A 44-year-old woman with hepatitis C cirrhosis presented with a week of heavy vaginal bleeding. Her obstetric history was significant for three cesarean sections. Her gynecologist made an initial diagnosis of menometrorrhagia exacerbated by thrombocytopenia and coagulopathy. Computed tomography (CT) angiography revealed splenic vein thrombosis and engorged pelvic veins which arose as collaterals from the splenic vein (Fig. 1). Hysteroscopy could not identify a culprit lesion due to the rapidity of bleeding.
A transjugular intrahepatic portosystemic shunt (TIPS) was created and thrombectomy of the splenic vein was performed and the residual partially occlusive thrombus was then stented. Hepatopedal flow was then noted from splenic vein to portal vein and through the TIPS. Hysteroscopy showed persistently engorged varices. Venous embolization of the varices was performed with a combination of embolization coils and a vascular plug (Fig. 2). Recovery was uneventful, and she was followed for 2 years in our clinic without further vaginal bleeding.
Vaginal bleeding due to portal hypertensive collaterals is a rare presentation in patients with cirrhosis. The first case was reported in 1967, in a cirrhosis patient with a history of total abdominal hysterectomy. They theorized that the varices formed within fibrous adhesions present as a result of the patient's prior surgery, a phenomenon that might have been similar to our case. At surgery, their patient had a dilated right hypogastric vein that communicated with several varicosities in the vaginal vault. Treatment consisted of ligation and partial vaginectomy.
When present, pelvic varices are typically multiple, ipsilateral, and dilated to at least 4 mm in diameter.2 The rarity of parauterine or vaginal varices is accounted for by several factors. First, portal hypertensive collaterals typically drain to the external iliac veins rather than to the internal iliac veins, whereby pelvic veins are dilated. Second, both the uterus and vagina have extensive venous plexuses draining to the hypogastric veins, which are part of the systemic circulation, thus adequately decompressing high pressure pelvic blood flow. However, perturbation or removal of the uterine plexus via surgery or scarring may leave the vaginal plexus insufficient to decompress shunted blood flow back into the systemic circulation, resulting in vaginal bleeding.7 Indeed, in the seven previously reported cases of vaginal variceal bleeding, six had previously undergone hysterectomy and another had received radiation for cervical cancer (Table 1). Our patient's risk factor seems to have been adhesions due to multiple prior cesarean sections. Another pathophysiologic change in our case was the development of splenic venous thrombosis with a resultant increase in left-sided portal hypertension leading to the rupture of the vaginal varices.
Table 1. Reported Cases of Bleeding Vaginal Varices
Treatment of bleeding vaginal varices generally parallels that of nonendoscopically manageable gastrointestinal variceal bleeding (Table 1). TIPS is likely to remain the best current option for well-selected patients with this rare presentation.
In summary, we report here only the seventh case of vaginal bleeding from varices in the setting of portal hypertension. Clinicians should maintain a high index of suspicion for this entity whenever a patient with portal hypertension presents with vaginal bleeding, specifically if the patient has undergone prior uterine surgery.