Objective: To evaluate whether eclampsia can be predicted in gestational hypertension or mild preeclampsia at term.
Methods: For this case-control study we selected 76 cases with eclampsia from the LEMMoN study and 1149 controls with mild hypertensive disease of pregnancy, who did not develop eclampsia, from the HYPITAT study. Risk indicators for eclampsia, identified in multivariable logistic regression, were used to assess the predictive capacity of our model with receiver-operating characteristic (ROC) curve analysis. Model optimism was assessed with bootstrapping.
Results: Maternal age, non-Caucasian ethnicity, systolic blood pressure >155 mmHg, ≥2+ protein on dipstick, elevated uric acid, creatinin >74 μmol/L, aspartate aminotransferase >30 U/L, and lactate dehydrogenase >400 U/L were significantly associated with eclampsia. Other factors included in the model were previous fetal loss, previous miscarriage, gestational age, and low platelet count. The area under the ROC curve was 0.92. Bootstrapping showed minimal overfitting of the model.
Conclusion: In women with gestational hypertension or mild preeclampsia at term eclampsia can be predicted.