Abstract
Pregnant women are universally screened for gestational diabetes (GDM) at 24-28 weeks gestation because
of the well-established link between hyperglycemia and adverse pregnancy outcomes. In the past decade,
hemoglobin A1c (A1C), which measures the percentage of glycated hemoglobin in red blood cells (RBCs), has
transformed the diagnosis of diabetes outside of pregnancy. While A1C has modernized diabetes diagnosis in
non-pregnant individuals, pregnant women continue to be diagnosed with GDM using cumbersome oral
glucose tolerance tests (OGTTs), which require fasting and multiple timed blood draws, and have problems
with intra-individual reproducibility. Despite its potential advantages, A1C has not been adopted to screen for
GDM because it is affected by pregnancy-related changes in RBC kinetics, rendering simple A1C-based
inferences of glycemia during gestation unreliable. We and others have demonstrated how pregnancy disrupts
the strong relationship between A1C and glycemia due to pregnancy-related changes in RBC kinetics. In
previous work, we have also successfully used mechanistic modeling to adjust A1Cs for non-glycemic variation
outside of pregnancy. The goal of this proposal, Glycemic Observation Using A1C for Gestational Diabetes
Diagnosis (GO A1C GDM), is to optimize A1C’s ability to detect hyperglycemia in pregnancy by adjusting for
gestational changes in RBC kinetics that affect A1C’s relationship with glycemia. We will leverage accurate
longitudinal glycemic measurements from continuous glucose monitoring (CGM) and rigorous ascertainment of
hyperglycemia-associated adverse outcomes in 2150 pregnant women participating in the Glycemic
Observation and Metabolic Outcomes in Mothers and Offspring study (GO MOMs) to accomplish this goal. In
GO MOMs, participants will have A1Cs measured and undergo serial 10-day periods of CGM monitoring at 4
time points across pregnancy. Our proposal, GO A1C GDM, adds serial hematologic measurements (CBCs,
reticulocyte counts, and ferritin) across gestation and employs mechanistic modeling to improve A1C-based
glycemia estimation during pregnancy. In Aim 1, we will adjust A1C for typical gestational changes in RBC
kinetics (GA-adjusted A1C). In Aim 2, we will personalize A1C adjustments, using hematologic measurements
to capture gestational changes in RBC kinetics specific to individuals (CBC-adjusted A1C). In Aim 3, we will
test the ability of A1Cs adjusted for RBC kinetics to predict hyperglycemia-associated adverse outcomes. We
will compare this predictive ability to that of traditional OGTT-based GDM diagnosis. The proposed
investigations have potential to greatly simplify the method by which we diagnose GDM, delivering advances in
precision diabetes screening to the entire obstetric population.