Improved maternal outcome in placenta accreta spectrum disorders: a single-institute observational cohort study.

Within the 11-year period, our study was performed based on the inclusion of 474 consecutive patients with clinically confirmed PAS disorders. Antenatal characteristics of the enrolled population from 2013 to 2023 were summarized in Table 2 . There were significant differences in maternal age, parity and gravity through the study period. Previous cesarean delivery was the predominant high-risk factor, with significant increase from 82.4% in initial period to 97.5% in transition period and a slight decrease to 91.6% in current period ( P  < 0.001). The rate of previous surgical scar, other than those undergoing cesarean delivery, gradually increased from 3.9% in 2013–2017 to 11.8% in 2020–2023. Nearly all patients had placenta previa, with the rates of 99.3%, 99.5% and 97.5% in initial, transition, and current periods, respectively ( P  = 0.307). Furthermore, 24.4% of the enrolled patients received MRI in current period, compared with 74.5% and 50.0% in initial and transient periods, respectively ( P  < 0.001). However, the diagnostic accuracy fluctuated around 85%, which seemed not to be affected by reduction of MRI referral. Meanwhile, patients in transition and current periods, in relative to the initial period, demonstrated higher median preoperative hemoglobin level [110 g/L (range 74–143 g/L) and 113 g/L (range 89–133 g/)] vs. 107 g/L (range 80–140 g/L]; P  = 0.001), along with decreased preoperative anemia rate ( P  = 0.015). Table 2 Antenatal characteristics of patients with PAS disorders Patient characteristics 2013–2023 N  = 474 2013–2016 N  = 153 2017–2019 N  = 202 2020–2023 N  = 119 P value Clinical characteristics  Maternal age (years) 33 (20–47) 32 (20–43) 32 (21–47) 33 (22–41) 0.005  Parity 4 (1–9) 4 (1–9) 3 (2–9) 4 (1–9) 0.080  Gravity 1 (0–4) 1 (0–3) 1 (0–3) 1 (0–4) <0.001  Maternal weight (kg) 70 (46–105) 70 (46–98) 70 (55–105) 70 (46–94) 0.406  Maternal height (cm) 160 (148–180) 160 (148–174) 160 (150–180) 160 (150–175) 0.868  BMI (kg/m 2 ) 27.2 (18.4–38.6) 27.3 (19.1–36.4) 27.3 (21.6–38.6) 37.0 (18.4–38.6) 0.535 High-risk factors  History of CS 432 (91.1) 126 (82.4) 197 (97.5) 109 (91.6) <0.001  Other surgical scar 29 (6.1) 6 (3.9) 9 (4.5) 14 (11.8) 0.012  IVF procedure 24 (5.1) 6 (3.9) 8 (4.0) 10 (8.4) 0.158  Other non-surgical scar 7 (1.5) 0 (0.0) 1 (0.5) 6 (5.0) 0.001  Adenomyosis 10 (2.1) 4 (2.6) 3 (1.5) 3 (2.5) 0.716 Prenatal diagnosis  Diagnostic method - Ultrasound 474 (100.0) 153 (100.0) 202 (100.0) 119 (100.0) - MRI 244 (51.5) 114 (74.5) 101 (50.0) 29 (24.4) <0.001  Type of placenta previa 0.307 Marginal 41 (8.6) 15 (9.8) 19 (9.4) 7 (5.9) Partial 7 (1.5) 1 (0.7) 5 (2.5) 1 (0.8) Complete 421 (88.8) 136 (88.9) 177 (87.6) 108 (90.8)  Suspected PAS 399 (84.2) 134 (87.6) 164 (81.2) 101 (84.9) 0.255  Preoperative Hb 110 (74–143) 107 (80–140) 110 (74–143) 113 (89–133) 0.001  preoperative anemia 227 (47.9) 86 (56.2) 95 (47.0) 46 (38.7) 0.015 Other surgical scar: myomectomy, surgical termination of cesarean scar pregnancy, transcervical resection of adhesions, incision of uterine septum and curettage Other non-surgical scar: uterine artery embolism, intra-uterine device and previous accreta Abbreviations : CS cesarean section, IVF In Vitro Fertilization, Hb  hemoglobin Antenatal characteristics of patients with PAS disorders Other surgical scar: myomectomy, surgical termination of cesarean scar pregnancy, transcervical resection of adhesions, incision of uterine septum and curettage Other non-surgical scar: uterine artery embolism, intra-uterine device and previous accreta Abbreviations : CS cesarean section, IVF In Vitro Fertilization, Hb  hemoglobin Table 3 presented the intraoperative characteristic of PAS disorders. At cohort level, the proportion of recommended-delivery(34–36 weeks) increased to 30.3%, and the proportion of late-delivery (≥ 36 weeks) decreased to 57.1% ( P  = 0.024), which was still the leading part. The composition of PAS disorders