[关键词]
[摘要]
目的:探讨视网膜中央静脉阻塞(CRVO)后发生新生血管性青光眼(NVG)的危险因素,在基础上构建预测模型。
方法:选择2016-02/2020-03我院眼科收治的483例527眼CRVO患者临床资料并随访至2021-06,统计CRVO合并NVG情况。采用多因素Logistic回归分析CRVO后发生NVG的危险因素。根据危险因素回归系数构建CRVO后发生NVG的风险预测模型,Hosmer-Lemeshow(H-L)检验、受试者工作特征(ROC)曲线法评价预测模型。
结果:患者失访15例23眼,随访15~64(中位数35)mo,其中70例86眼发生NVG(NVG组),398例418眼未发生NVG(非NVG组)。回归分析结果显示缺血型CRVO、首次就诊时IOP≥18mmHg、高血压、首次就诊时相对性瞳孔传入缺陷(RAPD)≥0.75logU、首次就诊时裸眼视力(UCVA)>0.30(LogMAR)是CRVO后发生NVG的危险因素(P<0.01),抗VEGF治疗是CRVO后发生NVG的保护因素(P<0.01)。CRVO后发生NVG的风险预测模型具有较好的符合度(H-L检验P>0.05)以及判别度(曲线下面积为0.877,95%CI:0.830~0.924,灵敏度为84.3%,特异度为88.9%)。进一步构建预测评分模型,其CRVO后发生NVG的临界值为5分,曲线下面积为0.844(95%CI:0.790~0.898),灵敏度为78.6%,特异度为87.4%。
结论:缺血型CRVO、首次就诊时IOP≥18mmHg、高血压、首次就诊时RAPD≥0.75logU、首次就诊时UCVA>0.30(LogMAR)、抗VEGF治疗与CRVO后发生NVG密切相关,据此建立预测模型具有较好的预测效能。
[Key word]
[Abstract]
AIM: To investigate the risk factors of neovascular glaucoma(NVG)after central retinal vein occlusion(CRVO), and to construct a predictive model.
METHODS: The clinical data of 483 patients(527 eyes)with CRVO admitted to the Department of Ophthalmology of our hospital from February 2016 to March 2020 were retrospectively selected and followed up until June 2021. CRVO combined with NVG were counted.The risk prediction model of NVG after CRVO was constructed according to the regression coefficient of risk factors, and the Hosmer-Lemeshow(H-L)test and receiver operating characteristic(ROC)curve method were used to evaluate the prediction model.
RESULTS: Fifteen patients(23 eyes)were followed up with 35(15-64)mo. NVG was happened in 70 patients(86 eyes)(NVG group)and no NVG was happened in 398 patients(418 eyes)(non-NVG group). Regression analysis showed that ischemic CRVO, IOP≥18mmHg at the first visit, hypertension,relative afferent pupillary defect(RAPD)≥0.75logU at the first visit, uncorrected visual acuity(UCVA)>0.30(LogMAR)at the first visit were risk factors for NVG after CRVO(P<0.01), and anti-vascular endothelial growth factor(VEGF)therapy were protective factors for NVG after CRVO(P<0.01). The risk prediction model for NVG after CRVO had good consistency(H-L test P>0.05)and discrimination(area under the curve was 0.877, 95%CI:0.830-0.924, sensitivity was 84.3%, specificity was 88.9%). The predictive scoring model was further constructed, and the critical value of NVG after CRVO was 5 points,the area under the curve was 0.844(95%CI:0.790-0.898), sensitivity was 78.6%, and specificity was 87.4%.
CONCLUSION: Ischemic CRVO, IOP≥18mmHg at the first visit,hypertension,RAPD≥0.75logU at the first visit, UCVA>0.30(LogMAR), and anti-VEGF therapy are closely related to NVG after CRVO, and the predictive model established on this basis has good predictive efficacy.
[中图分类号]
[基金项目]