Springer Nature [academic journals on nature.com], Journal of Human Hypertension, 2(24), p. 104-110, 2009
DOI: 10.1038/jhh.2009.45
Full text: Download
The study aimed to develop prediction algorithms for hypertensive disorders based on multivariate analysis of factors from the maternal history and compare the estimated performance of such algorithms in the prediction of early preeclampsia (PE), late-PE and gestational hypertension (GH) with that recommended by the National Institute for Clinical Excellence (NICE). Logistic regression analysis was used to determine which of the maternal characteristics and history had significant contribution in predicting early-PE, late-PE and GH. There were 37 cases with early-PE, 128 with late-PE, 140 with GH and 8061 cases that were unaffected by PE or GH. Predictors of early-PE were Black race, chronic hypertension, prior PE and use of ovulation drugs. Predictors of late-PE and GH were increased maternal age and body mass index, and family history or history of PE. Additionally, late-PE was more common in Black, Indian and Pakistani women. The detection rates of early-PE, late-PE and GH in screening by maternal factors were 37.0, 28.9 and 20.7%, respectively, for a 5% false positive rate. Screening as suggested by NICE would have resulted in a false positive rate of 64.1% with detection rates of 89.2, 93.0 and 85.0% for early-PE, late-PE and GH, respectively. Meaningful screening for hypertensive disorders in pregnancy by maternal history necessitates the use of algorithms derived by logistic regression analysis.