Research On FullPLERS Model Of Hypertensive Disorders In Pregnancy To Predict Adverse Outcomes

Objective:Discussion fullPIERS predictive model for adverse outcomes of pregnancy inducedhypertension predicted effect, then explore a number of factors that affect patients ofhypertensive disorders in pregnancy adverse outcomes.Method:1. Collected the clinical symptoms, relevant laboratory data and clinical outcomecorresponding to the situation of1,430cases of HDP patients with before its pregnancyoutcome appears within48h, and organized into an Excel spreadsheet.2.The use of fullPIERS predictive models to calculate HDP1430cases of patientswith risk values (see related calculations website:. Https://piers.cfri.ca), further assessmentfullPIERS model for patients with hypertensive disorders in pregnancy predict the effectsof adverse outcomes3. Use SPSS19.0software for maternal age, platelets, gestational age, total bilirubin,direct bilirubin, indirect bilirubin, hemoglobin, total protein, albumin, globulin, AST, ALT,creatinine, alkaline phosphatase, lactate dehydrogenase, urea, uric acid, an independentsample t test and multivariate Logistic regression analysis. And formula creatinine, AST,gestational age, platelet multivariate Logistic regression analysis. And then screened forChina’s hypertensive disorders in pregnancy to predict the risk of significant factors. Result:1. The receiver operating characteristic curve (ROC) area of1430patients in usingSPSS19.0software is0.768. Combined with the ROC curve, when the risk values≥4.5%of the patients with hypertensive disorders in pregnancy were269,18.81%of the totalnumber, including the number of adverse outcomes for144people, the proportion was53.53%, and its true positive rate54.96%, the false positive rate was0.107.2. The data collected in different time windows categories:(1)6hours time window:the area under the ROC curve was0.782. Combined ROC, when the risk is0.061, it is thehighest AUC. Risk value≥6.1%of patients are48people HDP,17.27%of the total, ofwhich the number of adverse outcomes for34people, a ratio of70.83%occurs,false-positive rate was0.056.(2) the time window of24hours: the area under the ROCcurve was0.770. Combined ROC, when the risk is0.054, it is the highest AUC. Risk value≥5.4%of patients with HDP101people,23.11%of the total number, including thenumber of adverse outcomes for60people, the proportion was59.41%, and its truepositive rate was60.61%, the false positive rate was0.12.(3) the time window of48hours:the area under the ROC curve was0.754. Combined ROC, when the risk is0.041, it is thehighest AUC. Risk value≥4.1%of patients with a total of113HDP,15.80%of the totalnumber, including the number of adverse outcomes was45, his false positive rate0.11.3. Because of the hospital found in drug use and treatment of severehypoproteinemia antihypertensive and/or severe anemia is not completely standardized,and the study of three drugs used to treat hypertension and blood transfusion therapy are inthe collection of medical records HDP as adverse outcomes in patients with one, it will usethe three drugs to treat hypertension after rounding draw this one area of the ROC curve0.768, and compare with the original data, P>0.05, no significant difference. And thisafter a transfusion therapy rounding the resulting ROC area of0.769, and its comparativeanalysis with the original data, derived P <0.05, statistically significant differences exist.4. Extremes: The study found a total of nine cases of extreme phenomena, including (1) one case of maternal mortality, the risk value of only0.086, if the adjustment inpatients with creatinine values or platelet values to the normal range, the risk values>99%.One case of adverse outcomes in patients with HDP did not occur, the risk value is as highas0.867, if the adjustment in patients with creatinine values or platelet values to thenormal range, the risk value dropped to29.4%.(2) the risk of an example of epilepsypatients,1case of pulmonary edema and HDP HELLP syndrome patients and4cases ofplacental abruption HDP value of0.002,1cases of patients with the risk of placentalabruption value of only0.001. Analysis of these extreme cases found in patients withgestational age, platelet, creatinine and AST were normal, and lactate dehydrogenase,alkaline phosphatase, uric acid, albumin are varying degrees of abnormality, there may beother factors to consider.(3) If after removing these extreme cases analyzed, the ROCcurve can be drawn from an area of0.792, more than0.768, comparison with previous dataobtained P <0.05, statistically significant differences exist.While rounding transfusiontherapy and extreme cases, the value of each risk and the corresponding final clinicaloutcomes of pregnant mothers case, Use SPSS19.0software come to draw AUC ROCROC curve was0.791, P <0.05. And statistical analysis to only rounding extreme cases,the P <0.05, statistically significant.5. Because if rounding using three or more antihypertensive medications that adverseoutcomes were statistically no significant difference, it is only while rounding the adverseoutcomes of transfusion therapy and in extreme cases, the area under the ROC curveobtained is0.793, and in extreme cases and only for statistical rounding analysis, the P <0.05, statistically significant.6. Use SPSS19.0software for independent samples t-test analysis, the resultgestational age, platelets, total bilirubin, direct bilirubin, indirect bilirubin, albumin, AST,ALT, alkaline phosphatase, lactate dehydrogenase enzymes, creatinine, uric acid P <0.05,a significant difference. Research on these factors Multivariate Logistic regression analysisresults suggest that HDP patients with gestational age, creatinine, platelets, AST, lactatedehydrogenase are meaningful factor, P <0.05. The lactate dehydrogenase with adverse outcomes drawn from the analysis area under the ROC curve0.615, cut-off point for the243.5U/L.Conclusion:1. fullPIERS of HDP model can be used to predict adverse outcomes in patientswhose best prediction time window is less than six hours, but not entirely suited to China.2. For subsequent verification fullPIERS derive suitable models or patients withadverse outcomes when our HDP prediction formula, the adverse outcomes of transfusiontherapy may need to be properly analyzed and choices.3. fullPIERS model creatinine and platelet values in the non-independence of thefactors that predict adverse outcomes in patients with HDP when there is interactionbetween the two. And fullPIERS model for the study of very severe patients did notproceed. Hence the need to further explore the relationship between the laboratoryparameters related variables and very severe patients how to adjust the HDP and limited.4. Related factors predictive of adverse outcomes in patients with HDP addition togestational age, creatinine, platelets, AST, we experimental results suggest that lactatedehydrogenase may also be meaningful factor. However, the sample size of this study andthe variable factors also too small, it still needs to be further explored and perfected.