First-trimester bleeding : can ultrasound and biochemical developmental markers deepen our understanding of threatened miscarriage?

Abstract

Introduction: Understanding the gestational behaviour of ultrasound and biochemical markers in early pregnancy is essential for improving clinical prediction. This study applies statistical modelling to explore how these markers evolve across gestation, with a particular focus on comparing their trajectories in women presenting with threatened miscarriage (TM) versus those with asymptomatic pregnancies. Methods: A prospective case–control study was conducted with 177 participants, divided into a study group (TM group) and a control group. Dynamic markers were modelled against gestational age using linear, quadratic, and segmented linear regressions. Statistical comparisons of model coefficients were used to evaluate differences both within pregnancy outcome groups (live birth vs loss) and between study and control cohorts. Results: Several markers, including fetal heart rate (FHR), trophoblast thickness, yolk sac (YS) wall thickness, beta-human chorionic gonadotropin (β-hCG), and mean amniotic sac diameter (MASD), were best represented by segmented linear models, reflecting distinct phase-dependent changes across gestational age. Linear models fit other markers such as progesterone, mean gestational sac diameter, and the soluble fms-like tyrosine kinase-1 to placental growth factor ratio. Statistically significant differences in regression coefficients were observed within the study group between outcome categories, as well as across groups, particularly for FHR, β-hCG, YS wall thickness and MASD. These differences support the notion that TM follows a physiologically distinct developmental trajectory compared to normal pregnancy. Conclusion: First-trimester markers in women with TM follow gestational trajectories that diverge from those in asymptomatic pregnancies. This physiological distinction underscores the need for tailored modelling to better reflect the unique risk profile in TM. Implications for practice: TM-specific models are crucial for accurate risk assessment and personalised care. Applying general thresholds may misclassify risk in TM, leading to unnecessary intervention or missed opportunities for reassurance.