摘要
目的 分析不孕症患者宫腔内人工授精妊娠结局的影响因素,并构建列线图预测模型。方法 采用病例对照研究选取2023年6月—2024年7月在吉林省妇幼保健院生殖中心接受宫腔内人工授精治疗的不明原因不孕症患者66例为研究对象,共行135个助孕周期,通过系统随访追踪至妊娠结局,分为活产组和非活产组。收集患者的年龄、不孕年限、体质指数、不孕类型、周期方案、优选后精子前向运动情况、抗米勒管激素(AMH)水平及基础卵泡刺激素(FSH)水平;将妊娠≥28周娩出存活胎儿,出生后存活>24 h定义为活产。通过Lasso回归分析和logistic回归分析筛选患者宫腔内人工授精妊娠结局的影响因素,使用R语言构建列线图预测模型。结果 非活产组患者年龄[(32.65±2.07)岁]高于活产组[(29.81±2.39)岁],体质指数[(20.61±1.38) kg/m^(2)]、周期方案为促排卵比例30.25%、优选后精子前向运动水平[(25.71±3.80)×10^(6)]及AMH水平[(3.45±0.68) ng/ml]均低于活产组[(21.93±1.26) kg/m^(2)、68.75%、(26.85±3.27)×10^(6)及(3.97±0.52) ng/ml],差异均有统计学意义(t/χ^(2)=10.529、5.037、45.289、4.255及5.819,均P<0.05)。两组不孕年限、不孕类型及基础FSH水平比较,差异均无统计学意义(均P>0.05)。将各临床指标作为变量纳入Lasso回归分析,共筛选出5个非零系数的变量:年龄、体质指数、周期方案、优选后精子前向运动及AMH;此时Lasso回归模型的拟合效果最理想。将Lasso回归分析筛选出的5个变量进行单因素logistic回归分析,共筛选出4个变量:年龄、体质指数、周期方案及AMH;进一步纳入多因素logistic回归分析,结果显示,年龄、周期方案及AMH均是不孕症患者宫腔内人工授精妊娠结局的独立影响因素(均P<0.05)。基于logistic回归分析筛选出的独立影响因素,使用R语言构建年龄、周期方案及AMH对不孕症患者宫腔内人工授精妊娠结局的列线图预测模型,用Bootstrap内部重复抽样法对建模人群数据重复抽样1 000次,验证模型的区分度,结果显示,一致性指数(C-index)=0.939,接近理想曲线,说明模型具有较高的校准能力。进一步行受试者工作特性(ROC)曲线验证,ROC曲线下面积(AUC)值为0.973,说明该模型在预测不孕症患者宫腔内人工授精妊娠结局时具有较高的性能。结论 年龄、周期方案及AMH均是不孕症患者宫腔内人工授精妊娠结局的影响因素,联合构建的列线图预测模型对不良妊娠结局具有较高的预测价值。
Objective To analyze the factors influencing pregnancy outcomes of intrauterine insemination in infertility patients and to construct a nomogram prediction model.Methods A case-control study was conducted,selecting 66 patients with unexplained infertility who underwent intrauterine insemination treatment at the Reproductive Center of Jilin Women and Children Health Hospital from June 2023 to July 2024,totaling 135 assisted cycles.Patients were followed up systematically until pregnancy outcomes were determined and were divided into a live birth group and a non-live birth group.Data collected included patient age,duration of infertility,body mass index(BMI),type of infertility,cycle protocol,forward motile sperm count after sperm processing,anti-Müllerian hormone(AMH)level,and basal folliclestimulating hormone(FSH)level.Live birth was defined as the delivery of a live fetus at≥28 weeks of gestation surviving for more than 24 hours after birth.Factors influencing intrauterine insemination pregnancy outcomes were screened using Lasso regression analysis and logistic regression analysis.A nomogram prediction model was constructed using the R language.Results The age of patients in the non-live birth group[(32.65±2.07)years]was higher than that in the live birth group[(29.81±2.39)years],the BMI[(20.61±1.38)kg/m^(2)],the proportion of ovulation induction cycles(30.25%),the level of forward motile sperm after processing[(25.71±3.80)×10^(6)],and the AMH level[(3.45±0.68)ng/ml]in the non-live birth group were lower than those in the live birth group[(21.93±1.26)kg/m^(2),68.75%,(26.85±3.27)×10^(6),and(3.97±0.52)ng/ml],the differences were statistically significant(t/χ^(2)=10.529,5.037,45.289,4.255,and 5.819,all P<0.05).There were no statistically significant differences between the two groups in the duration of infertility,type of infertility,or basal FSH level(all P>0.05).Various clinical indicators were included as variables in the Lasso regression analysis,which screened out 5 variables with non-zero coefficients:age,BMI,cycle protocol,forward motile sperm after processing,and AMH;at this point,the Lasso regression model demonstrated the most ideal fit.These 5 variables screened by Lasso regression were subjected to univariate logistic regression analysis,which identified 4 variables:age,BMI,cycle protocol,and AMH.These were further included in multivariate logistic regression analysis,the results showed that age,cycle protocol,and AMH were independent influencing factors for intrauterine insemination pregnancy outcomes in infertility patients(all P<0.05).Based on the independent influencing factors screened by logistic regression analysis,a nomogram prediction model incorporating age,cycle protocol,and AMH for predicting intrauterine insemination pregnancy outcomes in infertility patients was constructed using the R language.The Bootstrap internal resampling method was used to resample the modeling population data 1000 times to validate the model's discrimination.The results showed a consistency index(C-index)=0.939,close to the ideal curve,indicating high calibration ability of the model.Further validation using the receiver operating characteristic(ROC)curve showed an area under the curve(AUC)value of 0.973,indicating high performance of the model in predicting intrauterine insemination pregnancy outcomes in infertility patients.Conclusion Age,cycle protocol,and AMH are all influencing factors for intrauterine insemination pregnancy outcomes in infertility patients.The combined nomogram prediction model constructed has high predictive value for adverse pregnancy outcomes.
作者
李光
朱珊珊
唐克
张振武
LI Guang;ZHU Shan-shan;TANG Ke;ZHANG Zhen-wu(Jilin Women and Children Health Hospital(Jilin Provincial Obstetric Quality Control Center),Changchun,Jilin 130061,China)
出处
《中国妇幼保健》
2025年第24期4546-4550,共5页
Maternal and Child Health Care of China
基金
吉林省卫生健康科技能力提升项目(2023LC087)。
关键词
不孕症
宫腔内人工授精
妊娠结局
影响因素
列线图预测模型
Infertility
Intrauterine insemination
Pregnancy outcome
Influencing factors
Nomogram prediction model