remained statistically stable during the study period, with a decline to 19.3% in PAS1 and an increase to 48.7% in PAS3 ( P  = 0.181). Meanwhile, there was no statistical difference in the rates of initiating emergent surgery ( P  = 0.055). Similar to the use trend in intrauterine tamponade and uterus artery embolism, the rate of prophylactic balloon occlusion application was significantly declined from 81.7% to 0.0%. As one important outcome, uterine preservation rate, one important outcome in our study, steadily went up to 97.5%, compared with 81% in initial period and 93.7% in transition period ( P  < 0.001). We also observed a slight increase of the median intraoperative blood loss from 800 ml in initial period to 1200 ml in current period ( P  = 0.003)(Fig. 2 A). However, unlike the rise in blood loss, there was a small reduction in allogenic blood transfusion rate, revealing no statistical significance (33.3% vs. 29.7% vs. 27.7%; P  = 0.586). This study continued to analyze the characteristics of blood consumption, as shown in Table S2 and Fig. 2 B. Specifically, autologous transfusion rate fluctuated between 29.2% and 53.8%. With cell salvage implemented, the procedure became the main source of autologous blood donation (reaching 44.5%), while the rate of preoperative autologous blood donation decreased from 37.9% to 21.0% ( P  < 0.001 and P  = 0.001, respectively). Similarly, in terms of the characteristics of overall intraoperative allogenic blood transfusion, there was a 3% reduction every period, yet without any obvious difference. In addition, no differences were observed in transfusion rates and volume of packed red blood cell, fresh frozen plasma, as well as platelet and cryoprecipitate among different periods. Table 3 Intraoperative characteristics of patients with PAS disorders Patient characteristics 2013–2023 N  = 474 2013–2016 N  = 153 2017–2019 N  = 202 2020–2023 N  = 119 P value Basic characteristics  Gestational age (weeks) 36 (27–40) 36 (28–40) 35 (27–39) 36 (28–38) 0.002 Late-delivery (≥ 36 weeks) 269 (56.8) 102 (66.7) 99 (49.0) 68 (57.1) 0.024 Recommended-delivery (34–36 weeks) 149 (31.4) 38 (24.8) 75 (37.1) 36 (30.3)  Early-delivery (<34 weeks) 56 (11.8) 13 (8.5) 28 (13.9) 15(12.6)  Emergent surgery 99 (20.9) 22 (14.4) 49 (24.3) 28 (23.5) 0.055  General anesthesia 163 (34.4) 57 (37.3) 64 (31.7) 42 (35.3) 0.534 Intraoperative hemostatic procedures  Prophylactic ballon occlusion 190 (40.1) 125 (81.7) 65 (32.2) 0 (0.0) <0.001  Ligation of ascending branch of uterine artery 159 (33.5) 14 (9.2) 66 (32.7) 79 (66.4) <0.001  Satinsky clamp 68 (14.3) 0 (0.0) 45 (22.3) 23 (19.3) <0.001  Tourniquet 2 (0.4) 0 (0.0) 0 (0.0) 2 (1.7) 0.050  Intrauterine tamponade 228 (48.1) 108 (70.6) 105 (52.0) 15 (12.6) <0.001 Ballon tamponade 131 (57.5) 45 (41.7) 72 (68.6) 14 (93.3) <0.001 Gauze packing 97 (42.5) 63 (58.3) 33 (31.4) 1 (6.7)  Uterine artery embolism 7 (1.5) 7 (4.6) 0 (0.0) 0 (0.0) <0.001 Maternal outcomes  Intraoperative blood loss (ml) 1000 (200–13600) 800 (200–6000) 1000 (200–13600) 1200 (300–6800) 0.003  Intraoperative blood loss ≥ 1000 ml 239 (50.4) 65 (42.5) 103 (51.0) 71 (59.7) 0.019  Intraoperative blood loss ≥ 2000 ml 87 (18.4) 23 (15.0) 34 (16.8) 30 (25.2) 0.075  Intraoperative blood loss ≥ 3000 ml 49 (10.3) 13 (8.5) 18 (8.9) 18 (15.1) 0.139  Uterus Preservation 429 (90.5) 124 (81.0) 189 (93.6) 116 (97.5) <0.001  Intraoperative allogenic blood transfusion 144 (30.4) 51 (33.3) 60 (29.7) 33 (27.7) 0.586 Type of PAS 0.181  PAS1 106 (22.4) 42 (27.5) 41 (20.3) 23 (19.3)  PAS2 169 (35.7) 57 (37.3) 74 (36.6) 38 (31.9)  PAS3 199 (42.0) 54 (35.5) 87 (43.1) 58 (48.7) 3a 152 (76.4) 38 (70.4) 72 (82.8) 42 (72.4) 0.311 3b 46 (23.1) 16 (29.6) 14 (16.1) 15 (25.9) 3c 1 (0.5) 0 (0.0) 1 (1.1) 1 (1.7) Intraoperative characteristics of patients with PAS disorders 1000 (200–13600) 800 (200–6000) 1000 (200–13600) 1200 (300–6800) As for postoperative outcomes (Table 4 ), median overall blood loss increased from 970 ml in initial period to 1340 ml in current period ( P  = 0.034). Compared