Telephone: 859-323-4661; FAX: 859-257-3644
Images in Liver Transplantation
Vaginal varices with massive hemorrhage in a patient with nonalcoholic steatohepatitis and portal hypertension: Successful treatment with liver transplantation
Article first published online: 29 SEP 2008
Copyright © 2008 American Association for the Study of Liver Diseases
Volume 14, Issue 10, pages 1538–1540, October 2008
How to Cite
McHugh, P. P., Jeon, H., Gedaly, R., Johnston, T. D., DePriest, P. D. and Ranjan, D. (2008), Vaginal varices with massive hemorrhage in a patient with nonalcoholic steatohepatitis and portal hypertension: Successful treatment with liver transplantation. Liver Transpl, 14: 1538–1540. doi: 10.1002/lt.21506
- Issue published online: 29 SEP 2008
- Article first published online: 29 SEP 2008
- Manuscript Accepted: 9 MAR 2008
- Manuscript Received: 8 FEB 2008
We present a case of massive bleeding secondary to vaginal varices complicating portal hypertension in a patient awaiting orthotopic liver transplantation (OLT). The vagina is among the rarest of locations reported for portal hypertensive varices causing hemorrhage; in fact, our patient represents only the seventh such case since the first report by Kreek in 19671–5 and the only case in which the initial and definitive management was OLT.
A 58-year-old white female, gravida 3, para 3, was referred to our transplant center for liver transplant evaluation for nonalcoholic steatohepatitis. Her past medical history included diabetes mellitus type 2 and nephrolithiasis, and her surgical history was notable for a total abdominal hysterectomy performed 17 years earlier for endometriosis as well as open cholecystectomy, bilateral tubal ligation, and renal lithotripsy. On physical examination, she had mild ascites but was otherwise unremarkable. Upon completion of the transplant evaluation, she was activated on the transplant waiting list.
Approximately 1 year after she was listed, the patient underwent routine computed tomography imaging, which showed a small, cirrhotic liver without ascites, and patent vasculature. Three months later, she presented with a sudden onset of significant vaginal bleeding. Vaginal varices were identified by direct visualization, and after a 3-unit transfusion of packed red blood cells, she underwent emergent suture ligation of the varices with cessation of bleeding. However, bleeding recurred after an asymptomatic interval of 2 weeks. Shortly after her return to the hospital, the patient became hypotensive and was discovered to have a hematocrit of 22%. She was aggressively resuscitated and taken emergently to the operating room for examination under anesthesia. Actively bleeding varices were again seen at the upper aspect of the vaginal cuff; these were suture-ligated and packed, with cessation of bleeding. In all, she received 4 units of packed red blood cells, 1 unit of fresh frozen plasma, and a six-pack of platelets during this admission, with return of hemodynamic stability and no additional hemorrhage.
During this time, placement of a transjugular intrahepatic portosystemic shunt (TIPS) was considered, but in light of the patient's recent stability, this was not performed, and she was discharged to the care of her primary physician. Ten days later, the patient again developed hemodynamically significant vaginal hemorrhage. At this time, she was evaluated with reconsideration of TIPS placement. Computed tomography imaging showed a patent splenic vein but also the new finding of portal vein thrombosis, which was then confirmed by carbon dioxide angiogram; TIPS placement was therefore precluded. Pelvic varices were appreciated adjacent to the bladder and vagina (Fig. 1). Of note, no thrombosis of the iliac venous system on either side was observed. Abdominal and pelvic angiography was performed the next day to rule out an arteriovenous malformation. Although none was seen, thrombosis of the portal vein was again visualized, along with retrograde flow in the superior mesenteric vein and enlarged mesenteric venous collaterals draining into numerous pelvic varices. Hypertrophy of the inferior mesenteric vein was also noted, with probable retrograde flow (Figs. 2 and 3).
Fortunately, a liver allograft became available 2 days later, and the patient underwent successful OLT. During transplantation, a 5-cm organized thrombus was removed from the portal vein, with return of brisk portal flow. Portal vein continuity was able to be restored by end-to-end anastomosis between the donor and recipient portal veins. During the operation, the patient again developed significant vaginal hemorrhage, which necessitated gynecology consultation at the conclusion of the transplantation. The gynecologists performed additional suture ligation of the varices while the patient was still in the operating room. She had no additional episodes of vaginal hemorrhage during her uneventful postoperative recovery.
