- Pregnancy complicated by cirrhosis is uncommon but is associated with significant risk of maternal morbidity or mortality.
- Hemorrhage from varices, either esophageal or ectopic (varices not located in the esophagus or stomach), may be life-threatening for the pregnant patient.
- Ectopic varices in pregnant patients with portal hypertension are rare but may be identified during routine obstetric ultrasonography, allowing for further imaging and treatment, with the goal of decreasing morbidity associated with variceal hemorrhage at delivery.
Pregnancy in the setting of cirrhosis and portal hypertension is associated with significant maternal risk, including variceal hemorrhage, hepatic decompensation, and death.1–4 Routine screening for esophageal varices in the second trimester of pregnancy is recommended in women with suspected portal hypertension resulting from risk of variceal hemorrhage.4 Severe maternal morbidity or mortality also has been reported after rupture of varices found in locations other than the esophagus or stomach, termed ectopic varices.2,5–7 We present a case in which ectopic abdominal wall and pelvic varices were diagnosed on third-trimester obstetric ultrasonography in a woman with cirrhosis and portal hypertension, allowing for proactive variceal management to decrease hemorrhagic morbidity at delivery.
A 31-year-old woman, gravida 2 para 0010, presented at 18 weeks of gestation as a transfer of care to our facility. Her prior care had been out of state, and limited medical records were available. Her pregnancy was the result of in vitro fertilization, and her medical history was significant for cirrhosis and portal hypertension of unknown etiology. Her history included prior esophageal variceal hemorrhage, two attempted splenic artery embolizations, and thrombocytopenia secondary to portal hypertension and splenomegaly.
On presentation at 18 weeks of gestation, laboratory evaluation demonstrated an alanine aminotransferase level of 38 units/L, aspartate aminotransferase level of 50 units/L, international normalized ratio 1.3, and platelet count of 29×109/L. At 25 weeks of gestation, she underwent screening esophagogastroduodenoscopy, which found no significant esophageal or gastric varices. She was prescribed propranolol to control portal vascular pressures and prevent recurrent variceal bleeding. Her liver function test results remained stable, and her platelets increased over time to 50×109/L.
During fetal growth assessment at 26 weeks of gestation, pelvic varicosities adjacent to the uterus were noted, measuring up to 1.6 cm in diameter. A subsequent ultrasonogram at 30 weeks of gestation again demonstrated pelvic varicosities and a new finding of large varices anterior to the uterus. These vessels appeared to originate at the umbilicus and extend inferiorly within the abdominal wall, crossing the midline, and measured up to 1.4 cm in diameter (Fig. 1).
The suspicion for abdominal wall varices was discussed with a multidisciplinary team, which included representatives from maternal-fetal medicine, hepatology, radiology, and transplant surgery. Mode of delivery was of particular concern because the presence and location of the varices could result in catastrophic maternal hemorrhage with attempted abdominal entry if cesarean delivery was indicated. Vaginal delivery was not without risk either; a recent case report describes spontaneous rupture of intraabdominal varices after vaginal delivery, resulting in maternal death.2 Attempted vaginal delivery also would pose potential risks given the inability to safely proceed with urgent or emergent cesarean delivery.
A contrast-enhanced computed tomography angiogram was obtained to evaluate the patient's abdominal and pelvic vasculature and to assist with delivery planning. The study demonstrated evidence of cirrhosis, with portal hypertension, markedly enlarged portosystemic collaterals, and splenomegaly. A large, recanalized umbilical vein (16 mm) was noted to descend along the right anterior hemiabdomen. There was a periumbilical varix contiguous with bulky varices within the ventral abdominal wall and right rectus abdominis. This network of abdominal veins communicated with tortuous, bulky varices in the right anterior pelvis, ultimately draining into dilated right external iliac veins (Fig. 2). These vessels were noted to cross the midline at multiple points inferior to the umbilicus, prohibiting safe abdominal entry for cesarean delivery by a midline vertical or Pfannenstiel approach owing to risk of massive hemorrhage.
The consensus recommendation from the multidisciplinary conference was to first attempt treatment of the ectopic varices and then proceed with planned cesarean delivery, owing to the risk of uncontrolled or occult hemorrhage with unplanned cesarean or vaginal delivery. Vertical midline entry with assistance from a liver transplant specialist was recommended to allow for adequate visualization of the abdominal wall varices and to improve ability to control hemorrhage should it occur.