with initial and transition periods, patients in current period showed higher hemoglobin levels postoperatively than those preoperatively [−7 g/L (−67 to 22 g/L) vs. −3 g/L (−55 to 41 g/L) and − 2 g/L (−37 to 37 g/L); P  = 0.017], along with similar trend in median length of postoperative hospital stay. Early composite morbidity, halved from 23.5% initially to 11.8% currently ( P  = 0.044), with similar trend noticed in massive blood transfusion (≥ 4 units of RBC) and adverse complications/events ( P  = 0.032 and 0.006, respectively). Besides, the incidence of early reoperation declined significantly from 5.2% to 0.8% ( P  = 0.034), but coupled with a steady late composite morbidity. There were less complications/events during the study period (12.4% vs. 4.5% vs. 4.2; P  = 0.006). Table 4 Postoperative characteristics of patients with PAS disorders Patient characteristics 2013–2023 N  = 474 2013–2016 N  = 153 2017–2019 N  = 202 2020–2023 N  = 119 P value Overall blood loss (ml) 1117.5 (200–1360) 970 (200–6110) 1117.5 (250–13600) 1340 (368–7550) 0.034 Postoperative Hb (g/L) 107 (60–150) 107 (67–150) 107 (60–146) 106 (60–147) 0.715 Hb change (g/L) −3 (−67 to 41) −2 (−37 to 37) −3 (−55 to 41) −7 (−67 to 22) 0.017 Length of postoperative hospital stay (Days) 5 (2–31) 5 (2–31) 5 (3–20) 6 (3–18) 0.003 Postoperative allogenic blood transfusion 49 (10.3) 17 (11.1) 19 (9.4) 13 (10.9) 0.847 Early composite morbidity 86 (18.1) 36 (23.5) 36 (17.8) 14 (11.8) 0.044  Admission to intensive care unit 20 (4.2) 7 (4.6) 10 (5.0) 3 (2.5) 0.559  Transfusion of 4 units or more red blood cells 81 (17.1) 35 (22.9) 33 (16.3) 13 (10.9) 0.032  DIC 1 (0.2) 0 (0.0) 1 (0.5) 0 (0.0) 0.509  Ureteral injury 6 (1.3) 3 (2.0) 2 (1.0) 1 (0.8) 0.642  Early reoperation 12 (2.5) 8 (5.2) 3 (1.5) 1 (0.8) 0.034 Relaparotomy for hysterectomy 6 (1.3) 3 (2.0) 2 (1.0) 1 (0.8) - Relaparotomy for surgical hemostasis 1 (0.2) 0 (0.0) 1 (0.5) 0 (0.0) - Intrauterine tamponade 1 (0.2) 1 (0.7) 0 (0.0) 0 (0.0) - Uterine artery embolism 3 (0.6) 3 (2.0) 0 (0.0) 0 (0.0) - Dilation and curettage 1 (0.2) 1 (0.7) 0 (0.0) 0 (0.0) - Death 0 (0.0) 0 (0.0) 0(0.0) 0 (0.0) - Late composite morbidity 23 (4.9) 11 (7.2) 8 (4.0) 4 (3.4) 0.255  Infection 19 (4.0) 11 (7.2) 5 (2.5) 3 (2.5) 0.051  Readmission within 6 weeks 5 (1.1) 1 (0.7) 3 (1.5) 1 (0.8) 0.724  Delayed operation 1 (0.2) 1 (0.7) 0 (0.0) 0 (0.0) 0.350 Adverse complications/events 33 (7.0) 19 (12.4) 9 (4.5) 5 (4.2) 0.006  Bowel obstruction 8 (1.7) 7 (4.6) 1 (0.5) 0 (0.0) 0.003  Urinary retention 3 (0.6) 2 (1.3) 1 (0.5) 0 (0.0) 0.382  Hematoma of surgical site 1 (0.2) 0 (0.0) 1 (0.5) 0 (0.0) -  Impaired wound healing 1 (0.2) 0 (0.0) 1 (0.5) 0 (0.0) -  Interventional radiology complications 1 (0.2) 1 (0.7) 0 (0.0) 0 (0.0) -  Pulmonary embolism 1 (0.2) 0 (0.0) 0 (0.0) 1 (0.8) - Early composite morbidity: admission to intensive care unit, transfusion of 4 units or more red blood cells, DIC, ureteral injury, early reoperation or death Late composite morbidity: infection, hospital readmission within 6 weeks, or delayed reoperation Early reoperation: relaparotomy for hysterectomy or hemostatic procedure, intrauterine tamponade, uterine artery embolism and dilation and curettage Adverse complications/complications: bowel obstruction, urinary retention, hematoma of surgical site, impaired wound healing, interventional radiology complications and pulmonary embolism Abbreviations : HB hemoglobin, DIC  disseminated intravascular coagulation Postoperative characteristics of patients with PAS disorders 1117.5 (200–1360) 970 (200–6110) 1117.5 (250–13600) 1340 (368–7550) Early composite morbidity: admission to intensive care unit, transfusion of 4 units or more red blood cells, DIC, ureteral injury, early reoperation or death Late composite morbidity: infection, hospital readmission within 6 weeks, or delayed reoperation Early reoperation: relaparotomy for hysterectomy or hemostatic procedure, intrauterine tamponade, uterine artery embolism and dilation and curettage Adverse complications/complications: bowel obstruction, urinary retention, hematoma of surgical site, impaired wound healing, interventional radiology complications and pulmonary embolism Abbreviations : HB hemoglobin, DIC  disseminated intravascular coagulation

