Can miscarriage be predicted in pregnant women who present to the emergency department in the first trimester?

To evaluate the relationship between physician-assessed vaginal bleeding and subsequent miscarriage in first-trimester pregnancies, and to identify independent clinical predictors that could form a simple risk estimation model. This prospective cohort study included pregnant women in their first trimester who presented to the emergency department, with or without vaginal bleeding, between January 2023 and January 2024. All participants underwent standardized clinical and ultrasonographic evaluation by the same physician. Patients were categorized into two groups: those with active vaginal bleeding (defined as bleeding similar to menses) and those without bleeding. Binary logistic regression was used to identify predictors of miscarriage. Of the 451 participants, 260 (57.6%) presented with vaginal bleeding and 191 (42.4%) without bleeding. Miscarriage occurred in 69 (26.5%) of the bleeding group and 11 (5.8%) of the non-bleeding group (p < 0.01). Gestational age at presentation, vaginal bleeding, and history of vaginal delivery were identified as independent predictors of miscarriage. The regression model—0.296 (vaginal delivery) + 0.998 (bleeding) – 0.433 (gestational age)—showed good fit (Nagelkerke R² = 0.231, p = 0.01). Active vaginal bleeding, early gestational age, and prior vaginal delivery are significantly associated with miscarriage. These parameters may serve as preliminary, easily obtainable predictors for early identification of high-risk pregnancies, though external validation in larger, multi-center studies is warranted.