Prediction of progression to severe disease in women with late preterm hypertensive disorders of pregnancy.

ABSTRACT

INTRODUCTION: If hypertensive disorders of pregnancy are diagnosed before term, the benefits of immediate delivery need to be weighed against the neonatal consequences of preterm delivery. If we are able to predict which women are at high risk of progression to severe disease, they could be targeted for delivery and maternal complications might be reduced. In addition, this may prevent unnecessary preterm births in women at low risk. MATERIAL AND METHODS: We developed a prediction model using data from the HYPITAT-II trail, which evaluated immediate delivery vs. expectant monitoring in women with non-severe hypertensive disorders of pregnancy between 34 and 37 weeks of gestation. Univariate and multivariate logistic regression analysis were used to identify relevant variables from clinical and laboratory parameters. The performance of the resulting prediction model was assessed by receiver operating characteristic analysis, calibration and bootstrapping, using the average predicted probabilities. RESULTS: We included 519 women, 115 (22.2%) of whom developed severe hypertensive disorders of pregnancy. The prediction model included maternal age (odds ratio 0.92 per year), gestational age (odds ratio 0.87 per week), systolic blood pressure (odds ratio 1.05 per mmHg), the presence of chronic hypertension (odds ratio 2.4), platelet count (odds ratio 0.996), creatinine (odds ratio 1.02) and lactate dehydrogenase (odds ratio 1.003). The model showed good fit (p = 0.64), fair discrimination (area under the curve 0.76, 95% confidence interval 0.73-0.81, p < 0.001) and could stratify women in three risk groups of average, intermediate and high risk (predicted probabilities <0.22, <0.44 and >0.45, respectively). CONCLUSION: In women with non-severe hypertension in pregnancy near term, progression to severe disease can be predicted. This model requires external validation before it can be applied in practice.