Combined laparoscopy and hysteroscopy vs. uterine curettage in the uterine artery embolization-based management of cesarean scar pregnancy: a retrospective cohort study

The CSP patients, who were diagnosed between March 1, 2005 and March 1, 2010 at First Maternity and Infant Health Hospital, Tongji University; Obstetrics and Gynecology Hospital, Fudan University and Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University; Shanghai, China, comprised the patient pool of this study. The research protocol was approved by the ethics committee of the three hospitals above. The inclusion criteria were (i) history of previous cesarean delivery before hospitalization, (ii) first-visit ultrasonography revealing an empty uterine and cervical canal and a myometrial defect at the caesarean scar site that was surrounded by a rich blood supply and in which a gestational sac was embedded and (iii) being immediately treated with intra-arterial MTX + UAE and followed by combined laparoscopy and hysteroscopy or curettage. The exclusion criteria were (i) receiving MTX treatment or curettage before hospitalization and/or presenting a massive uterine hemorrhage and (ii) being treated with conservative or other surgical measures, eg. intramuscular MTX, curettage, hysteroscopy, etc., as the first-line therapeutic methods. A total of 58 patients were therefore included and then divided into two cohorts (groups), i.e., 25 patients in the study group, who received intra-arterial MTX + UAE + combined laparoscopy and hysteroscopy, and 33 patients in the control group, who received intra-arterial MTX + UAE + ultrasound-guided curettage. The medical records and follow-up information of these patients were carefully and thoroughly reviewed.

The initial treatment for patients in both groups was the same; the right femoral artery was punctured, and 5.0-F Cobra catheters (Cordis, Brentford, Middlesex, UK) were inserted into the bilateral uterine arteries under angiographic guidance. A dose of MTX (50 mg/kg, Hengrui, Lianyungang, Jiangsu, China) was injected into each catheter, and blood flow was blocked by gelatin sponge microbeads of two sizes: 500–700 μm and 700–900 μm (Alicon, Hangzhou, Zhejiang, China). Patients were given one of the following treatments: curettage or combined laparoscopy and hysteroscopy. Both procedures were conducted by a same group of clinicians within 48 hours after UAE (we noted the hemostatic effect of UAE reached the maximum within 48 hours and declined thereafter in our previous clinical practice; data not published). (i) Uterine curettage: The internal cervical os was dilated; the conceptus was detached with a curette under ultrasound guidance and carefully pulled to the cervical canal, and then it was dragged out of the external cervical os with a vascular clamp. The curettage was terminated if the intraoperative bleeding reached 200–400 mL, and this condition was treated accordingly (e.g., vaginal gauze, uterine tamponade, intravenous injection of aminomethylbenzoic acid or reptilase). If conservative treatments failed and/or the cumulative bleeding reached 800 mL, wedge resection or hysterectomy was adopted. Laparotomy was immediately performed in cases of uterine perforation, and the uterus was either repaired or removed according to the severity of the lesion. (ii) Combined laparoscopic and hysteroscopic surgery: Laparoscopy was used for intraoperative surveillance and intra-peritoneal surgery. The vesico-uterine excavation was exposed by laparoscopy to visualize the uterine isthmus. Hysteroscopy was used to determine the exact location of the ectopic conceptus and to estimate the minimal distance from it to the uterine serosa. The conceptus was removed with a diathermy loop using the coagulating resection and tearing maneuver under the direct visualization of hysteroscopy. The local tissue weakness or defect was further evaluated based on the degree of translucency of the isthmus wall to hysteroscopic light. The defect was closed using 1–0 absorbable suture under laparoscopy (using an interrupted suturing method; generally, 1–3 stitches could be enough to close the defect). The quality of the tissue repair was evaluated by hysteroscopy. Patients in both groups received 3-day postoperative antibiotics, and the volume of vaginal bleeding and the severity of abdominal pain (mild: bearable without any sedatives; moderate: bearable with a common dose of paracetamol; severe: unbearable and requiring pethidine) were recorded. The serum β-hCG level was monitored every other day. Patients with <10 mL of daily vaginal bleeding (determined by weighting the sanitary napkins used every day) and a decrease in serum β-hCG over two consecutive tests were discharged. The main operative outcomes were divided into two types: success and failure. A successful treatment was defined as a complete/partial removal of the ectopic conceptus, cease of the abnormal vaginal bleeding and preservation of the uterus. A treatment failure was defined as one of the following conditions: ongoing growth of the ectopic conceptus, continuous vaginal bleeding and loss of the uterus.

All patients had been followed up on a regular basis for two years. The initial follow-up schedule was once per week from the first day after discharge. This schedule was adjusted to once every three months if the serum β-hCG level returned to normal as demonstrated by two consecutive β-hCG tests. The follow-up items that were assessed included abdominal pain, vaginal bleeding, menstrual status and serum β-hCG level, as well as the uterine ultrasonography. All patients were advised to use contraception for at least one year. Information regarding repeat CSP and placenta implantation of the patients who became pregnant during the follow-up period was obtained by medical record review.

The differences in categorical data between the two groups were compared using the two-sided χ² test or Fisher’s exact test, as appropriate. A two-sided Student’s t test was used to compare the differences in the β-hCG level, the size of the gestational sac and the depth of implantation between the control patients with massive bleeding (>800 mL) and the remaining patients in the control group. SPSS 12.0 software (IBM, Armonk, New York, USA) was used, and p < 0.05 was considered statistically significant.

