Elective uterine artery embolization prior to laparoscopic resection of interstitial pregnancy: two cases and literature review

Ectopic pregnancy (EP) is any type of pregnancy in which the fertilized ovum implants outside the uterine cavity. The vast majority of EPs are situated in the fallopian tube, typically in the ampullary region (70%), less likely in the isthmic (12%), fimbrial (11%), or interstitial part (2–4%). Other uncommon locations include ovarian (1–3%), abdominal (< 1%), cervical (< 1%), rudimentary horn (< 0.5%), and cesarean scar pregnancies (1–3%) [1‐4].

In 1989, EPs occurred at an estimated prevalence of 1–2% worldwide. This is two to three times higher than in 1970 [5]. The increase is presumably related to an increased prevalence of risk factors directly or indirectly leading to decreased tubal passage. The prevalence has since not significantly changed [6, 7].

Pregnancies that are situated in the interstitial portion of the fallopian tube are referred to as interstitial [8, 9]. The intramural or interstitial part of the tube is approximately 0.7 mm wide and 1–2 cm long, often with a tortuous course [8]. Interstitial pregnancies (IPs) are also referred to as “cornual,” though some reserve this entity to pregnancies located within a rudimentary horn of an abnormal uterine cavity [8, 9]. While the generic risk factors displayed in Table 1 may also apply, specific risk factors to this type of EP are previous ipsilateral or bilateral salpingectomy, previous EP, in vitro fertilization, and tubal damage from previous EP [8]. Historically, the mortality rate of this condition was around 2.5%, which is approximately seven times higher than that of EPs in general. It is assumed that this can be explained by the greater expansion capacity at this location, the richer vascularization of the area, eventually leading to life-threatening hemorrhage when rupture occurs [8].

There is to our knowledge no consensus on the best treatment modality of IP. Herein, we provide a literature review which we did on the occasion of treating two patients with uterine artery embolization (UAE) immediately prior surgical treatment, because of an anticipated high risk for bleeding.

A 28-year-old gravida 5 para 1 was referred for a second opinion on an evolutive IP. She had a history of a primary cesarean section for vasa previa, a spontaneous first trimester miscarriage, two EPs treated by salpingectomy, and a hysteroscopic cesarean scar niche repair. The latter niche repair was done because of ultrasound signs of fluid in the niche before starting in vitro fertilization (IVF) treatment. On hysteroscopy, blood and debris were confirmed and a repair was performed 4 months prior to the index event (IP). Control hysteroscopy 1 month after the procedure showed normal findings. The index pregnancy was by IVF. On early scan at 6 + 6 weeks, an IP was suspected. We confirmed this at 7 + 1 weeks to be a left IP with a gestational sac of 19 × 20 mm, CRL of 6.8 mm, β-hCG of 38,000 IU/L, and heart activity. There was no abdominal fluid. The referring center opted for a single-dose methotrexate (MTX) protocol (75 mg; 50 mg/m²). She presented on day one post-injection with stinging and cramping abdominal pain, yet without hemodynamic impact or peritoneal signs. On day six post-injection, she was referred because of raising β-hCG and persisting heart activity, spotting, along with intermittent abdominal pain. Figure 1 displays the ultrasound, β-hCG, and hemoglobin findings over the reporting period. We decided to proceed with surgical intervention yet opted for prior bilateral UAE during the same general anesthesia to reduce the risk for hemorrhage based on the apparent high vascularization around the pregnancy. Access was gained through the right femoral artery with catheterization of the left internal iliac artery followed by selective catheterization of the left uterine artery. Polyvinyl particles (Contour 250-350, Boston Scientific, Diegem, Belgium) were injected under 3D angiography control. The same procedure was followed on the contralateral side. Then, a laparoscopic cornual resection was performed and the uterine defect was closed in two layers using Vicryl 2-0 (Fig. 2). Blood loss was negligible, yet operating time was 140 min. Histopathology confirmed an IP. She was discharged on day two, and β-hCG became unmeasurable 4 weeks later. She had a withdrawal bleeding 3 weeks after the operation and had another period 5 weeks later. A waiting period of at least 6 months [10, 11] was advised to allow maximal healing of the uterus. She conceived 8 months after the IP in the first IVF cycle. She presented again with right fossa pain at 5 + 3 weeks, yet ultrasound confirmed an intracavitary position without any signs of IP.

