Placenta accreta spectrum (PAS) is a complex disorder of uterine wall disruption with significant morbidity and mortality, particularly at time of delivery. Both physician and physical hospital resource allocation/utilization remains a challenge in PAS cases including intensive care unit (ICU) beds. The primary objective of the present study was to identify preoperative risk factors for ICU admission and create an ICU admission prediction model for patient counseling and resource utilization decision making in an evidence-based manner. This was a case-control study of 145 patients at our PAS referral center undergoing cesarean hysterectomy for PAS. Final confirmation by histopathology was required for inclusion. Patient disposition after surgery (ICU vs post-anesthesia care unit) was our primary outcome and pre-/intra-/postoperative variables were obtained via electronic medical records with an emphasis on the predictive capabilities of the preoperative variables. Uni- and multivariate analysis was performed to identify independent predictive factors for ICU admission. In this large cohort of 145 patients who underwent cesarean hysterectomy for PAS, with histopathologic confirmation, 63 (43%) were admitted to the ICU following delivery. These patients were more likely to be delivered at an earlier gestational age (34 vs 35 weeks, P < 0.001), have had >2 episodes of vaginal bleeding and emergent delivery compared to patients admitted to patients with routine recovery care (44% vs 18.3%, P = 0.009). Uni- and multivariate logistic regression showed an area under the curve of 0.73 (95% CI: [0.63, 0.81], P < 0.001) for prediction of ICU admission with these three variables. Patients with all three predictors had 100% ICU admission rate. Resource prediction, utilization and allocation remains a challenge in PAS management. By identifying patients with preoperative risk factors for ICU admission, not only can patients be counseled but this resource can be requested preoperatively for staffing and utilization purposes.
Read full abstract