Quantifying placenta accreta spectrum severity and its associated blood loss: a novel transvaginal ultrasound scoring system

Abstract

BackgroundHemorrhage associated with placenta accreta spectrum (PAS) is a leading cause of maternal morbidity and mortality. Estimating blood loss in these individuals is a critical component of comprehensive preoperative planning. ObjectiveA semiquantitative score based on transvaginal ultrasound was developed and tested to predict PAS, estimate its severity, and blood loss in individuals with clinical and ultrasound evidence suggesting PAS. Study DesignA secondary analysis was conducted of prospectively collected data from a quaternary center of patients with suspected accreta on 2D ultrasound and clinical suspicion. A predetermined scoring system was applied based on three components: (1) uterine wall (score 0: no loss of hypo-translucent uterine wall with overlying placenta in the lower uterine segment; 1: loss of hypo-translucent <3-cm defect; 2: 3–6-cm defect; and 3: >6-cm defect); (2) arterial vascularity at the uterine wall defect (score 0: no vessels observed; 1: 1–2 vessels over the defect; 2: 3–5 vessels; and 3: >5 vessels); and (3) cervical involvement (score 0: normal cervical length without previa; 1: previa with normal cervical length; 2: short cervix with previa, minimal vascularity and small lacunae; 3: short cervix with previa, increased vascularity and large lacunae). Each patient's three domain scores determined a cumulative, final score of 0–9. Patients were managed at the discretion of a multi-disciplinary team and patient's preference among the following options: cesarean delivery with placenta removal, cesarean delivery with placenta in-situ (conservative) with or without delayed hysterectomy, or cesarean hysterectomy. The frequency of different degrees of placental invasion per pathology examination per score unit was registered. Multiple linear regression analysis was performed for association of blood loss according to score adjusted by risk factors for PAS. ResultsA total of 73 patients were evaluated. All 11 patients who had a score of 0 had cesarean delivery with placenta removal without evidence of intraoperative PAS, thus resulting in a 100% negative predictive value. The remaining 62 had scores between 1 and 9. Among patients with scores 0–3 (n=20), only one had intraoperative PAS, yielding a negative predictive value of 97%. Higher scores were associated with severe PAS forms (r=0.301, P=.02). Based on the associations between PAS scores, clinical correlation, and blood loss, we divided patients into four categories: Category 0: PAS score 0; Category 1: scores 1–3; Category 2: scores 4–6; and Category 3: scores 7–9. The median blood loss in Category 0=635±352 mL, Category 1=634±599 mL, Category 2=1549±1284 mL, and Category 3=1895±2106 mL (P<.001). On multivariable analysis, Category 2 (β=0.97, P<.01) and Category 3 (β=1.26, P<.003) were associated with significantly greater blood loss than Category 0, irrespective of type of surgery. ConclusionThe transvaginal ultrasound score separates groups at low risk (Category 0) and at higher risk of PAS (Categories 1–3). Categories 1–3 may provide important clinical information to estimate the risk of severe forms of PAS and of blood loss during surgery. Video Abstract [Display omitted]