Establishment and validation of predictive model for histologic chorioamnionitis during expected treatment of preterm premature rupture of membrane before 34 weeks of gestation
10.3760/cma.j.cn113903-20220106-00018
- VernacularTitle:孕34周前未足月胎膜早破期待治疗组织学绒毛膜羊膜炎预测模型的建立和验证
- Author:
Yufei GAO
1
;
Lei ZHANG
;
Hengyu TIAN
;
Yangyu ZHAO
;
Yan ZHANG
Author Information
1. 北京大学第三医院妇产科，北京 100191
- Keywords:
Fetal membranes, premature rupture;
Chorioamnionitis;
Forecasting
- From:
Chinese Journal of Perinatal Medicine
2022;25(12):891-897
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To establish and verify a model to predict histologic chorioamnionitis (HCA) for women during expected management of preterm premature rupture of membranes (PPROM) at 24-34 weeks of gestation.Methods:This retrospective study enrolled 493 pregnant women who were diagnosed with PPROM at 24-33 +6 weeks of gestation and delivered in Peking University Third Hospital from January 1, 2012, to December 31, 2020. They were randomly divided into the modeling set ( n=345) and validation set ( n=148) at a ratio of 7∶3. Basic information, risk factors, clinical treatment, and maternal and infant outcomes were compared between participants with and without HCA using Chi-square test, t test, or Mann-Whitney U test. Multivariate logistic regression analysis was performed to analyze the independent risk factors for HCA. The predictive values of different indexes for HCA were compared and the predictive model was then established and verified using the receiver operating characteristic (ROC) curve and area under curve (AUC). Results:There were no significant differences in the basic information, common risk factors for premature rupture of membranes (PROM), or the use of tocolytics, antibiotics or dexamethasone between women in the HCA and non-HCA groups (all P>0.05). Compared with the non-HCA group, the HCA group showed an earlier onset of PROM [31.3 (24.0-33.9) vs 32.3 (27.0-33.9) gestational weeks, U=4 103.00, P=0.017], longer expected treatment [66.5 (0.7-895.3) vs 18.0 (0.3-1 123.0) h, U=1 791.00, P<0.001] and higher incidence of neonatal asphyxia, sepsis, and intracranial hemorrhage [24.3% (58/239) vs 13.2% (14/106), χ 2=5.44; 9.6% (23/239) vs 2.8% (3/106), χ 2=4.86; 41.0% (98/239) vs 17.9% (19/106), χ 2=17.45; all P<0.05]. Moreover, the positive rate of bacterial culture, C-reactive protein (CRP) level, neutrophil/lymphocyte ratio (NLR), white blood cell (WBC) and neutrophil counts were significantly increased in the HCA group [37.2% (89/239) vs 22.6% (24/106), χ 2=7.10; 8.2 (0.0-273.0) vs 5.0 (0.0-218.9) ng/ml, U=2 419.00; 5.6 (1.2-58.6) vs 4.6 (1.7-18.7), U=2 357.50; 11.9 (4.5-30.0)×10 9/L vs 10.1 (5.8-21.8)×10 9/L, U=4 074.50; 9.5 (2.5-28.1)×10 9/L vs 7.6 (3.5-18.5)×10 9/L, U=4 021.50; all P<0.05], while the lymphocyte count was decreased [1.5 (0.5-3.7)×10 9/L vs 1.6 (0.7-3.9)×10 9/L, U=4 237.00, P=0.017]. CRP level, NLR, the gestational week at the onset of PROM and the duration of expected treatment were independent risk factors for HCA ( OR=1.069, 95% CI: 1.024-1.117; OR=1.192, 95% CI: 1.048-1.356; OR=0.906, 95% CI: 0.867-0.947; OR=1.017, 95% CI: 1.007-1.026). Based on the four risk factors, the predictive model was established and ROC curve was drawn. AUC for evaluating the performance of the predictive model was 0.880, which indicated a clinical significance. Conclusion:The model established based on the four risk factors, which were CRP level, NLR, the gestational week at onset of PROM and the duration of expected treatment, performs well in the prediction of HCA in women with PPROM during expected treatment and has good clinical practical value.