This retrospective cohort study was approved by the Ethic Review Committee of GL Hospital (Approval No.: 2023−239-01). The study was conducted following the rules of Revised Standards for Quality Improvement Reporting Excellence (SQUIRE 2.0, 2015) and the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline [ 31 , 32 ]. Initially, this study included 498 patients between January 1 st, 2013 and December 31 st,2023, including 418 patients suspected prenatally and 80 patients with identified condition during cesarean delivery. Eventually, 474 pregnancies of clinically confirmed PAS disorders (Fig. 1 ) were incorporated for subsequent analyses. The excluded cases included 2 patients less than 26 weeks’ gestation, 5 patients for intrauterine fetal demise, 5 duplicated cases and 12 patients for non-invasion of placenta during surgery. During this study time, our analyses were conducted within three distinct period: an initial period (January 1 st, 2013 - December 31 st, 2016), a current period (January 1 st, 2020 - December 31 st, 2023), and a 3-year transition period (2017–2019), with staged introduction of various elements of improvement (Table 1 ). Fig. 1 Flow chart for the inclusion and exclusion of study population Table 1 Quality improvement elements for PAS disorders Quality improvement elements Periods Description Peripartum care Multidisciplinary team 2012-Up to now Formulation of principles for patients’ assessment, management and follow-up. Programmatic update of peripartum care plan, checklist, standardized surgical and anesthetic approach by regular team audit. Optimization of iron supplementation 2012–2016 Measurement of ferritin with hemoglobin level < 110 g/L, followed by oral and/or parenteral iron supplementation where indicated. 2017-Up to now Routine examination of hemoglobin and ferritin around 20 weeks of gestational age, followed by oral and/or parenteral iron supplementation where indicated. Prenatal diagnosis Ultrasound screening 2012–2016 Screening based on the sonographer’s experience, mainly by observation of loss of “clear zone”, myometrial thinning, placental bulge and bridging vessels. No application of uniform sonographic signs or format. 2017-Up to now In accordance with the proposal by the European Working Group on Abnormally Invasive Placenta, emphasis on focal exophytic mass and utero-vesical hypervascularity with sensitivity of 98% and specificity of 56% in predicting PAS > 3a base on a Chinese article published by our Center. Simultaneous application of placenta mapping, drawing boundary of placenta localization and area of different grades of invasion, to guide the incision of the uterus. Adjunctive MRI 2012–2017 Use of MRI as an adjunct to most of patients suspected of PAS disorders. 2018-Up to now Application of MRI in predicting placenta increta and percreta with sensitivity of 79.7% and specificity of 83.1%, respectively. No priority of MRI than ultrasound in diagnostic accuracy, with the use of MRI in the case of requirement for further examination of the depth of invasion in certain patients. Surgical plan Prophylactic ballon occlusion 2012–2017 Application of prophylactic balloon occlusion to most of patients suspected of placenta increta and percreta preoperatively. 2018-Up to now Gradual removal of the procedure from common practice due to the clash of schedule on hybrid operating room, alongside potential complications. Surgical practice 2012–2017 Infiltration of the balloon to achieve temporal pelvic devascularization if needed. Manual removal of placenta and compression suture of invasion area without dissecting vesicouterine peritoneum and putting down the bladder. No intentional preservation of the placenta in situ as a strategy of conservative management. 