Application of the Surgical APGAR Score to Predict Intensive Care Unit Admission and Post-Operative Outcomes in Cesarean Hysterectomy for Placenta Accreta Spectrum

Background and Objective: Placenta Accreta Spectrum (PAS) encompasses a continuum of abnormal placentation conditions associated with significant maternal and fetal morbidity. Management of PAS requires coordinated cesarean hysterectomy. Associated morbidities include blood transfusion, coagulopathy, and intensive care unit (ICU) admission. Accurate prediction of ICU admission allows for enhanced multidisciplinary management, coordination of care and utilization of resources. Scoring systems exist in other surgical specialties that can predict the likelihood of ICU admission, but these have not been applied to an obstetric population. The SAS is a 10-point scale that has been validated for the prediction of ICU-level care requirements within 72 h post-operatively in numerous surgical specialties. The purpose of this study was to apply the Surgical APGAR Score (SAS, version 9) to patients undergoing management of PAS to determine if it can predict ICU admission in this population. Materials and Methods: This is a case–control study. We retrospectively analyzed 127 cases of pathology-confirmed PAS patients who underwent cesarean hysterectomy in singleton, non-anomalous, viable pregnancies. Our primary outcome was ICU admission. In addition, secondary outcomes included antepartum characteristics, operative time, intraoperative events as well as post-operative complications and total postoperative length of stay. SAS was assigned by extracting estimated blood loss (EBL), and the lowest mean intraoperative heartrate (HR and mean arterial pressure (MAP) from intraoperative documentation. Categorical and continuous factors were summarized using frequencies and percentages or means ± SD or median and range as appropriate. Pearson’s chi-square, Fisher’s exact tests, and Mann–Whitney U and t-tests were applied when appropriate. Logistical regression to assess the impact of SAS on ICU admission was performed. p-values < 0.05 were considered significant for two-tailed analysis. Statistical analysis was performed using Graphpad software (version 9). Results: Fifty-eight patients (45%) were admitted post-operatively to the ICU, while 69 patients (55%) were admitted for routine care to the post-anesthesia care unit. Baseline demographics were similar between groups. Forty-four patients (52%) admitted to the ICU had a SAS score < 4. SAS < 4 was associated with greater blood loss (3000 vs. 2500 mL, p = 0.03) and longer operative time (198 vs. 175 min, p = 0.03). Logistic regression analysis of SAS score and ICU admission revealed a low predictive value (OR 2.28, AUC = 0.599). Conclusions: The SAS system is a poor tool for the prediction of ICU admission in patients with PAS undergoing cesarean hysterectomy. A risk calculator that accounts for the unique physiologic changes in pregnancy and high risk for pregnancy is needed.