Prospective Evaluation Of First Trimester Prediction Methods For Pre-eclampsia

Objective1.To evaluate the characteristics and performance of the prediction models in screening for early onset pre-eclampsia.2.To evaluate the performance of first trimester risk prediction methods for pre-eclampsia in a population of singleton pregnancy in China.And to provide a reference for further exploration of pre-eclampsia prediction methods.Methods1.Systematic review:?1?Databases of PubMed,Medline,Embase,CNKI,Wanfang data were searched from inception to October 2016.Studies that established models of predicting early onset pre-eclampsia during the first trimester were included.?2?Two reviewers independently checked potentially eligible article,assessed risk of bias and extracted data.?3?Subtotal the performance of models and analysis the property of models included independent risk factors;Comparing the gap between simple models?using collected maternal characteristics?and specialized models?included blood pressure,uterine artery doppler and biomarkers?by forest plot developed by SAS 9.4.2.Prospective evaluation:?1?It was included pregnant women in the 11⁺⁰ to 13⁺⁶ weeks who underwent prenatal examination in our hospital from March 2017 to September 2017.Moreover,we collected the risk factors of pregnant women's medical history,mean arterial pressure and uterine artery pulse index,followed up regularly and recorded whether the pregnant women developed pre-eclampsia and other pregnancy outcomes.?2?Compare the performance of maternal risk factors,MAP,UtA-PI,clinical guidelines and prediction models to predict the development of pre-eclampsia.?3?Statistical Analysis:using viewpoint6.0 software to enter the maternal characteristics,UtA-PI,MAP data and pregnancy outcome,built a database.Moreover,using SPSS16.0 statistical analysis software to implement the statistical analysis.The measurement data was described by means of mean±SD,and student t-test or t-test was used to complete the comparison between groups.The count data showed as the rate,and the comparison between groups using Pearson Chi-Square test,Continuity Correction chi-square test and Fisher's exact test.P<0.05 was considered to be statistically significant.Results1.Systematic review:?1?17 eligible model development studies were identified,included 9prospective cohort studies,2 case-control studies,6 nested case-control studies.The pregnant women came from 10 different populations were assessed in model development studies.?2?.In the13 simple models,the AUC were between 0.640 and 0.806 and the sensitivity was from 21.0%to60.0%with a 10%false positive rate?FPR?.The AUC of 17 specialized models in predicting early onset pre-eclampsia ranged from 0.770 to 0.978 and the sensitivity was between 48.0%and 95.2%with the FPR fixed at 10%.?3?Compared with simple models,we found that specialized models could ensure a promotion of 0.171?range from 0.060 to 0.245?in the median AUC,and a promotion of 40.8%?16.0%to 52.2%?in the sensitivity with a 10%FPR.Add mean arterial pressure?MAP?to the simple model,its AUC and sensitivity of simple models increased by0.092?0.079 to 0.104?and 28.7%?16.2%to 55.0%?.Added uterine artery pulsatility index?UtA-PI?,the AUC and sensitivity increased by 0.106?0 to 0.137?and 31.8%?-1%to 41.9%?.And added combination the MAP and UtA-PI,the AUC and sensitivity increased by 0.157?0.094 to0.218?and 31.6%?12.0%to 52.2%?.2.Prospective evaluation:?1?700 singleton pregnancy were recruited,and 664 cases were obtained,including 15 cases of preeclampsia,6 cases of chronic hypertension combined with preeclampsia,and 8 cases of gestational hypertension.?2?Compared with the unaffected group,there was a statistical difference in the history of initial diagnosis,BMI,preeclampsia,and history of chronic hypertension.?3?The medical history factors predict preeclampsia have poor performance except history of clinical hypertension.?4?Using the MAP alone to predict preeclampsia,AUC 0.757,when the false positive rate was 10%,the sensitivity was 61.9%,UtA-PI level had no significant predictive value.?5?The three guidelines used to predict preeclampsia were not performing well.?6?The AUC of FMF simple model can reach 0.810?95%CI:0.732-0.887?,and the predicted performance will not improve after the inclusion of MAP and UTA-PI.Conclusions1.Specialized prediction models performed better than simple models in predication of early onset pre-eclampsia while it needs a further confirmation in different populations.2.Chinese population was involved to validate the performance of different predictive methods for preeclampsia prospectively in first trimester.Based on single risk factors of medical history or the guidelines,the prediction methods for pre-eclampsia perform moderate effect.A simple model with the combination of maternal risk factors may improve the performance of guideline in providing screening methods.3.The prediction tests and models reported abroad need a further confirmation in Chinese populations.