Selective intraoperative multi-vessel embolization for placenta accreta spectrum: a doubly robust propensity-weighted analysis of a histologically confirmed cohort.

OBJECTIVE

To evaluate the association between selective intraoperative multi-vessel embolization (SiM-VE) and perioperative transfusion during cesarean hysterectomy for histologically confirmed placenta accreta spectrum (PAS).

STUDY

DESIGN

This retrospective cohort study (2012-2024) included patients undergoing cesarean hysterectomy for histologically confirmed PAS, comparing SiM-VE with surgical devascularization alone. All SiM-VE patients underwent bilateral uterine artery embolization, with additional selective embolization of collateral/neovascular supply performed in the majority. The primary outcome was perioperative packed red blood cell transfusion. Doubly robust inverse probability of treatment weighting mitigated confounding by indication, and Oaxaca-Blinder decomposition partitioned the difference in operative duration.

RESULTS

Of 268 patients (SiM-VE n = 234; No SiM-VE n = 34), the SiM-VE group had higher anterior previa (161 [68.8%] vs 14 [41.2%]; P = 0.003), percreta (109 [46.6%] vs 7 [20.6%]; P = 0.005), and lower preoperative hemoglobin (10.9 ± 1.2 vs 11.4 ± 1.7 g/dL; P = 0.03). In the primary doubly robust model, SiM-VE was associated with lower odds of perioperative transfusion (adjusted odds ratio 0.25; 95% CI, 0.11-0.56; P = 0.001). Operative duration was longer in the SiM-VE group (315.0 ± 66.8 vs 180.2 ± 55.2 min; P < 0.001), with decomposition attributing approximately 30% of the difference to baseline complexity. Access-site arterial thrombosis requiring surgical intervention occurred in 3 (1.3%) patients.

CONCLUSIONS

SiM-VE was associated with lower adjusted odds of perioperative transfusion in this single-center cohort, representing a trade-off between lower allogeneic blood exposure and longer operative duration alongside a 1.3% risk of access-site arterial thrombosis.