Identification of high-risk patients for adverse pregnancy outcome based on multivariate logistic regression analysis at 20–23 gestational weeks

Aims: To establish and test a multivariate approach for identification of patients with high risk for adverse pregnancy outcome (APO) by a combination of sonographic (20+0–23+6 weeks) and biometrical data in a test group and to prove its effectiveness in a validation group. Methods: In order to develop a multivariate approach for identifying patients at high risk for APO, data from 15,855 non-treated singleton pregnancies (test group A), with well-known outcomes, were analysed retrospectively. Registered parameters were parity, body mass index (BMI), mean of pulsatility index of both uterine arteries (meanPI), presence of notch, and depth of notch described as the mean of the notch index of both uterine arteries (meanNI). In a second step, the model was validated in 3678 subsequent non-treated singleton pregnancies (validation group B). Results: In test group A, the calculated probabilities of APO were found between 1.3% and 82.8%, with the 95th centile at 11.2% and the 99th centile at 28.3%. In validation group B, in patients with a low calculated risk for APO 28%, complications were found in 39.1%. Conclusion: A multivariate approach toward risk-assessment of APO seems to be a promising method for identifying patients who are at high risk for APO. Besides the individual consequence of intensive surveillance, the identification of groups of patients at high risk for APO might serve as a basis for interventional studies.