E-Mail karger@karger.com Original Paper Fetal Diagn Ther DOI: 10.1159/000369970 First-Trimester Screening for Gestational Diabetes Mellitus Based on Maternal Characteristics and History Argyro Syngelaki a, b Alice Pastides a Reena Kotecha a Alan Wright a Ranjit Akolekar a, b Kypros H. Nicolaides a, b a Harris Birthright Research Centre for Fetal Medicine, King’s College Hospital, London, and b Department of Fetal Medicine, Medway Maritime Hospital, Gillingham, UK characteristic curve of the new model was higher (p < 0.0001) than that of the previous five models (0.823 vs. 0.688–786). Conclusions: Early effective screening for GDM can be achieved based on maternal characteristics and history. © 2014 S. Karger AG, Basel Introduction Gestational diabetes mellitus (GDM) is associated with an increased risk of maternal and perinatal short- and long-term complications [1–6]. The condition is di- agnosed by a positive oral glucose tolerance test (OGTT), which is either carried out in all pregnant women [7] or in a selected group of women identified by their demo- graphic characteristics and obstetric history as being at high risk for GDM [8]. In the UK, OGTT is offered to women with any one of the following risk factors: body mass index (BMI) >30, development of GDM in a previ- ous pregnancy, previous delivery of a macrosomic baby ( ≥4.5 kg), first-degree relative with diabetes mellitus, or racial origin with a high prevalence of diabetes such as South Asian, African-Caribbean and Middle Eastern [8]. We have previously suggested that in screening for GDM it would be preferable to combine the various ma- ternal factors into a multivariate logistic model, rather Key Words Gestational diabetes mellitus · First-trimester screening · Pyramid of pregnancy care Abstract Objectives: To develop and validate a prediction model for gestational diabetes mellitus (GDM) at 11–13 weeks’ gesta- tion based on maternal characteristics and history and to compare its performance with the method recommended by the National Institute of Health and Care Excellence (NICE) and five other published prediction models. Methods: A pre- dictive logistic regression model for GDM was developed from 1,827 cases (2.4%) who developed GDM and 73,334 un- affected controls. A 5-fold cross-validation study was per- formed to validate this model and to compare its perfor- mance with those of the NICE guidelines and the previously published models. Results: In the logistic regression model, maternal age, weight, height, racial origin, family history of diabetes, use of ovulation drugs, birth weight, and previous history of GDM were found to be significant predictors of GDM. In screening for GDM in the 5-fold cross-validation study, detection rates (DRs) were higher (p < 0.0001) for the proposed model (DR = 83.2%) than for the NICE guidelines (DR = 77.5%) for a false positive rate of approximately 40% (determined by NICE). The area under the receiver operating Received: October 21, 2014 Accepted: November 17, 2014 Published online: December 18, 2014 Prof. K.H. Nicolaides Harris Birthright Research Centre for Fetal Medicine King’s College Hospital, Denmark Hill London SE5 9RS (UK) E-Mail kypros  @  fetalmedicine.com © 2014 S. Karger AG, Basel 1015–3837/14/0000–0000$39.50/0 www.karger.com/fdt Downloaded by: UCL 82.23.63.209 - 12/23/2014 10:28:04 PM