The pathophysiologic effects of cirrhosis can lead to serious complications in patients awaiting OLT. Portal hypertension, which frequently accompanies end-stage liver disease, can be caused by anything that compromises portal-systemic drainage. In a patient with cirrhosis, this is typically due to increased vascular resistance within the liver,6 although outflow and inflow impairments (such as portal vein thrombosis in our patient) can also be contributory.
The formation of venous collaterals to decompress the portal system occurs as a direct consequence of increased portal pressure, typically in one or more areas in which there is native communication between portal and systemic circulation (distal esophagus/proximal stomach, rectum, umbilicus, and retroperitoneum). Varices are essentially venous collaterals in these transition zones that have become prominently dilated. Varices are most commonly found in the esophagogastric region, and the presence of these varices poses a high risk of developing hemorrhage, up to 30% within 2 years.6 Numerous ectopic locations have been reported as well, including the small bowel, colon, and biliary tree.7
The anatomy of the vagina and uterus makes them unlikely locations to develop varices for two reasons. First, the uterus has an extensive venous plexus, which primarily drains into the uterine veins and later the hypogastric veins (part of systemic circulation). The vagina also has a venous plexus, which similarly drains into the hypogastric veins via bilateral vaginal veins. The plexuses are in communication with each other and with the vesical and hemorrhoidal plexuses. Second, the nearest and only native communication between these networks and portal drainage is the superior portion of the hemorrhoidal plexus. Although anorectal varices may occur in up to 44% of patients with cirrhosis,8 the combined span of the vaginal and uterine plexuses provides numerous venues for decompressing venous hypertension, without the consequence of varix formation.
We present a case of hemorrhagic vaginal varices complicating portal hypertension in a patient awaiting OLT, a problem that is rarely described in the literature. With one exception, all reported cases of vaginal variceal hemorrhage occurred in patients who had previously undergone hysterectomy; at 19 years, our patient had the longest interval between hysterectomy and onset of hemorrhage among these cases. In 2005, Orlando et al.5 described the 2 most recent cases and conducted a thorough review of the literature. They hypothesized that the loss of the uterine venous plexus creates the possibility of prominent venous congestion in the vaginal network, ultimately leading to variceal development.
Only one reported patient with bleeding vaginal varices had not undergone hysterectomy, although she had a history of cervical cancer treated with radiotherapy 17 years earlier.4 Because a well-known effect of radiation is decreased vascularity,9 focused radiation to the cervix may have had the same result as hysterectomy in terms of disrupting communication between the vaginal and uterine venous plexuses. It appears that in certain patients, the vaginal plexus alone may not be sufficient to decompress severe portal hypertension, and this can result in venous engorgement and development of vaginal varices with hemorrhagic potential.
The therapeutic approach to bleeding vaginal varices in a patient with cirrhosis does not differ greatly from the approach to varices in any other location. Initial management is usually directed toward the local control of bleeding using suture ligation, banding, or sclerotherapy, together with compression dressing and simultaneous correction of any underlying coagulopathy. Temporizing measures such as TIPS are also beneficial in reducing variceal pressure and possibly improving other complications of portal hypertension. Liver transplantation remains the definitive treatment for end-stage liver disease and severe portal hypertension resulting in bleeding varices and provides excellent outcome.
As we identify potential risk factors for developing this complication, it may be advisable to consider including a thorough gynecologic examination as part of the transplant evaluation process for patients with cirrhosis with a previous hysterectomy. Early recognition of vaginal varices may allow for surveillance and preemptive therapy before life-threatening hemorrhage occurs. However, if this type of patient indeed presents with otherwise unexplained vaginal bleeding, varices must be a diagnostic consideration.
- 4Vaginal variceal hemorrhage in a patient with primary biliary cirrhosis: a case successfully treated by balloon-occluded retrograde transvenous obliteration. Am J Gastroenterol 1999; 94: 3081–3083., , , , , , et al.Direct Link: