A total of 925 pregnant women with singleton and placental accreta spectrum and who were subjected to cesarean section at the first affiliated Hospital of Zhengzhou University from January 2018 to June 2020 were involved in this study. There were 336 cases of placental percreta. A total of 142 pregnant women were subjected to PAMRR, while 194 pregnant women were treated with other suture methods (control group). These suture methods included figure of eight suture, coarctation suture, and simple continuous suture among others.
Before surgery, we performed ultrasonic and Magnetic Resonance Imaging (MRI) examination to all PAS suspected patients. Data about the placenta, including its localization, thickness, loss of hypoechoic retroplacental zone and numerous large and irregular lacunae in the placenta, abnormal uterine bulging, dark intraplacental bands on T2-weighted images, heterogeneous signal intensity and disorganized placental vasculature were recorded. Inclusion criteria: All the patients were confirmed by ultrasound and MRI as well as clinical diagnosis. i. PAS score under ultrasound ≥ 10. Placental accreta score refers to the scoring system designed by Peking University Third Hospital  (Figure 1); ii. Placenta percreta confirmed by MRI  (Figure 2); iii.Surgical manifestations (FIGO grade 3a, 3b and 3C) , such as invasion of placental villi into the uterine serosa, urinary bladder or broad ligament, vaginal wall, pelvic sidewall or any other pelvic organ. Pregnant women who met all the above criteria were diagnosed with placental percreta. Cases with large areas of placental percreta were included in this study. The exclusion criteria were: i. PAS score under ultrasound <10; ii. Patients with severe obstetric complications and serious internal and surgical diseases, including cardiac disease, liver disease or pre-eclampsia; iii. Patients with twins or multiple pregnancy. All surgical procedures were performed by the same surgeon, Xianlan Zhao, with several assistants that remained unchanged. The whole process was managed by our multidisciplinary team, including doctors who were experts in ultrasound, MRI, anesthesiology, invasive technology and obstetrics. At 42 days after surgery, patients were returned to the outpatient clinic of the hospital. Clinical follow-ups were made by phone calls at three months, six months, one year and two years after surgery.
Based on clinical evidence, individual schedules were made for every patient. The schedules considered gestational week to pregnancy termination, the amount of prepared blood, and the necessity for placing a prophylactic abdominal aorta balloon among others. For the following three cases, the abdominal aorta balloon was preset before surgery: i. Patients with more than one previous cesarean section; ii. Severe abdominal adhesion in previous cesarean section predicted the difficulty of tourniquet placement during this operation; iii. Patients with cervical involvement as indicated by ultrasound or MRI. Surgical procedures were performed in the Digital subtraction angiography (DSA) room. Before cesarean section, the interventional doctor inserted a 5F balloon catheter into the distal abdominal aorta beneath the opening of the renal arteries through the right femoral artery. After confirming the good position of the balloon, cesarean section was performed, Figure 3.
i. The primary incision was recut, and the scar was removed from all patients that had been subjected to at least one C-Section before. Majority of the cases were transverse incisions of the lower abdomen while minority of the cases were longitudinal incisions. To expose the upper border of the placenta, our incision was about 3 cm longer than the previous one.
ii. Uterine incision was performed transversely over the upper border of the placenta, which is obvious after opening the abdomen. As shown in Figure 4.A, there were engorged and tortuous vessels on the serosa surface corresponding to the area of the placental increta beneath. To minimize bleeding, the incision was pulled tightly by the surgeon and the first assistant to maintain some tension.
iii. After delivering the fetus, the uterus was immediately exteriorized by the first assistant whose left hand pulled the anterior wall and the right hand dragged the fundus in the abdominal cavity. Then, we checked whether there was adhesion between the uterus and abdominal wall. The adhesion should be separated before opening the uterus.
iv. The aortic balloon that had been placed below the kidney artery before cesarean section was inflated by simultaneous delivery to occlude blood flow to the uterus. Meanwhile, we put a tourniquet as low as possible at the lower segment of the uterus to further block blood flow to the uterus. It was challenging to establish a tourniquet on patients with serious adhesions in which the aortic balloon was crucial and irreplaceable. Both treatments significantly reduced the amount of bleeding.
