Methods:
In this study we used demographic, clinical and biobank data from the Screening for Pregnancy Endpoints (SCOPE) study (trial registration ACTRN 12607000551493).35 Ethical approval was obtained from local ethics committees. SCOPE was a prospective multicentre cohort study of healthy nulliparous women, undertaken between 2004 and 2011 in study centres in Australia (Adelaide), Ireland (Cork), NZ (Auckland) and the United Kingdom (Leeds, London and Manchester). Informed consent was obtained from each individual participant. The primary objective of the SCOPE study was to develop screening tests to predict pre-eclampsia, small for gestational age births and spontaneous preterm birth, based on a detailed set of clinical risk factors and early pregnancy biomarkers measured at 15±1 weeks gestation. Therefore, all women with major clinical risk factors for one of these adverse outcomes were excluded from SCOPE, including women with pre-existing hypertension or diabetes. Other exclusion criteria were based on obstetric and gynaecological history, and complications of the index pregnancy.35Women with uncomplicated singleton pregnancies were recruited before 15 weeks gestation and assessed for a wide range of demographic, psychological and lifestyle factors.36 Physical measures (such as height, weight, waist and hip circumference, and blood pressure) and the collection of peripheral blood samples, was performed at two time points: at 15±1 and 20±1 weeks gestation. All women were followed until delivery.
Maternal body mass index (BMI) was calculated as per the international World Health Organization definitions.37 In addition, ethnic-specific BMI criteria for Asian populations were reported,38 to adjust for variations in body composition by ethnicity such as the ratio of body fat and lean tissue mass.39 In the SCOPE study the maternal socioeconomic index was calculated using the NZ Socioeconomic Index 1996,40 including for the non-NZ study centres. Socioeconomic status is described by this index on a continuous scale from 10 to 90, with 10 being the lowest socioeconomic score and 90 the highest. This continuous score is subsequently divided into six classes, with women of lowest socioeconomic status in the 6thclass (score <24). Maternal anaemia at 15±1 weeks gestation was defined as haemoglobin <110 g/L. Diet was categorised as not vegetarian, demi-vegetarian (i.e. does not eat meat, but may eat fish; as defined by SCOPE), and vegetarian (a diet without meat or fish). GDM was diagnosed based on national guidelines of the respective countries included in SCOPE. At the time of recruitment it was universal practice to screen for GDM in Australia and NZ, but not in the United Kingdom or Ireland, where only risk factor screening was performed. The composite variable for hypertensive disorders diagnosed during pregnancy included chronic hypertension, gestational hypertension, preeclampsia and superimposed preeclampsia. Birthweight was adjusted for maternal height, booking weight, parity, ethnicity and infant sex, to create customised birthweight centiles.41
Serum samples were tested by Luminex or ELISA immunoassays (Alere, San Diego) on 57 biomarkers. In 2014, stored samples were additionally tested for serum lipids, insulin, adiponectin and alanine aminotransferase. For this study we were interested in early pregnancy biomarkers in relationship to metabolic health. Placental, metabolic, inflammatory and cardiovascular biomarkers were therefore selecteda priori on the basis of a possible association with the metabolic syndrome, diabetes or GDM. Five biomarkers were subsequently excluded, as over 50% of women had a value below the lowest limit of detection (insulin, human placental growth hormone, pentraxin-3, plasminogen activator inhibitor-1, and endothelin-1). See supplemental table S1 for a list of biomarkers included in this study, with more detail on immunoassay methods. Finally, 21 original biomarkers were included, plus lipid variables. Not all tests were performed on all women (see table 2 for n per biomarker), but data missingness appeared to be random by study centre. Women of Indian ethnicity were analysed as the population of interest, and mothers of European ethnicity were included as the referent group.