In this study, RPOC with bleeding were treated by UAE using GS with high technical and clinical success rates. In one technical failure case, we couldn’t embolize right UA because of a direct shunt between the right UA and the right internal iliac vein. However clinical success was obtained; the bleeding source from RPOC was considered as left UA and embolized successfully. The arteriovenous shunt itself might not be a source of bleeding because uterine arteriovenous shunt can be asymptomatic and result in spontaneous lesion resolution (Degani et al., 2009).
In recent case series of UAE for RPOC using mainly microspheres (size: 700-1200 μm) (Bazeries et al., 2017), 25.8% of patients required additional treatment for re-bleeding. Microspheres may be unsuitable for bleeding from RPOC with markedly dilated spiral arteries in the inter-villous space of the placental tissue (Roberts, 2014), which may be up to 2 mm in diameter (Espinoza et al., 2006). We experienced ovarian supply in one case, which was managed without embolization. The ovarian artery supplied only small part of RPOC on angiography and both uterine artery embolization was enough to stop bleeding caused by RPOC. However, ovarian arteries might be one of the causes of re-bleeding: ovarian artery supply was confirmed in 12% of PPH patients (Kim et al., 2018). Therefore, it is suggested that we should visualize both ovarian artery on initial aortography at the level of the renal arteries and selective embolization might be needed if its contribution to bleeding is confirmed (Kim et al., 2018; Maassen et al., 2009). Pseudoaneurysm was more frequently seen on angiography in patients after D&C/E in this study. This suggests that the traumatic procedure of D&C/E can cause pseudoaneurysm due to arterial rupture of a hypervascular RPOC. Thus, vascularity should be evaluated with color Doppler before D&C/E for RPOC (Kitahara et al., 2011). Also, dynamic gadolinium-enhanced MRI may support the US findings regarding the vascularity of the mass to avoid bleeding complication related to gynecologic interventions (Dohke et al., 2000).
A standard of care for RPOC is usually careful observation because RPOC often disappear spontaneously without treatment (Jain & Fogata, 2007; Lee et al., 2014). When RPOC remains symptomatic, gynecologic options such as D&C/E and HR (Ben-Ami et al., 2014; Golan et al., 2011) can be one of options. However, it might have the possibility of life-threatening hemorrhage because 18% of RPOC have marked vascularity on doppler US (Kamaya et al., 2009). There is no reliable means to predict future spontaneous massive bleeding (Kitahara et al., 2011), therefore conservative management may be preferable to emergency UAE, if the patient’s condition allows. As not-embolization methods to manage RPOC, ‘maximum laminaria procedure’ with surgical RPOC removal was recently reported by Usui et al. (Usui et al., 2018). They performed overnight insertion of two laminaria tents into cervical canal of patients with expect of the compression of uterine artery. This might be one of choices to treat RPOC with non-urgent bleeding, but UAE might be preferred for cases of massive bleeding.
Uterine arteriovenous malformation (AVM) is a differential diagnosis of RPOC, although rare with 0.1% of after abortion or delivery to 4.5% of patients with abnormal vaginal bleeding (Yazawa et al., 2013; O’Brien et al., 2006). The torturous abnormal feeding arteries and early venous filling seen in RPOC in this study are similar angiographic findings to those of uterine AVM (Timmerman et al., 2003; Ghai et al., 2003). To discriminate congenital uterine AVM from RPOC, US finding of mass-like lesion in the endometrium or thickened endometrial echo complex is a key feature of RPOC (Sellmyer et al., 2013), but strict differentiation may be difficult, because RPOC and AVM can have overlapping and can coexist (Iraha et al., 2018). Management of symptomatic uterine AVM resembles RPOC with bleeding: surgery or UAE. UAE provides a safe and effective alternative to surgery, with 88–100% success rate using coils, GS and n-butyl-2-cyanoacrylate (Badawy et al., 2001; Wang et al., 2012; Picel et al., 2016).
In this study, menstruation after UAE was observed in nine out of 11 patients (82%) whose follow up period was longer than 1 month. Pregnancy was observed in six of the nine (67%) patients with menstruation after UAE. A recent study showed that 92.5% of patients resumed menstruation after UAE (Gaia et al., 2009). The difference in our study might be because these two patients experienced normal delivery although follow-up period was only two and 3 months. After UAE using GS, 62% of patients were reported to achieve pregnancy (Gaia et al., 2009), almost the same rate as UAE using NBCA; 60% pregnancy rate (Picel et al., 2016). UAE might have possibility of pregnancy, but future research should address unknown effect for uterine function and complication in long period of time.
There are some limitations in our study. First, this is a retrospective study with a small number of patients. RPOC with massive bleeding are rare disease, and a multicenter registry may be warranted to assess the embolization techniques and outcomes. Second, histopathological confirmation of RPOC was not undertaken except in one case. In addition, despite RPOC spontaneously detached or diminished after UAE, further examination (hysteroscopy or follow-up CE-MRI) might be needed because remnant tissues degeneration has potentials for future malignancy. Yet it might be addressed that RPOC is one of a main cause of late PPH and may be managed by UAE using GS which might have possibility of fertility preservation.