Abstract
BACKGROUND: Optimal delivery timing in the placenta accreta spectrum (PAS) remains a clinical challenge, particularly within the broad 34 to 35(+6) weeks' window recommended by current guidelines. Cervical length and antepartum bleeding have been associated with earlier delivery and increased maternal risk. This study aimed to evaluate whether a measurable algorithm based on these parameters could guide individualized delivery timing within this gestational range. METHODS: We conducted a retrospective cohort study including 145 PAS cases managed at a single tertiary center between May 2015 and December 2024. PAS cases diagnosed based on prenatal imaging, confirmed intraoperatively and by histopathology. A standardized algorithm incorporating cervical length and the number of antepartum bleeding episodes was used to determine delivery timing between 34 and 35(+6) weeks of gestation. Cervical length was assessed either serially or at least once prior to delivery. Patients were stratified into six groups based on cervical length (CL: ≥25 mm or < 25 mm) and number of antepartum bleeding episodes (0, 1, or ≥ 2). Elective and urgent deliveries were compared in terms of maternal and neonatal outcomes using independent t-test, Mann-Whitney U test, Chi-square test, and Fisher's exact test. RESULTS: A total of 145 PAS cases were included. Patients with a cervical length ≥ 25 mm and no bleeding delivered at a median of 35(+5) weeks. When one bleeding episode was present, delivery occurred earlier at 35(+1) weeks and further declined to 34(+1) weeks with two or more episodes. Among patients with a cervical length < 25 mm, median gestational age at delivery was 34(+1) weeks with no bleeding, 34(+2) weeks with one episode, and 33(+1) weeks with two or more episodes. Urgent delivery was associated with higher rates of NICU admission, ICU transfer, and transfusion of > 4 units of pRBCs (p < 0.01 for all). In ROC analysis, the optimal cut-off for cervical length was 26 mm (AUC: 0.653, 95% CI: 0.569-0.730) and for ≥ 1 bleeding episode (AUC: 0.629, 95% CI: 0.503-0.754). Combining both variables improved discrimination (AUC: 0.734, 95% CI: 0.623-0.845), with 71.4% sensitivity and 70.2% specificity. CONCLUSIONS: Cervical length and antepartum bleeding history may assist in individualizing delivery timing for PAS within the recommended 34 to 35(+6) weeks' gestational window. Further prospective studies are warranted to validate this algorithm and assess its potential clinical utility.