绝经后女性的生殖特点与骨质疏松症的相关性分析及个体化模型预测

Correlation analysis and individualized model prediction between reproductive characteristics of postmenopausal women and osteoporosis

目的探究绝经后女性的生殖特点及骨质疏松症的相关性,确定影响因素并为之建立个体化预测模型。方法回顾性分析350例绝经后女性一般资料、OSTA评分、生殖特点及骨密度资料,通过独立样本t检验及二元Logistic回归分析确定预测因子,使用预测因子构建绝经后骨质疏松症患病风险的诺模图,并进行了内验证。结果年龄、BMI、初潮年龄、绝经年龄及妊娠次数与绝经后骨质疏松症的发生存在相关性,二元Logistic回归分析准确度显示,以年龄、BMI、初潮年龄、绝经年龄及妊娠次数预测骨质疏松症发病的准确率为72.6%,高于以OSTA评分预测骨质疏松症发病的准确率(69.4%)。将5个因素作为女性绝经后骨质疏松症的独立预测因素,合并到诺模图中,构建的诺模图显示了良好的校准性和区分度,原始数据集AUC为0.7926,内验证集的AUC为0.8141。结论我们设计的对绝经后女性患骨质疏松风险的预测模型,综合绝经后女性年龄、BMI、初潮年龄、绝经年龄及妊娠次数,对个体进行患病预测,可以方便快速地对绝经后女性患有骨质疏松症风险进行个体化预测,该模型有助于预测绝经后女性患有骨质疏松症的风险大小以及决定是否需要对患者进行进一步骨质疏松症的相关检查,在收集的实验数据中,本模型有较高的准确度。

Objective To explore the relationship between reproductive characteristics and osteoporosis in postmenopausal women,to identify the influencing factors,and to establish an individualized prediction model.Methods The general data,OSTA score,reproductive characteristics,and bone mineral density data of 350 postmenopausal women were retrospectively analyzed.The independent sample t test and binary logistic regression analysis were used to determine the predictors.The predictors were used to construct the prevalence of postmenopausal osteoporosis and the nomogram.The internal risk was validated.Results Age,BMI,age of menarche,age of menopause,and number of pregnancies were correlated with the occurrence of postmenopausal osteoporosis.The binary logistic regression analysis showed that the accuracy of the incidence of osteoporosis predicted with age,BMI,age of menarche,age of menopause,and number of pregnancies was 72.6%,which was higher than the accuracy of predicting the incidence of osteoporosis by OSTA score(69.4%).Five factors as independent predictors of postmenopausal osteoporosis in women were merged into a nomogram.The constructed nomogram showed good calibration and discrimination.The AUC was 0.8141.Conclusion We have designed a prediction model for the risk of osteoporosis in postmenopausal women.By combining age,BMI,age of menarche,age of menopause,and number of pregnancies in postmenopausal women,we can predict the disease of individuals,which can be used conveniently and rapidly for postmenopausal women.The model helps to predict the risk of osteoporosis in postmenopausal women and to decide whether patients need further osteoporosis related tests.The data collected in the study show that this model has higher accuracy.