Predictors of intraoperative conversion to cesarean hysterectomy during attempted conservative management of placenta accreta spectrum: a multicenter cohort study.

BACKGROUND: Conservative management of placenta accreta spectrum (PAS) is increasingly attempted to reduce maternal morbidity and preserve fertility. However, in a substantial proportion of cases, uterine preservation cannot be maintained and conversion to cesarean hysterectomy is required before skin closure, a distinct clinical event whose predictors and associated morbidity remain poorly described. OBJECTIVE: We aimed to determine the incidence of intraoperative conversion to cesarean hysterectomy during attempted conservative PAS management, to identify its preoperative predictors, and to characterize the associated maternal morbidity. METHODS: Retrospective multicenter cohort study conducted between January 2013 and December 2023 across three French tertiary PAS referral centers. Women with antenatal PAS suspicion in whom conservative management was attempted were included. Conservative management comprised leaving the placenta partially or completely in situ or performing focal resection of the invaded myometrium and adherent placenta, followed by uterine reconstruction. Intraoperative conversion was defined as cesarean hysterectomy performed before skin closure due to uncontrolled hemorrhage refractory to hemostatic measures. Preoperative predictors were identified by univariate and multivariable logistic regression analysis. RESULTS: Among 189 women with antenatal PAS suspicion, 101 in whom conservative management was initially attempted were included. Intraoperative conversion to cesarean hysterectomy occurred in 29 of 101 women (28.7%), corresponding to nearly one in three women undergoing attempted conservative management. In the conversion group, median estimated blood loss was 3,600 mL (IQR 2,500-6,000) versus 600 mL (IQR 300-1,525) in the successful group (P<.001), with bladder injuries in 13.8% versus 0% (P=.006), and requirement for conversion to general anesthesia in 82.8% versus 33.3% (P<.001). On multivariable analysis, antenatal vaginal bleeding was independently associated with intraoperative conversion (aOR 2.85, 95% CI 1.10-7.39; P=.031). The model had an AUC of 0.70 (95% CI 0.59-0.80), indicating modest discriminative ability. CONCLUSION: Intraoperative conversion to cesarean hysterectomy occurs in nearly one in three women undergoing attempted conservative PAS management and is associated with substantially higher maternal morbidity than successful uterine preservation. Antenatal vaginal bleeding independently predicts this event. It may help guide preoperative counseling and surgical planning.