Early-Pregnancy Red Cell Distribution Width (RDW-SD) and Gestational Diabetes Mellitus: A Prospective Cohort Study in Aden, Yemen

Background: Gestational diabetes mellitus (GDM) is a common pregnancy complication with significant maternal and fetal consequences. Red cell distribution width (RDW) has been proposed as an accessible biomarker for metabolic dysregulation, but studies have largely used RDW-CV, a relative measure influenced by mean corpuscular volume (MCV). The comparative predictive performance of RDW-SD (absolute distribution width) and RDW-CV for early GDM prediction remains unresolved. Objectives: This study aimed to: (1) compare the discriminatory performance of baseline RDW-SD and RDW-CV for subsequent GDM; (2) assess whether RDW-SD adds incremental predictive information beyond baseline glucose measurements; and (3) evaluate the specificity of RDW-SD by examining its performance for a distinct outcome (incident anemia). Methods: In a prospective cohort of 300 pregnant women in Aden, Yemen, we measured RDW-SD, RDW-CV, and glucose parameters at 13–18 weeks (T1) and classified GDM at 24–28 weeks (T2) using a sequential screening pathway. We compared discrimination (AUC) using the DeLong test, adjusted associations, and incremental value of RDW-SD beyond glucose measurements using multivariable logistic regression and nested cross-validation. Internal validation was performed with out-of-fold predictions and paired bootstrap uncertainty analysis. Decision curve analysis evaluated net benefit.