A 32-year-old gravida 5 para 3 spontaneously conceived. She was referred because on elective dating ultrasound at 9 + 2 weeks a right evolutive IP was found. She had a history of a spontaneous first trimester miscarriage and three uncomplicated term vaginal deliveries. On ultrasound, the surrounding myometrium was 2.2 mm which was strongly vascularized (Fig. 3). Because of the size (CRL = 24.5 mm), the thin myometrial layer, and a β-hCG of 121,758 IU/L, we advocated immediate surgery, yet because of the vascularization we first offered bilateral UAE. Polyvinyl particles (Contour 355-500, Boston Scientific; Embosphere 500-700 and 700-900, Merit Medical, Brussels, Belgium) and spongostan plugs (Ethicon, Diegem, Belgium) were used (Fig. 4). On laparoscopy, a 6-cm pregnancy in the right uterine horn was observed. The pregnancy was removed by cornuostomy, and the myometrial defect was sutured in three layers (first V-loc 2-0, second and third Vicryl 2-0). Blood loss was negligible, and operating time was 180 min. Two months later she still had some brown vaginal discharge. Ultrasound showed normal findings with a strong proliferative endometrium along with a corpus luteum on the left ovary and a normal looking scar at the resection site. β-hCG was 3.2 IU/L.

Both patients explicitly consented to have their history being reported in the literature.

For the literature review, we searched the PubMed on this matter, published until February 2018, using the following key terms “Pregnancy, Interstitial”[Mesh], “Therapeutics”[Mesh], “Interstitial Pregnancy,” and “Pregnancy Treatment” (953 papers). Sources of relevant articles in the references were screened as well (> 100 papers). All English-, French-, Dutch- and German-language articles were retrieved and screened on title and abstract for relevance (Appendix 1, 2, 3, 4, 5, and 6). Articles in which the location of the EP was unclear or in which the outcome was not clearly specified or objectively measured were excluded. We empirically decided to further discuss outcomes of series with 10 patients or more as to have reasonable denominators for calculating overall outcomes. The only exception to that was Table 4, which displays the entire published experience with UAE. There was not a single series with ≥ 10 patients treated with UAE.

There is considerable experience with primary systemic medical therapy in asymptomatic hemodynamically stable patients with IP. Table 2 summarizes studies describing ten or more patients with IP treated by primary systemic MTX. Dosing and regimen of MTX are inconsistent, and success rates are typically over 70%, except in one series [12]. In case of failure (persisting β-hCG leading to additional treatment), surgery was offered, except in one series by Hiersch et al., where second-line local MTX was combined with UAE. Out of five patients, two still required surgery as a third step. Tanaka et al. described 33 cases treated with a very consistent scheme of slowly intravenously injected, yet a fixed dose MTX. The success rate was 94%; two patients required surgery. The opposite was true in the experience of Kim et al. (n = 30) administering intramuscular MTX, yet with an inconsistent dosing regimen. Sixteen (53%) required additional surgery.

Local injection of MTX, potassium chloride (KCL), etoposide, and actinomycin D under laparoscopic, ultrasound, or hysteroscopic guidance have all been reported as effective (Table 3 and Appendix 2). These injections are usually given into the gestational sac, occasionally in the surrounding myometrium or locally intra-arterial. These are invasive procedures, compared to systemic MTX. Benifla et al. used MTX for IP locations and KCl for heterotopic presentations, out of concerns for teratogenicity. Of the three eutopic pregnancies associated to a heterotopic location, two were eventually lost. Further details on outcomes are missing. The calculated success rate was 88%. Treatment failures were not offered a second MTX injection, yet successfully managed by surgery.

Table 4 displays reports on patients managed with selective UAE combined with any administration regimen of MTX. The actual indication for secondary UAE was refusal of surgery (Ophir et al., Yang et al.; each n = 1) or not mentioned (Deruelle et al., Tamarit et al., Berretta et al., Hiersch et al.). Primary UAE combined with MTX was either part of a standard protocol (n = 9; Krissi et al.) or because of the suspicion of increased risk for hemorrhage (n = 1; Valsky et al.). The paper does however not mention how that increased risk was estimated. Table 4 also includes two cases managed by UAE followed immediately by planned surgery (either laparoscopic or hysteroscopic). The argument for UAE was made based on increased vascularization on 3D CT angiography. In one of those two cases, a subsequent spontaneous conception and cesarean delivery of a healthy baby at 37 weeks was reported. Overall success rate in all the series in this table is 91%.