The multidisciplinary team recommended transjugular intrahepatic portosystemic shunt placement and umbilical vein embolization to decrease the pressure in the collateral vessels before delivery, providing safe surgical access to the abdomen. Given the risk of massive maternal hemorrhage if cesarean delivery were to be performed before decompression of these vessels, the fetus was not continuously monitored during the transjugular intrahepatic portosystemic shunt procedure and the patient was appropriately counseled regarding both maternal and fetal risks. The interventional radiology procedures were performed under general anesthesia at 35 weeks of gestation. Two right internal jugular vein sheaths were placed, one into the right hepatic vein for transjugular intrahepatic portosystemic shunt creation and one into the inferior vena cava for intravascular ultrasound guidance by intracardiac echocardiography catheter. After portal vein access was obtained, a VIATORR transjugular intrahepatic portosystemic shunt endoprosthesis was deployed, reducing the portosystemic gradient from 15 to 5 mm Hg (Fig. 3). The maternal umbilical vein then was accessed percutaneously and embolized. The estimated procedure-related radiation dose to the fetus from scatter was 3 mGy. Figure 1 shows the drastic change in ultrasound appearance of the varices before the transjugular intrahepatic portosystemic shunt (Fig. 1A and B) compared with after the transjugular intrahepatic portosystemic shunt (Fig. 1C and D).
On postoperative day 2 after the transjugular intrahepatic portosystemic shunt, the patient reported lower extremity swelling with 3+ edema. Lower extremity Doppler ultrasonography was negative for deep venous thrombosis bilaterally. Echocardiogram showed no evidence of cardiac dysfunction or elevated right heart pressures, and liver ultrasonogram revealed patent transjugular intrahepatic portosystemic shunt vasculature. New-onset edema was therefore attributed to physiologic redistribution of blood and plasma volume previously within ectopic varices. The patient's liver function test results remained stable after the transjugular intrahepatic portosystemic shunt. However, there was a drop in the platelet count to a nadir of 41×109/L. She received intravenous immunoglobulin therapy and a course of corticosteroids immediately before planned delivery, with improvement in platelet counts to 65×109/L.
After successful variceal treatment, the decision was made to proceed with delivery. At 36 5/7 weeks of gestation, the patient underwent cesarean delivery under general anesthesia, with a liver transplant specialist assisting with abdominal entry. The abdominal wall was inspected with bedside ultrasonography before incision, and no enlarged abdominal wall vessels were visualized, consistent with posttransjugular intrahepatic portosystemic shunt imaging. A midline infraumbilical vertical skin incision was made. On abdominal entry, enlarged veins were noted below the fascia and rectus muscles, particularly in the periumbilical area but were not involved in the incision and thus avoided. Several large, desiccated vessel lumens were observed along the path of surgical entry. A viable female neonate was delivered, weighing 2,310 g (30th percentile), with Apgar scores of 9 and 9 at 1 and 5 minutes. The estimated blood loss was 1,200 mL. The patient returned to the labor and delivery floor postpartum, and her recovery was uncomplicated. She was discharged home on postoperative day 3, with close multidisciplinary follow-up. She was counseled regarding the significant risks associated with cirrhosis in pregnancy and that future pregnancies would not be recommended as a result of maternal risks.
Portal hypertension resulting from cirrhotic liver disease poses significant maternal risks in pregnancy. Although pregnancy outcome data are limited as a result of the rarity of pregnancy in patients with cirrhosis, reported complications include variceal hemorrhage, hepatic decompensation, and maternal mortality.1–4 Patients with portal hypertension or cirrhosis should be carefully counseled on maternal risks before pregnancy. In a cohort of women with cirrhosis primarily resulting from viral hepatitis, Rasheed et al found that, compared with pregnant women without cirrhosis, pregnant women with cirrhosis had a significantly increased risk of maternal mortality (7.8 vs 0.2%, P=.001). When compared with nonpregnant women with cirrhosis, pregnant women with cirrhosis were more likely to have hepatic decompensation, variceal bleeding, and death.3 Although the generalizability of these findings is limited, they do demonstrate that significant maternal risks are present in the setting of cirrhosis in pregnancy.
