wileyonlinelibrary.com/journal/anzjog 251 © 2018 The Royal Australian and New Zealand College of Obstetricians and Gynaecologists INTRODUCTION The prevalence of gestational diabetes mellitus (GDM) is increas- ing due to advanced maternal age, increasing proportions of women entering pregnancy who are overweight or obese 1 and implementation of the International Association of the Diabetes and Pregnancy Study Groups (IADPSG) diagnostic criteria. 2 GDM is associated with adverse maternal and neonatal outcomes, 3 and DOI: 10.1111/ajo.12833 ORIGINAL ARTICLE Role of serum biomarkers to optimise a validated clinical risk prediction tool for gestational diabetes Sally K. Abell 1,2 , Soulmaz Shorakae 1,2 , Jacqueline A. Boyle 1,3 , Barbora De Courten 1,2 , Nigel K. Stepto 4 , Helena J. Teede 1,2 and Cheryce L. Harrison 1 Aust N Z J Obstet Gynaecol 2019; 59: 251–257 1 Monash Centre for Health Research and Implementation, School of Public Health and Preventive Medicine, Monash University, Melbourne, Victoria, Australia 2 Diabetes and Vascular Medicine Unit, Monash Health, Melbourne, Victoria, Australia 3 Monash Women’s Services, Monash Health, Melbourne, Victoria, Australia 4 Institute of Sport, Exercise and Active Living, Victoria University, Melbourne, Victoria, Australia Correspondence: Dr Cheryce L. Harrison, Monash Centre for Health Research and Implementation, School of Public Health and Preventive Medicine, Locked Bag 29 Clayton, Vic. 3168, Australia. Email: cheryce.harrison@monash.edu Conficts of interest: The authors report no conficts of interest. Received: 11 February 2018; Accepted: 2 May 2018 Background: Clinical risk prediction tools for gestational diabetes (GDM) may be enhanced by measuring biomarkers in early pregnancy. Aim: To evaluate a two-step GDM risk prediction tool incorporating fasting glu- cose (FG) and serum biomarkers in early pregnancy. Materials and methods: High molecular weight (HMW) adiponectin, omentin-1 and interleukin-6 (IL-6) were measured at 12–15 weeks gestation in women with high risk of GDM from a randomised trial using a clinical risk prediction tool. GDM diagnosis (24–28 weeks) was evaluated using 1998 Australian Diabetes in Pregnancy (ADIPS) criteria and newer International Association of the Diabetes and Pregnancy Study Groups (IADPSG) criteria. Associations between biomarkers and development of GDM were examined using multivariable regression analy- sis. Area under the receiver-operator curve (AUC), sensitivity and specifcity were calculated to determine classifcation ability of each model compared to FG and maternal characteristics. Results: HMW adiponectin improved prediction of ADIPS GDM (AUC 0.85, sensi- tivity 50%, specifcity 96.2%, P = 0.04), compared to FG and maternal factors (0.78, 35% and, 98.1%, respectively). HMW adiponectin <1.53 μg/mL further improved the model (AUC 0.87, sensitivity 75%, specifcity 88.2%, P = 0.01). HMW adiponec- tin did not improve prediction of IADPSG GDM (AUC 0.84, sensitivity 64%, specifc- ity 97.9%, P = 0.22) compared to FG and maternal factors (0.79, 56%, 93.8%). Omentin-1 and IL-6 did not signifcantly improve classifcation ability for GDM. Conclusions: A two-step approach combining FG and HMW adiponectin to a vali- dated clinical risk prediction tool improved sensitivity and predictive ability for ADIPS GDM. Further research is required to enhance GDM prediction using IADPSG criteria for application in clinical practice. KEYWORDS adipocytokines, biomarkers, gestational diabetes, risk prediction