Table 5 displays the experience with primary surgery, typically by minimally invasive access. Success rate was 94%; transfusion need was 9%. Primary laparotomy was performed for tubal rupture, in case of severe adhesions (Tulandi et al.) or because of surgeon’s preference (Hwang et al.). Conversions were because of significant hematoperitoneum or because of uncontrolled bleeding perioperatively (n = 7; 2%).

Today the diagnosis of EP is usually made by ultrasound. In high-risk patients or countries where access to early ultrasound is easy, the diagnosis can be made prior to the development of symptoms. This allows careful planning of management. We surgically managed two cases of IP, which both were initially asymptomatic. One had typical risk factors and the other one did not. One had prior MTX therapy, and the second one had a very high β-hCG level. Both the ultrasound examination raised the suspicion of a highly vascularized lesion. Therefore, we decided to perform primarily bilateral UAE and surgery in the same anesthesia. This is different than the cases managed in Table 4. Though it is impossible to prove that UAE reduces the risk for hemorrhage, it seems that our surgery in both cases was nearly bloodless. Treatment was apparently also effective given that β-hCG levels fell as expected.

When systematically searching the literature, a gap of knowledge is identified on the use of UAE or in a broader perspective, the management of IP. This is probably because of the rarity of the condition. The data around do neither allow a proper meta-analysis, so that we limited ourselves to summarize the findings in somewhat larger series for each management option. There is quite some experience with primary medical therapy in asymptomatic hemodynamically stable patients. In analogy to other ectopic locations [13], the variability of MTX administration protocols is wide, including systemic single shot (either promptly or slowly infused), repetitive doses, and local administration [13]. Medical therapy has also been combined with UAE, mostly successful, yet Hiersch et al. reports on two cases where second line local MTX treatment combined with UAE failed. In those, we would guess the patient would have had more benefit of surgery.

Our literature review learns that the most frequent complication of surgery is hemorrhage, either with or without transfusion. The overall transfusion rate in IP is not judgeable since no reference to that outcome was made in any of the medically treated cases. However, 9% of laparoscopically managed IPs required blood transfusion. Therefore, it seems logical to take measures to reduce that risk. Surgically, one can use prophylactic coagulation by electrosurgery or ligation of the feeding artery, yet this may compromise viability of the tissue. Alternatively, vasoconstrictors have been described to reduce blood loss and operating time, yet they may have their own side effects and have only been reported to be effective for IPs with an average β-hCG of 10,000–25,000 IU/L [8, 14, 15]. Conversely, these are very cheap agents.

Modern invasive radiologic techniques are becoming increasingly popular, and those services become more widely accessible even in a semi-acute setting. Embolization techniques have found their place in modern obstetrics and gynecology. The experience with uterine myomas is meanwhile very large, and subsequent conception seems to be possible and relatively safe [16]. Torre et al. described an insignificant change in fertility rate and ovarian reserve after UAE for uterine fibroids in women with no other infertility factors [16]. Krissi et al. reported on the subsequent fertility after MTX administration with UAE in the treatment IP. Out of five women who tried to conceive, four did so, and three delivered successfully. Disadvantages of UAE are the higher cost in comparison to vasopressin, the longer duration of anesthesia, the more complicated logistics, and the additional local morbidity (e.g. ischemic pain, Asherman syndrome) [17, 18].

We report on the use of elective UAE prior to laparoscopic resection of IP, which coincided with a nearly bloodless operation. A literature search shows a wide variety of treatment options, yet most cases seem to be following the typical approach to EP. The overall success rate of surgical treatment of IP is higher than that of medical treatment. When performing laparoscopy, good hemostatic techniques are recommended since the operation takes place in a strongly vascularized region [8, 14, 15]. Our experience with two cases of UAE is yet another approach. It seems safe and reliable and does not preclude future conception.