v. We pushed down the bladder and partially resected the anterior myometrium with the placenta unseparated (Figure 4.B). All of our patients had placenta percreta covering a large area. The bladder was pushed down to a horizontal line of internal cervix os or to the accreta area if the posterior wall of the bladder muscle layer was invaded. The anterior myometrial from the lower lip of the uterine incision to the area 1 cm above the cervix was resected using a scalpel or scissor together with the bulk of implanted placenta. Due to the prophylactic abdominal aorta balloon and/or tourniquet, bleeding hardly ever occurred during this process. At the same time, the interventional doctor injected 1.2 ml of saline into the balloon to block the abdominal aorta.
vi. A single continuous suture was made along the lower lip of the uterine incision, as shown in Figure 4.C. Since the new-formed lower lip was one part of placenta in the deeply implanted area that was full of tortuous vessels, which can cause massive bleeding in a short time, this single continuous suture significantly minimized the bleeding. Other hemostatic methods were performed on the placental bed to prevent bleeding, including the area around the internal cervix os and the area where the placenta could not be excised. The figure-of-eight suture was the first choice for us. In the absence of active bleeding points, the tourniquet was removed. After tourniquet removal, the new bleeding site was sutured again. Based on hemostasis, the saline in the balloon was gradually extracted until the balloon was completely deflated.
vii. The lower transverse incision at the uterine was sutured in one layer in the common way, or two layers if necessary. As shown in Figure 4.D, this area was still permeated with tortuous vessels. In cases in which suturing the second layer was considered dangerous, we preferred monolayer suturing.
viii. The balloon catheter was removed after surgery. When necessary, uterine arterial ligation was performed on patients with consistent vaginal bleeding.
Demographic characteristics including age, gravidity, time of previous c-sections, gestational age and placenta accreta score were recorded. We also recorded the perioperative information of patients, including estimated blood loss (EBL), operative time, blood transfusion type and volume, neonatal weight, postoperative hospital stay period, costs and whether they had uterine arterial embolism (UAE) and ICU transfer. Autologous blood transfusion technology was performed for all study participants based on their wishes.
Short-term prognosis was evaluated. Short-term follow-ups for all patients at the clinic and in the outpatient department were performed 42 days after the operation. Short-term complications included bladder injury, fever (temperature ＞37.3 ℃), late puerperal hemorrhage, pyometra, peritonitis, uterine necrosis, sepsis and perioperative infection, hematoma around the puncture location, false aneurysm, arteriovenous fistula and thrombus. Patients with an abdominal aorta arterial balloon were subjected to an echo examination after surgery to determine whether there was thrombus in the vessels (arteries, from the common iliac artery to the dorsalis pedis artery, and veins from the dorsalis pedis vein to the common iliac vein) of the pelvis and extremities.
Long-term follow ups were done by phone. Long-term follow-up was performed at three months, six months, one year, and at two years after surgery. Breast feeding situation, menstrual quantity, menstrual period, chronic pelvic pain, intrauterine adhesion and lower extremity discomfort were assessed. There was one pregnant case after PAMRR.
The SPSS 25.0 software was used for data analysis. Age, number of pregnancies, previous cesarean section times, gestational weeks, placenta accreta score, cost, operation time, EBL, CRBC transfusion, autologous blood transfusion, postoperative hospital stay and neonatal weights were measurement data. Age, cost and neonatal weights were normally distributed and were presented by mean ± standard deviation (x
± s). The independent sample t-test was used for between group comparisons. Data for the other indicators did not conform to normal distribution, and were presented by median (25th -75th percentile) [M (p25-p75)]. The rank sum test was used for inter group comparisons. The remaining statistical indicators were count data, expressed by frequency or percentage, and the chi square test was used for inter group comparisons. p≤0.05 was set as the threshold for statistical significance.