2018–2020 Dissection of vesicouterine peritoneum to ligate newly formed vessels and put down bladder. Placement of a Satinsky clamp or tourniquet as low as possible at the lower segment of the uterus before manual removal of the placenta for uterus conservation. No excision of the area where placenta was deeply adherent. 2021-Up to now Excision of the anterior wall of the uterus where placenta was deeply adherent. Implementation of other procedures according to the practice between 2018 and 2020. Autologous blood transfusion Preoperative autologous blood donation 2012-Up to now Implementation of preoperative autologous donation for high-risk patients (hemoglobin > 110 g/L), with the maximum collection of 300 ml per time, and the minimum interval of repeated collection set to 2 weeks to allow recovery from anemia. Planning of the surgery with 5 weeks from the first collection, but not within 1 week from last collection. Initiation of the collection schedule at 32 weeks of gestation for patients with PAS disorders generally. Acute normovolemic hemodilution 2012–2016 Implementation of acute normovolemic hemodilution to patients (hemoglobin > 110 g/L) and planning of the surgery within 1 week. Cell salvage 2019-Up to now Implementation of cell salvage as a main source of autologous blood transfusion. Flow chart for the inclusion and exclusion of study population Quality improvement elements for PAS disorders In accordance with the proposal by the European Working Group on Abnormally Invasive Placenta, emphasis on focal exophytic mass and utero-vesical hypervascularity with sensitivity of 98% and specificity of 56% in predicting PAS > 3a base on a Chinese article published by our Center. Simultaneous application of placenta mapping, drawing boundary of placenta localization and area of different grades of invasion, to guide the incision of the uterus. A multidisciplinary team was established in 2012 to provide routine consultation services involving Obstetrics, Imaging (Ultrasound and Radiology), Anesthesiology, Neonatology and Transfusion medicine. Services consulted on as needed were Gynecologic Oncolgy, Urology, Critical Care Medicine and Interventional Radiology. The multidisciplinary team was responsible for patient-specific peripartum care, prenatal diagnosis and surgical plan, as well as advancement of the quality improvement program. Mos of the quality elements were initiated round 2017, including iron supplementation optimization, institutional diagnostic protocol development, surgical plan refinement, and multiple autologous blood preparation protocol implementation (Fig. 2 ). Fig. 2 Time trend of blood loss and RBC transfusion during the study period. A Time trend of mean blood loss (ml) by year. B Time trend of Mean volume of RBC (red blood cell) (ml) administrated per patient. In China, one unit of red blood cell was collected from 200ml of whole blood. Therefore, allogenic units are converted to milliliters here using the Chinese average volume per unit (200 mL) in order to compare uniform standard measures between autologous and allogenic blood. Purple lines indicate the temporal trends calculated by cubic spline for non-linear fit Time trend of blood loss and RBC transfusion during the study period. A Time trend of mean blood loss (ml) by year. B Time trend of Mean volume of RBC (red blood cell) (ml) administrated per patient. In China, one unit of red blood cell was collected from 200ml of whole blood. Therefore, allogenic units are converted to milliliters here using the Chinese average volume per unit (200 mL) in order to compare uniform standard measures between autologous and allogenic blood. Purple lines indicate the temporal trends calculated by cubic spline for non-linear fit The primary outcomes were uterine preservation rate and early maternal morbidity, defined as ≥ 1 occurrence of the following: admission to intensive care unit, transfusion of ≥ 4 red blood cells, disseminated intravascular coagulation, ureteral injury, early reoperation or death. The secondary outcomes included estimated intraoperative blood