Establishment Of A Predictive Model For Intensive Care Unit Admission In Obstetric Posterior Reversible Encephalopathy Syndrome

BackgroundPosterior reversible encephalopathy syndrome(PRES)is a clinical imaging syndrome with a variety of symptoms,including headaches,visual abnormalities,seizures,nausea,vomiting,consciousness impairment,and focal neurological signs.Neuroimaging is vital for diagnosing and evaluating PRES,with typical features of vasogenic oedema in the white matter of the portions of both cerebral hemispheres,especially the posterior-predominant regions.PRES is associated with a variety of pathologies,including hypertension,preeclampsia(PE),eclampsia,renal failure,transplantation,sepsis,autoimmune disorders and chemotherapy.And Obstetric PRES,triggered by obstetric conditions,is a neuro-imaging syndrome and a critical maternal complication that is highly prevalent among PE or eclampsia women.The pathophysiology and treatment of obstetric PRES is different from other cause-related PRES,it may endanger the maternal and fetal lives in the acute disease stage,and sometimes requires admission to the intensive care unit(ICU)for aggressive monitoring and management.Therefore,rapid identification of obstetric PRES patients who may need to be admitted to the ICU is critical to improving maternal and fetal outcomes.However,due to the nonspecific clinical manifestations of obstetric PRES,clinicians may not be able to identify and manage the critical patients timely,and thus leading to adverse pregnancy outcomes.Presently,the risk factors for admission to the ICU of obstetric PRES patients remain unclear,and there have been no studies on predictive models for admission to the ICU of obstetric PRES patients.ObjectiveThe main purpose of our study was to explore the risk factors related to ICU admission of obstetric PRES patients through retrospective analysis of clinical manifestations,laboratory examination indicators and imaging changes,then construct a prediction model and to provide a more intuitive and objective quantitative reference tool for clinicians to identify and attach the critical pregnant women that may admit to ICU timely,so as to achieve the purpose of early identification and early intervention,thereby improving the prognosis of patients.MethodsAccording to the inclusion criteria and exclusion criteria,366 pregnant women who were hospitalized in the Guangzhou Medical Center for Critical Pregnant Women from October 2012 to December 2021 and met the diagnostic criteria of PRES were included for analysis.According to whether the pregnant women with PRES were admitted to the ICU during hospitalization,94 patients were grouped into the ICU admission group and 272 patients were grouped into the non-ICU admission group.And we retrospectively collected all patients’ clinical data,laboratory examination indicators and imaging data through the hospital’s electronic medical record system and imaging system,including maternal age,gestational week and blood pressure at the onset of symptoms,prenatal body mass index(BMI),pregnancy history of maternal,previous diseases,acute stage comorbidities,clinical symptoms and signs,hospital days,pregnancy outcome,complete blood count,blood biochemistry,coagulation function,liver and kidney function,lactic acid level and the location and severity of neuroimaging cerebral edema,etc.The general clinical data,laboratory test indicators,imaging examinations and pregnancy outcomes of the two groups were compared.The independent predictors of ICU admission of obstetric PRES patients were identified by univariate analysis and multivariate regression logistic analysis,and the risk prediction model of obstetric PRES patient admission to ICU was constructed by R software,and receiver operating characteristic(ROC)curve,decision curve analysis(DCA)and calibration curve was used to assess the calibration and clinical utility of the predictive model.Results1.Among the 366 obstetric PRES patients,the average age was 32.3±6.2 years,the average gestational week was 30.7±6.0 weeks,the median BMI was 22.7 kg/m2,the mean systolic blood pressure was 167.8±25.0 mm Hg and the diastolic blood pressure was 106.0±16.1mm Hg.Of all the patients,94(25.7%)patients were admitted to the ICU in the acute stage,and a total of 176(48.1%)patients presented with adverse pregnancy outcomes,including 47(50.5%)in the ICU admission group and 129(47.4%)in the non-ICU admission group,with no significant difference between subgroups.2.Univariate analysis of two groups showed that the gestational week,the rates of eclampsia and HELLP syndrome,the rates of seizures and altered consciousness,neutrophil to lymphocyte ratio(NLR),C-reactive protein level,white blood cell(WBC)count,lactate dehydrogenase(LDH),aspartate amino transaminase(AST),alanine amino transferase(ALT),lactic acid,Uric acid(UA),urea nitrogen(BUN),cerebral edema involved area points and oedema degree scores were much higher in the ICU admission group than those in the non-ICU admission group,while the patients age,albumin(ALB)level and platelet(PLT)count were significantly lower than those in the non-ICU group(P < 0.05).3.Multivariate regression logistic analysis of relevant risk factors in the two groups showed that the levels of ALB(OR=0.915,95%CI 0.850-0.985,P=0.018),CRP(OR=1.011,95% CI 1.002-1.020,P=0.015),lactic acid(OR=1.828,95%CI1.258-2.656,P=0.002),oedema degree score(OR=1.505,95% CI 1.092-2.074,P=0.012)and altered consciousness(OR= 5.137,95% CI 2.198 to 12.002,P <0.001)were independent predictors of ICU admission in obstetric PRES.4.The levels of CRP,ALB,lactic acid,altered consciousness and oedema degree score were included in the risk prediction model of obstetric PRES patients admitted to ICU,a nomogram was constructed,and ROC curve analysis was performed to test the prediction ability of this model,the area under curve(AUC)was 0.874,the Bootstrap method was used for internal verification.The calibration curve analysis results showed that there has high-fitness between the prediction curve and the standard model curve with an average absolute error of 0.024,the DCA analysis results showed that the risk model had high clinical application value.ConclusionsThe levels of ALB,CRP,lactic acid,the occurrence of altered consciousness and the oedema degree score were independent predictors of ICU admission in maternal with PRES.And the risk prediction model established by the above risk factors had a good predictive value,which was helpful for early identification of high-risk obstetric PRES patients requiring ICU transfer,thereby improving clinical outcomes.