19 Accuracy of fullpiers score for prediction of morbidity/mortality in patients with severe preeclampsia in Northeastern Brazil

Introduction Hypertension is the first cause of maternal mortality in Brazil. In the manegement of these patients it is important to have tools for predicting those with more chance for developing complications. The objective of this study is to evaluate the accuracy of fullPIERS score for prediction of morbidity/mortality in patients with severe preeclampsia in northeastern Brazil. Methods An accuracy assessment study was conducted from January to December 2014, in Instituto de Medicina Integral Prof. Fernando Figueira (IMIP), including patients with severe preeclampsia and excluding those with chronic hypertension or other chronic diseases. The fullPIERS score was calculated at admission and the patients were followed until hospital discharge. Statistical analysis was performed using Medcalc 16.4.3. A ROC (Receiver Operator Curve) curve was constructed to determine the accuracy of fullPIERS score for predicting complications and the Area Under the Curve (AUC) was calculated. Multiple logistic regression analysis was performed to determine the ajusted risk of maternal complications. Results Of the 325 women enrolled, 55 had one or more complications (16.9%); there were no maternal deaths. The area under the curve was 72% ( p  0.0001), with a cut off > 1.7%, for which the accuracy parameters were determined: sensitivity of 60%, specificity 65.2%, positive likelihood ratio of 2.69 and negative likelihood ratio of 0.5. After multivariable analysis, the only variable significantly associated with maternal complications was fullPIERS score > 1.7% (OR = 5.87; 95% CI = 3.16–10.89). Conclusion The fullPIERS model is a useful tool to predict complications in women with severe preeclampsia, improving the quality of care. We suggest its implementation in the clinical practice with special care to patients with initial score greater than 1,7% in our service.