Abstract: Objective To explore the risk prediction model of diabetic retinopathy (DR) in female patients with type 2 diabetes mellitus (T2DM). Methods A total of 253 female patients with confirmed T2DM were included in the study. The patients were divided into DR group and non-DR group according to whether DR occurred or not according to the results of the fundus examination. The baseline data of the two groups were compared, and multivariate logistic regression was used to analyze the case data of the two groups. ROC curves were drawn using graphpad prism 8 to analyze the predictive performance of the model. Results Multivariate logistic regression analysis showed that the duration of diabetes mellitus (DM), glycosylated hemoglobin (HbA1c), neutrophil-to-lymphocyte ratio (NLR), and pulse pressure were the main risk factors for the occurrence of DR in women. Subcutaneous fat area (SFA) was the protective factor in the development of DR for females. Construct a female DR risk prediction model Y=1/[1+e^{-(0.138X1+0.265X2+0.477X3+0.023X4-0.007X5-5.304)}], X1 refers to the course of DM, X2 refers to HbA1c, X3 refers to NLR, X4 refers to pulse pressure, X5 refers to SFA. The area under the ROC curve (AUC) of the risk prediction model was 0.806, the 95% CI was 0.752~0.860 (P<0.001), the maximum Youden index was 0.480, the best critical value was 0.335, the corresponding sensitivity was 0.798, and the specificity was 0.682. Conclusion The occurrence of DR in female T2DM patients is related to the course of DM, HbA1c, NLR, SFA, and pulse pressure. The risk prediction model of DR in female T2DM patients is feasible.

Key words: type 2 diabetes, female, diabetic retinopathy, risk factors, prediction model

摘要： 目的 探究2型糖尿病（T2DM）女性患者并发糖尿病视网膜病变（DR）的风险预测模型。方法 将确诊T2DM的253例女性患者纳入研究,根据眼底检查结果是否发生DR将患者分为DR组和非DR组。比较2组患者基线资料,应用多因素Logistic回归分析2组病例资料。利用Graphpad Prism 8绘制ROC曲线,分析该模型的预测效能。结果 多因素Logistic回归分析的结果显示,糖尿病（DM）病程、糖化血红蛋白（HbA1c）、中性粒细胞淋巴细胞比值（NLR）、脉压是女性DR 发生的主要危险因素,皮下脂肪面积（SFA）是DR发生的保护性因素。构建女性DR风险预测模型Y=1/[1+e^{-(0.138X1+0.265X2+0.477X3+0.023X4-0.007X5-5.304)}],X1指DM病程,X2指HbA1c,X3指NLR,X4指脉压,X5指SFA。风险预测模型的ROC曲线下面积（AUC）为0.806,95%CI为0.752~0.860（P<0.001）,约登指数最大值为0.480,最佳临界值为0.335,对应灵敏度为0.798、特异度为0.682。结论 女性T2DM患者DR的发生与DM病程、HbA1c、NLR、SFA、脉压相关;模型预测女性T2DM患者DR患病风险具有可行性。

关键词: 2型糖尿病, 女性, 糖尿病性视网膜病变, 危险因素, 预测模型