loss, allogenic blood transfusion, and late composite morbidity (including infection, readmission within 6 weeks, or delayed reoperation). The grade of PAS was made based on the clinical criteria as detailed in the FIGO Clinical Classification system, which was introduced clinically in our Center since 2019. Diagnosis prior to 2019 were updated based on the description of operation records by two core obstetricians in the multidisciplinary team. The definitions of variables and outcomes were provided in Table S1 [ 33 – 36 ]. All cases were followed up at 42 days postpartum, with all the analyzed variables sourced from medical records. Analysis was performed by R 4.3.3. Given their non-normally distributed characteristics, continuous variables were reported as median and range, and analyzed by Kruskal-Wallis H tests. Categorized variables expressed as frequencies (proportion) were compared by chi-square test or Fisher’ exact test. Statistical significance was defined as P  < 0.05. Figures were graphed using mean values to interpret average blood loss and transfusion volumes. Moreover, trends of graphed lines were calculated by restricted cubic splines to accommodate a nonlinear relationship.

Placenta accreta spectrum (PAS) is a life-threatening obstetric complication characterized by abnormal implantation of the villi into/through the myometrium [ 1 ]. As a trigger of major obstetric hemorrhage, this condition is associated with severe maternal morbidity, such as multi-organ dysfunction, acute respiratory distress, disseminated intravascular coagulation and even death [ 2 , 3 ]. PAS disorders exhibit steady rise in their incidence from 1/2510 in the 1980 s to 1/533 in 2000 s in the United States [ 4 , 5 ]. In a large survey from 2015 to 2016 in China, the prevalence of PAS reached 2.20% due to a high cesarean section rate [ 6 ]. Indeed, prior cesarean delivery has been recognized as the strongest risk factor for PAS, presenting with an increase observed with the escalated numbers of prior cesareans [ 7 , 8 ]. Meanwhile, the risk of PAS disorders in subsequent pregnancies may also be attributable to procedures which may disturb the integrity of the uterine lining, such as uterine curettage, hysteroscopic surgery and uterine artery embolization [ 9 ]. Critically, strategies in peripartum management have been outlined in guidelines on PAS published around 2018 [ 10 – 15 ]. However, it remains controversial with regard to the optimal management of this disorder [ 16 ]. According to existing evidence, patients receiving cesarean hysterectomy might experience improved maternal outcomes when undergoing systemic multidisciplinary management [ 17 – 19 ]. Meanwhile, scheduled delivery, rather than urgent delivery, can also reduce hemorrhagic complications [ 20 , 21 ]. Currently, there has been a definition of the suggested criteria for “center of excellence” regarding multidisciplinary team, intensive care unit, blood services, antenatal diagnosis, patient referral and intraoperative management at laparotomy. Nevertheless, given limited resource settings, especially in low- and middle-income countries, there may still be variance among protocols in different institutes [ 16 , 17 , 22 ]. Our Research Center has set up a multidisciplinary team to handle patients with suspected PAS disorders, with a purpose to mitigate the impact of clinical and logistical dilemma. Preoperative and intraoperative care plans were primarily established around 2012. With reference to published literature and guidelines, we have held comprehensive self-audit including monthly multidisciplinary team meeting and implemented postoperative follow-up to discuss items for improvement suitable for clinical practice [ 23 – 30 ]. Accordingly, the present study was carried out to describe the preoperative and intraoperative improvements in women with clinically confirmed PAS disorders over a 11-year period, and to assess the effect on maternal outcomes from the aspects of composite maternal morbidity and uterine preservation rate.

