Analysis of risk factors for hearing loss caused by neonatal cytomegalovirus infection and construction of a nomogram prediction model
10.3760/cma.j.cn112866-20230407-00036
- VernacularTitle:新生儿巨细胞病毒感染致听力受损的危险因素分析及列线图预测模型构建
- Author:
Yan GUO
1
;
Li XU
;
Qinliang ZHENG
;
Baiting NI
;
Yanling TAO
Author Information
1. 济宁医学院附属医院儿科，济宁 272029
- Keywords:
Cytomegalovirus;
Neonatal;
Hearing impairment;
Nomogram
- From:
Chinese Journal of Experimental and Clinical Virology
2023;37(4):398-403
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To analyze the risk factors of hearing loss caused by neonatal cytomegalovirus (CMV) infection, to construct a nomogram prediction model, and to evaluate its discrimination and consistency.Methods:Data of a total of 250 neonates with congenital and perinatal CMV infection who were treated in our hospital from June 2018 to March 2021 were collected, and grouped into hearing-impaired group of 50 cases and hearing-unimpaired group of 200 cases according to whether the hearing was impaired or not. Logistic regression analysis was performed to analyze the risk factors affecting hearing loss caused by CMV infection in neonates. R software was applied to construct a nomogram model for predicting the risk of hearing loss caused by CMV infection in neonates, and the ROC curve and calibration curve were analyzed to verify the discrimination and consistency of the nomogram model.Results:Among 250 CMV-infected neonates in the modeling group, the incidence of hearing impairment was 20.00%. The proportion of primary maternal infection before 14 weeks of gestation, the proportion of urine CMV viral load >5×10 4 copies/ml, the proportion of symptomatic CMV infection and the proportion of abnormal head MRI in the hearing-impaired group were significantly higher than those in the hearing-unimpaired group ( χ2=12.545, 11.812, 11.801, and 9.025, P=0.000, 0.001, 0.001, and 0.003). That urine CMV viral load>5×10 4 copies/ml, symptomatic CMV infection and abnormal head MRI were the risk factors affecting hearing loss caused by CMV infection in neonates ( OR=2.388, 6.765, and 4.993, P=0.022, 0.000, and 0.001), the prediction model of the line graph was constructed using Bootstrap internal verification, Hosmer-Lemeshow goodness of fit test =6.055, P=0.301, area under ROC curve was 0.817, 95% CI=0.754~0.880, and external verification result showed that the predicted probability of the calibration curve was close to the actual probability and has a good agreement, with an AUC of 0.791 (95% CI=0.716~0.866). Conclusions:The nomogram model constructed in this study to predict the risk of hearing loss caused by CMV infection in neonates can be used for the prediction of hearing loss in neonates after CMV infection.
