Prediction model of risk factors for hypertensive disorder complicating pregnancy

Objective To investigate the risk factors of hypertensive disorder complicating pregnancy (HDCP), and establish a model to predict the risk of developing HDCP over gestational weeks. Methods From July 2016 to August 2017, we recruited 300 pregnant women patients with HDCP (disease group) and 390 normal pregnant women (control group) who were routinely examined and delivered in Beijing Haidian Maternal and Child Health Hospital. Univariate and multivariate analysis were performed to compare the epidemiologic, hemodynamics, and laboratory variables between the two groups. Results There were significant differences in epidemiologic variables (age, multiple pregnancy, previous spontaneous abortion, HDCP history, pre-pregnancy BMI, BMI during pregnancy), hemodynamics variables (systolic blood pressure, diastolic blood pressure, pulse pressure, mean arterial pressure, K value, Cardiac index, total peripheral resistance) and laboratory variables (hematocrit, platelet, mean platelet volume, uric acid, creatinine) between the disease group and the control group at different stages of pregnancy, which facilitated the early identification of HDCP. The combination of epidemiologic, hemodynamics and laboratory variables led to great improvement in the accuracy(84.88%), sensitivity (79.75%) and specificity (89.25%) of the risk prediction model for HDCP. The prediction model was more reliable over gestational weeks, with an accuracy rate of 77.78%, lowest sensitivity of 44.44% in early pregnancy (pregnancy≤13 weeks), an accuracy rate of 83.5%, sensitivity of 71.34% in mid-pregnancy (21-27 weeks of pregnancy), while an accuracy rate of more than 91%, sensitivity of 87% and specificity of more than 94% in late pregnancy (≥28 weeks). Conclusion The establishment of a dynamic model for HDCP risk prediction model based on the associated risk factors may facilitate the early identification and intervention of HDCP, finally improve the adverse pregnancy outcome.