The demographic and clinical characteristics of the patients in the study and control groups have been listed in Table 1; there were no significant differences found between the two groups.

The bilateral UAE were successfully performed in all the patients. In the study group, for the majority of the patients, laparoscopy observed an isthmus with normal appearance. Under hysteroscopy, a conceptus with a diameter of 1–6 cm (including the gestational sac, placenta, and villus), which protruded toward the uterine cavity, was found at the incision in 14 patients. In eight patients, an indentation, which contained placenta- or villous-like tissue (1–2 cm in diameter), was observed. In the remaining three patients, a conceptus (placenta, villus) with a diameter of 2–4 cm, which partially protruded into the internal cervical os and/or obstructed the cervical canal, was observed. No complications occurred during the surgeries, and the average volume of blood loss was 78.0 mL (range: 20–200 mL, Table 2). For all patients, the final hysteroscopic examination revealed a good incision repair.

In the control group, the conceptus was successfully removed by a single curettage in 26 patients. The average volume of intraoperative bleeding in this group was 258.5 mL (range: 20–1600 mL, Table 2). Seven patients underwent multiple curettages because the blood loss during the first curettage reached or exceeded 400 mL. Their average blood loss was 507.5 mL (range: 200–680 mL). Of them, one patient exhibited excessive intraoperative bleeding (>800 mL), which was ceased conservatively (e.g., uterine tamponade, hemostatic drugs) in 72 hours. This patient was transfused with 400 mL of blood, and her total blood loss was 1200 mL. Another patient exhibited persistent vaginal bleeding (100–200 mL/day, which increased over time) with severe abdominal pain for two days following the curettage. After all conservative methods failed, an exploratory laparotomy was performed and revealed a perforation at the uterine isthmus. Due to the severely damaged tissue, the attempt to repair the perforation was unsuccessful, and a hysterectomy was performed. The total blood loss of this patient was 1600 mL, and the volume of blood transfused was 800 mL. The preoperative serum β-hCG levels of these two patients were 8,200 IU/L and 32,000 IU/L, respectively, which had no significant difference from that of the other patients (t = -0.774, p = 0.442, two-sided Student’s t test) of this group. Additionally, no significant differences in the size of the gestational sac (12 mm and 57 mm for the two patients, respectively) or the implantation depth (distances to the serosa were 0.7 mm and 0.5 mm for the two patients, respectively) were observed between these two patients and the other patients (for gestational sac size, t = -0.831, p = 0.409; for implantation depth, t = 1.551, p = 0.127; two-sided Student’s t test).

The average surgical time was 48 minutes (range: 35–62 minutes) in the study group and 34.6 minutes (range: 15–130 minutes) in the control group, and no significant difference was found (Table 2). The average hospital stay was 6.5 days (range: 5–9 days) in the study group and 8.7 days (range: 5–22 days) in the control group, and this difference was statistically significant (Table 2).

During the first three postoperative days, the daily volume of vaginal bleeding among patients in the study group did not exceed 50 mL (average: 21.9 mL/day; range: 0–50 mL/day). However, among patients in the control group it was 10–120 mL (average: 28.1 mL/day). Both the volume and duration of vaginal bleeding were significantly lower in the study group than in the control group (Table 3).

Mild to moderate abdominal pains were reported by all patients. The average duration of postoperative pain was similar between the two groups (Table 3). During the follow-up period, one patient in the study group and eight patients in the control group reported a recurrent dull and/or occasionally intensified pain. The difference in the rate of occurrence of such chronic pain was significant between the groups (Table 3).

Patients in both groups (except one case of hysterectomy) resumed menstruation within one month of surgery (Table 3). However, more patients in the control group reported irregular (reduced or increased) menstrual flows, elongated menstrual periods or abnormal inter-menstrual bleeding than those in the study group (Table 3).

The average serum β-hCG regression time was significantly shorter in the study group (19.6 days, range: 15–23 days) than in the control group (29.3 days, range: 17–46 days, Table 3).

A postoperative ultrasonographic evaluation was performed for all patients one month after the surgery. High-density light spots, which suggest the existence of a residual conceptus, were found in two patients in the study group and seven patients in the control group (Table 3). Expectant management was recommended to eight patients because their β-hCG levels were normal, and no signs of bleeding were found. For one patient in the control group, whose serum β-hCG level was elevated, a course of intramuscular injections of 50 mg MTX (first day) and 5 mg folic acid (second day) were applied for four cycles. Her serum β-hCG level declined to normal after one month, but at that time a small mass of conceptus could still be detected by ultrasonography.

Regular uterine ultrasonography was performed for each patient. All of the residual conceptuses disappeared by the end of the follow-up. However, unhealed incision defects were observed in a few patients in both groups. Specifically, in one patient from the study group, a small submucosal dehiscence (2 × 7 × 10 mm³) was found. In two patients from the control group, tissue fissures (or diverticula), which reached the uterine serosa, were detected. Additionally, incision dehiscence or indentations at different degrees were found in 22 patients from the control group. Compared to the study group, the rate of incision defects was significantly higher in the control group (Table 3).

After CSP, few patients wanted to become pregnant again. Only one patient in the study group desired pregnancy. However, this patient ultimately failed to conceive. Three patients from the study group and two from the control group reported induced abortion due to unplanned pregnancies. Their treatments all went smoothly, and no cases of repeat CSP were encountered (Table 3).