotein a), hypertension, diabetes mellitus, coronary heart disease, and cerebral infarction. Neutrophil/lymphocyte ratio and platelet/lymphocyte ratio were calculated. The single logistic regression was used to analyze the clinical parameters of the two groups of patients in the modeling set, and the stepwise regression method was used to screen the variables, and the column graph for predicting the risk of RVO was constructed. The Bootstrap method was used to repeated sample 1 000 times for internal and external verification. The H-L goodness-of-fit test and receiver operating characteristic (ROC) curve were used to evaluate the calibration and discrimination of the nomogram model. Results:After univariate logistic regression and stepwise regression analysis, high density lipoprotein, neutrophil count and hypertension were included in the final prediction model to construct the nomogram. The χ2 values of the H-L goodness-of-fit test of the modeling set and the validation set were 0.711 and 4.230, respectively, and the P values were 0.701 and 0.121, respectively, indicating that the nomogram model had good prediction accuracy. The area under the ROC curve of the nomogram model for predicting the occurrence of post-stroke depression in the modeling set and the verification set was 0.741 [95% confidence interval ( CI) 0.688-0.795] and 0.741 (95% CI 0.646-0.836), suggesting that the nomogram model had a good discrimination. Conclusions:Low high density lipoprotein level, high neutrophil count and hypertension are independent risk factors for RVO. The nomogram model established based on the above risk factors can effectively assess and quantify the risk of post-stroke depression in patients with cerebral infarction.
