Background: There is little knowledge about neonatal complications in GH and PE and induction at term, we aim to assess whether they can be predicted from clinical data. Methods: We used data of the HYPITAT trial and evaluated whether adverse neonatal outcome (Apgar score < 7, pH < 7.05, NICU admission) could be predicted from clinical data. Logistic regression, ROC analysis and calibration were used to identify predictors and evaluate the predictive capacity in an antepartum and intrapartum model. Results: We included 1153 pregnancies, of whom 76 (6.6%) had adverse neonatal outcome. Parity (primipara OR 2.75), BMI (OR 1.06), proteinuria (dipstick +++ OR 2.5), uric acid (OR 1.4) and creatinine (OR 1.02) were independent antepartum predictors; In the intrapartum model, meconium stained amniotic fluid (OR 2.2), temperature (OR 1.8), duration of first stage of labour (OR 1.15), proteinuria (dipstick +++ OR 2.7), creatinine (OR 1.02) and uric acid (OR 1.5) were predictors of adverse neonatal outcome. Both models showed good discrimination (AUC 0.75 and 0.78), but calibration was limited (Hosmer-Lemeshow p = 0.41, and p = 0.20). Conclusions: In women with GH or PE at term, it is difficult to predict neonatal complications, possibly since they are rare in the term pregnancy. However, the identified individual predictors may guide physicians to anticipate requirements for neonatal care.
|Number of pages||7|
|Journal||The Journal of Maternal-Fetal and Neonatal Medicine|
|Publication status||Published - 1 Jan 2015|
- Neonatal outcome
- Prediction models