基于血清补体因子及血脂水平评估湿性年龄相关性黄斑变性

An Evaluation of Wet Age-related Macular Degeneration Based on Serum Complement Factors and Blood Lipid Level

目的探讨湿性年龄相关性黄斑变性(W-AMD)与血清补体因子及血脂水平的相关性;构建基于补体因子及血脂水平的W-AMD预测模型。方法收集2021年1月至2022年1月首都医科大学附属北京同仁医院经临床确诊的146例W-AMD患者作为研究组;选取同时期89例健康体检者作为对照组。分别测定各组样本血清补体因子,包括C1q、C3、CFB及CFH;血脂成分,包括TC、TG、HDL-C及LDL-C。应用SPSS 25.0统计软件对W-AMD组与对照组的各项研究指标进行差异性比较;进一步对W-AMD组中存在统计学差异的各项研究指标之间进行相关性分析。使用R Studio软件(R 3.6.0),构建基于补体因子及血脂水平的W-AMD预测模型,并验证模型预测效能。结果W-AMD组与对照组比较,C1q、C3、CFB以及TC、TG、LDL-C差异存在统计学意义,P值<0.05;CFH和HDL-C差异无统计学意义,P>0.05。进一步分析W-AMD组中C3与CFB、TC、TG、LDL-C之间,CFB与TC、TG之间,TC与TG、LDL-C之间,以及TG与LDL-C之间存在关联性,P<0.01。应用R Studio软件(R 3.6.0)综合C1q、C3、CFB及TC、TG、LDL-C构建Logistic回归模型,W-AMD=-12.835+8.753×CFB+6.098×TG。模型训练集特异性0.758、灵敏度0.843、准确度0.811、阳性预测值0.852、阴性预测值0.746、AUC值0.901;验证集特异性0.630、灵敏度0.909、准确度0.803、阳性预测值0.800、阴性预测值0.810、AUC值0.943。结论W-AMD患者血清补体因子C1q、C3、CFB及外周血脂水平TC、TG、LDL-C均高于健康人群水平,推测与其发病机制相关,有助于寻找治疗靶点;基于补体因子及血脂水平建立的W-AMD预测模型能够准确识别W-AMD,为临床评估提供了一种客观的实验室参考指标。

Objective To investigate the association between wet age-related macular degeneration(W-AMD)and serum complement components and blood lipid levels,and to further establish the prediction model of W-AMD based on complement factors and blood lipid levels.Methods A total of 146 clinically diagnosed W-AMD patients from January,2021 to January,2022 were retrospectively collected as the study group,while 89 healthy subjects were selected as control group.Serum complement components,including C1q,C3,CFB,and CFH,as well as lipid components,including TC,TG,HDL-C,and LDL-C were collected in each group.SPSS 25.0 statistical software was used to investigate and compare multiple research indicators between the study group and the control group.Further,the correlation analysis was carried out among identified research indicators with statistical differences in the study group.R Studio software(version 3.6.0)was used to construct a prediction model of W-AMD based on complement components and lipid levels,and the prediction efficiency of the model was verified.Results Compared with the control group,there were significant differences in C1q,C3,CFB and TC,TG and LDL-C(P<0.05).There was no significant difference between CFH and HDL-C(P>0.05).Further analysis showed that there were correlations between C3 and CFB,TC,TG,LDL-C,CFB and TC TG,TC and TG,LDL-C,and TG and LDL-C in the study group(P<0.01).Logistic regression model was constructed,which integrated variables of C1q,C3,CFB,TC,TG and LDL-C as follows:W-AMD=-12.835+8.753×CFB+6.098×TG.The specificity of the training set was 0.758,while the sensitivity was 0.843,the accuracy was 0.811,the positive predictive value was 0.852,the negative predictive value was 0.746 and the AUC value was 0.901.For validation set,the specificity was 0.630,sensitivity was 0.909,accuracy was 0.803,positive predictive value was 0.800,negative predictive value was 0.810,and AUC value was 0.943.Conclusion Serum complement factors C1q,C3,CFB and peripheral blood lipid levels of TC,TG and LDL-C in W-AMD patients are higher than those in healthy people.The potential reason could be related to its pathogenesis.Therefore,this finding will be useful for future studies to search for therapeutic targets.The prediction model of W-AMD based on complement factor and blood lipid level can accurately identify W-AMD,which provides an objective laboratory reference index for clinical evaluation and management.