Patient Selection
Our hospital is the tertiary referral centre providing a range of subspecialty services for more than 60 hospitals in China. From January 2012 to September 2018, 481 patients with diagnosed CSP underwent UAE followed by hysteroscopic curettage in our hospital. In 24 of 481 (5.0%, 24/481) patients, non-gonadal collaterals supplying GS were identified by arterial angiography performed immediately after UAE. These 24 patients underwent attempt collateral embolization in addition to UAE, and were included for retrospective analysis.
This study complies with current ethical consideration and was approved by our institutional ethical committee. The combined procedure of UAE and additional embolization of non-gonadal collateral supplying GS followed by hysteroscopic curettage was explained in full to each patient, and each patient gave an informed consent.
The diagnostic criteria for CSP included an increased serum β-human chorionic gonadotropin (β-hCG) level (normal range less than 5 mIU/mL), history of a prior cesarean scar, and ultrasound image that met the criteria proposed by Godin et al [7], as follows: (i) absence of GS in uterine cavity and cervical canal; (ii) presence of a GS located at the anterior isthmus of uterus with or without cardiac activity; (iii) thinning or absence of myometrium between bladder and the GS (Fig. 1).
All the patients were hospitalized in the department of gynecology. The baseline characteristics of the patients including maternal age, time interval since last cesarean section, gestational age, GS diameter, myometrium thickness (the thickness of the myometrium between the GS and the bladder), and serum β-hCG level were recorded.
Arterial embolization and hysteroscopic curettage
The patients underwent UAE in our interventional department by two experienced interventional radiologists. Using Seldinger technique, the right common femoral artery access was achieved under local anesthesia. A 5-Fr Cobra catheter (Cordis, USA) was advanced distally to selectively catheterize the left internal iliac artery (IIA) negotiating with a hydrophilic guide-wire. The subsequent digital subtraction angiography (DSA) was performed at the IIA using 16 mL of nonionic contrast medium (Ultravist; Bayer, Germany) at a flow rate of 4 mL/s.
The DSA performed at the IIA was attempt to visualize the vascular anatomy of uterine artery, uterine perfusion, and the hypervascularity enhancement in GS region. When necessary, DSA of the IIA was performed with ipsilateral anterior oblique projection (30°) to reduce the imaging superimpositions and detect the origin of uterine artery. Then, a coaxial microcatheter (Stride 2.6 F, Asahi Intecc or Progreat 2.7 F, Terumo ) was introduced coaxially into the uterine artery and advanced distally beyond the origin of the cervicovaginal branch with the digital roadmapping of the IIA. Selective uterine artery angiography was performed to ensure the trajectory of the uterine artery, uterine perfusion, and the hypervascularity enhancement in GS region. The subsequent embolization was performed by injecting the embolic agent slowly under continuous fluoroscopic guidance without reflux to undesired arteries. The embolic agents were 1000–1400μm sized gelatin sponge particles (Alicon Co. Ltd., Hangzhou, China), which were absorbable agents. Each vial of gelatin sponge particles (1 mL) was mixed with 50 mL solution of contrast medium and saline solution at a 1:1 ratio. Embolization of the uterine artery was performed to the point of complete stasis of blood flow achieved in the ascending uterine artery.
Then postembolization angiongraphy at IIA after the removal of microcatheter was performed immediately to confirm the arterial occlusion of uterine artery and identify the possible presense of collateral supplying GS. The artery with the angiographic appearance of neovascularity and corresponding enhancement in the GS region was interpreted as a collateral supplying GS. Once a collateral supplying GS was identified, further selective microcatheterization and angiography was performed in the collateral to ensure the corresponding enhancement, followed by embolization with aforementioned gelatin sponge particles to the point of complete stasis of blood flow. After the embolization of the left uterine artery and collateral was finished, a Waltman loop was created on the Cobra catheter and the embolization of the right uterine artery and collateral was performed in the same way (Fig. 2). Analgesics and antiemetics were administered when needed.
After the embolization, the patient received 100 mg of mifepristone per day for cervical preparation. The following removal of the GS using hysteroscopic curettage technique was conducted by experienced gynecologists within 48 hours post-embolization. Under epidural anaesthesia, an operative hysteroscope with a 10-mm external diameter was placed inside the uterus following the cervical dilatation by Hegar dilators. Uterine distension was obtained using 5% glucose solution propelled by a uterine expansion instrument. Subsequently, an operative 26F hysteroscopic resectoscope with an electric wire loop electrode was introduced under transabdominal sonography guidance. The GS was pushed with the wire loop to expose the blood vessel bed of the implantation site. Then, the GS was removed using placenta forceps under direct vision. The following suction curettage with the wire loop was perfomed to clear the residual gestational tissue from the uterine wall. Coagulating the blood vessels with hysteroscopic rolling ball was used to control bleeding if required. This process continued until the gestational tissue was cleared completely and the myometrium was visualized (Fig. 3).
According to the protocols, ultrasonography was performed 3 days after the procedure, and serum β-hCG level was assessed before the procedure and every 3 days after procedure until discharge. Blood count was monitored before the procedure and 24 hours after procedure. Patients were discharged when (i) vaginal bleeding ceased or was mild; (ii) steadily decreasing in serum β-hCG levels was observed; and (iii) pelvic pain was absent. Follow-ups after discharge included repeat ultrasonic examination and serum β-hCG level measurement that were performed every 2 weeks until the serum β-hCG level returned to normal, and no residual GS was detected.
Outcome evaluation
The items evaluations were as follows: anatomic characteristics of the collateral supplying GS, technical success of collateral embolization, and clinical outcomes including time for β-hCG levels normalization and side effects.
Anatomic characteristics of the collateral supplying GS consisting of the origin of each collateral and whether unilateral or bilateral collaterals identified in each patient were recorded. Technical success of collateral embolization was considered when selective embolization in all identified collaterals supplying GS, and complete disappearance of uterine perfusion and hypervascularity enhancement in the GS region on post-embolization angiography at IIA were achieved. Side effects included post-embolization pain, the amount of intraoperative blood loss during curettage, secondary anemia, and whether need for additional blood transfusion or further hysterectomy or laparotomy.
In order to assess the role of the additional collateral embolization, the included patients were divided into two groups based on whether they underwent technically successful collateral embolization (UAE-SCE group) or failed collateral embolization (UAE-FCE group) in addition to UAE. The baseline characteristics and clinical outcomes were compared between the two groups.
Statistics
The paired t test and Man Whitney test were used for comparisons of discrete and numerical variables, respectively. All statistics were determined using SPSS (SPSS, Chicago, USA) version 22.0 and P value of less than 0.05 was considered to be statistically significant.