To sum up, the application of multiple quality improvement elements is crucial for patients with PAS disorders receiving uterine preservation. The quality improvement program involves two key changes, namely, the removal of certain elements such as routine MRI and invasive hemostatic procedures; and the introduction of specific measures, including iron supplementation and various forms of autologous blood preparation.

In our study, the implementation of various elements of quality improvement program could contribute to increased uterine preservation and reduced early composite morbidity. Moreover, there was steady rate of allogenic transfusion, despite an increase in intraoperative blood loss, possibly attributable to elevated preoperative hemoglobin level and subsequent autologous blood donation. According to previous retrospective cohort studies, management by core teams within regional centers of excellence can yield better outcomes clinically [ 19 , 33 ]. Through management within these dedicated multidisciplinary programs, patients with PAS disorders were less likely to receive massive blood transfusion, admit to intensive care unit, occur reoperation or adverse complications, outperforming the outcomes of standard obstetric care [ 19 , 33 , 37 ]. So far, there is still a lack of identification regarding the ideal structure of the multidisciplinary team and care plan, despite the proposal of criteria for center of excellence. Moreover, even with gradually accumulated experience on PAS management, quality improvements depend largely on reflective changes in care plan by regular team audit, especially for patients with increta and percreta after cesarean hysterectomy [ 37 ]. However, further assessment is necessitated to clarify how these elements of quality improvement affect the outcomes of patients mainly receiving uterine preservation. Principally, patient blood management can be developed from the aspects of preoperative hemoglobin optimization and autologous blood preparation utilization, which may enable the minimization of potential complications related to allogenic transfusion [ 38 , 39 ]. Nearly 38% of pregnancies were estimated to be complicated by iron deficiency and/or anemia, demonstrating increased risk of postpartum hemorrhage [ 40 , 41 ], while it could be corrected through oral and parenteral iron supplementation [ 42 ]. Furthermore, anemia/iron deficiency was independent risk factor of blood transfusion in PAS patients [ 43 , 44 ]. Nevertheless, the US Preventive Services Task Force reviewed and declared limited benefit of routine iron supplementation and no studies addressed screening for iron deficiency and iron deficiency anemia, which, however, could be improved by prenatal iron supplementation during pregnancy [ 45 ]. In the present study, a program of iron supplementation optimization was implemented by our multidisciplinary team, allowing earlier screening around 20 weeks of gestational age. Consequently, it was demonstrated with elevated preoperative hemoglobin levels and reduced anemia during pregnancy, facilitating subsequent autologous blood preparation. In accordance with the guideline issued by British Committee for Standards of in Haematology, we formulated corresponding criteria on preoperative autologous blood donation [ 29 ]. Our previous data on high-risk patients has documented the feasibility and safety of preoperative autologous blood donation during delivery, with median blood loss of 677.5 ml (255–8400 ml) followed by 17.9% allogenic blood transfusion, similar to the rate of 16.7% in placenta previa [ 38 , 46 ]. But as reported by Lili Steblovnik, only 13.6% of patients could benefit from preoperative blood donation [ 38 ]. As a result, there was a gradual reduction of preoperative blood donation from 37.9% in initial period to 21.0% in current period. Furthermore, in terms of intraoperative cell salvage, which has been recommended in different guidelines, prior observational study has unveiled its role on minimizing allogenic blood transfusion [ 39 , 47 , 48 ]. Currently, it serves as a predominant plan in autologous blood preparation. In our study, acute normovolemic hemodilution was primarily regarded as a supplementary protocol in autologous blood preparation, sharing similar role in previous case reports on Jehovah’s Witness [ 49 , 50 ]. However, it was quickly abandoned given its risk of decreasing oxygen delivery to the fetus. Through flexible strategy on autologous blood preparation, our study reported an overall rate of autologous blood transfusion of reaching 53.8%. However, significant reduction of allogenic transfusion was not expected as previous studies, despite the observation of a decreasing trend and less massive blood transfusion [ 39 , 51 ]. It is believed that preoperative balloon occlusion may contribute to mitigating intraoperative bleeding in cesarean hysterectomy [ 52 – 54 ]. But some studies failed to demonstrate any benefit and criticize their use [ 55 , 56 ]. Simultaneously, given its potential complications, the application was gradually replaced by surgical improvements and other blood conservation techniques. The implementation of uterine incision without penetrating placenta combined with the dissection of vesicouterine peritoneum to put down bladder provided the surgical foundation for uterine preservation. The previous one has been applied as the first step in Triple-P procedure, as well as one-step conservative surgery for PAS disorders, facilitating the delivery of fetus without interrupting placenta [ 57 , 58 ]. Moreover, the dissection of vesicouterine peritoneum to ligate newly formed vessels and putting down bladder could provide convenience for the explosion of lower segment of uterus [ 58 ]. The adoption of such technique could allow the utilization of eight-figure suture to reconstruct the lower segment of uterus, an area frequently implicated in abnormal placental implantation during transition period. In junction with ligation of the ascending branch of uterine artery, the application of preoperative balloon occlusion and uterine balloon tamponade were reduced from 81.7% to 71.6% in initial period to 32.2% and 52.0% in transition period, respectively. Further modification of this surgery was realized through the resection of myometrium where placenta was deeply attached and manual removal of the remaining placenta. On the basis of thorough hemostasis of the surgical field, it was observed with further reduction of preoperative balloon occlusion and uterine balloon tamponade utilization. However, this study revealed an augmentation of 400 ml blood loss, which contradicted other research on quality improvements on PAS [ 37 , 59 ]. It was speculated that the elevated blood loss may be explained by the severity of PAS and less invasive hemostatic procedures (e.g., preoperative balloon occlusion and intrauterine tamponade). Furthermore, one randomized controlled trial reported slightly greater blood loss in the conservative surgery group compared to the cesarean hysterectomy group, which may account for the shifting trend in blood loss observed during our study [ 60 ]. However, this phenomenon did not lead to additional consumption of allogenic blood products, but with decreased early composited morbidity and adverse complications/events. Despite a relatively large sample size, this study still has several limitations, such as the retrospective design and the inability to quantify the effect of each quality improvement element. It should be acknowledged that the implementation of quality improvement element for PAS disorders is actually not instantaneous. It requires time for adaption, especially considering the learning curve of surgical skills. In response to this issue, our study established a transition period, including the introduction of most of quality improvement elements, which filled the gaps of different outcomes from initial to current period, compare with other two-phase studies [ 37 , 59 ]. In addition, part of patients undergoing surgery finally in our Center were referred from other institutes at different gestational weeks, which might produce more significant impact on our prenatal care plan and findings than we can measure. These included iron supplementation, preoperative autologous blood preparation, as well as schedule of surgery, especially in patients with vaginal bleeding and threatened preterm labor. Besides, the implementation of these quality improvement elements was completed in a single high-volume tertiary center with multidisciplinary teams or resources. However, this advantage may compromise the applicability of these measures in resource-limited settings. In the present study, our general method was to achieve uterine preservation, rather than cesarean hysterectomy or expectant management, exhibiting significant difference from other Centers. The rationale of our general attempt relied on a Canadian study, which proposed that due to sepsis, severe vaginal bleeding and coagulopathy, 40% of hysterectomy were implemented after expectant management [ 61 ]. Besides, cesarean hysterectomy might not be preferred given the risk of fertility loss and future mental health [ 62 , 63 ]. Anyway, with respect to the above, there is a need for multi-center cohort study to further evaluate the impact of quality improvement on maternal outcomes receive uterine preservation.

Supplementary Material 1. Supplementary Material 1. Supplementary Material 2. Supplementary Material 2.