Shaheen et al1 evaluated pregnancy outcomes in women with cirrhosis (n=339) compared with matched controls (n=6,625) using data from the U.S. Nationwide Inpatient Sample from 1993 to 2005. Of note, the number of pregnancies in this patient population increased steadily over the time period studied. Women with cirrhosis were at higher risk for both maternal (1.8% vs 0%, P<.001) and fetal (5.2% vs 2.1%, P<.001) mortality. Hepatic decompensation occurred in 15% of women with resultant 12% mortality, and variceal hemorrhage occurred in 5%.
An uncommon but highly morbid manifestation of portal hypertension is ectopic varices. When hemorrhage of ectopic varices occurs in the nonpregnant population, the risk of mortality may be as high as 40%.5 There are few reports of ectopic varices in pregnancy; however, they suggest a high risk of maternal morbidity and mortality. Described sequelae include hemorrhage from abdominal wall varices during cesarean delivery,2,6 maternal death after spontaneous vaginal delivery with intraabdominal variceal rupture,2 and maternal death after postpartum gonadal vein rupture.7
Our case describes identification of ectopic abdominal wall and pelvic varices in a patient with cirrhotic portal hypertension by routine obstetric ultrasonography. This allowed for further investigation with dedicated imaging and treatment with transjugular intrahepatic portosystemic shunt placement, leading to a significant decrease in portal pressures and size of collateral vessels before delivery. The transjugular intrahepatic portosystemic shunt procedure has previously been reported in pregnancy, primarily for treatment of esophageal varices.8 In our patient, it was used as a prophylactic measure to treat ectopic varices and to minimize hemorrhagic complications at the time of delivery. Additionally, identification of the ectopic varices allowed for careful surgical planning before delivery with a plan for midline vertical incision, involvement of transplant surgery for assistance with abdominal entry, and availability of appropriate blood products at delivery.
Based on our experience, we recommend that careful attention be paid to abdominal ultrasonography in women with portal hypertension in pregnancy with interrogation of the abdominal wall and pelvis for evidence of ectopic varices performed by obstetric or radiology services. Additionally, screening esophagogastroduodenoscopy for esophageal varices in the second trimester for women with suspected portal hypertension is recommended by the American College of Gastroenterology.4
At least one ultrasonogram for evaluation of ectopic varices should take place in the late second or third trimester, when the hemodynamic changes of pregnancy have reached their maximum and compression of the venous system by the gravid uterus may increase the presence of portosystemic shunts. If there is concern for ectopic varices, further imaging should be considered. We demonstrated that treatment of varices through transjugular intrahepatic portosystemic shunt insertion before delivery can decrease portosystemic gradients and variceal size and thus could be considered for women with ectopic varices. Careful delivery planning is needed, and cesarean delivery should be considered if an urgent or emergent cesarean delivery cannot be safely performed as a result of the presence and location of varices. Management of these women should occur in a tertiary care center with a multidisciplinary approach.
1. Shaheen AA, Myers RP. The outcomes of pregnancy in patients with cirrhosis: a population-based study. Liver Int 2010;30:275–83.
2. Rosenfeld H, Hochner-Celnikier D, Ackerman Z. Massive bleeding from ectopic varices in the postpartum period: rare but serious complication in women with portal hypertension. Eur J Gastroenterol Hepatol 2009;21:1086–91.
3. Rasheed SM, Abdel Monem AM, Abd Ellah AH, Abdel Fattah MS. Prognosis and determinants of pregnancy outcome among patients with post-hepatitis liver cirrhosis. Int J Gynaecol Obstet 2013;121:247–51.
4. Tran TT, Ahn J, Reau NS. ACG clinical guideline: liver disease and pregnancy [published erratum appears in Am J Gastroenterol 2016;1668]. Am J Gastroenterol 2016;111:176–94.
5. Almadi MA, Almessabi A, Wong P, Ghali PM, Barkun A. Ectopic varices. Gastrointest Endosc 2011;74:380–8.
6. Blackburn P, Patel SS, Mikhail E. Abdominal wall varices in pregnancy—a case report and review of the literature. J Preg Child Health 2016;3:266.
7. Manzl J. Pregnancy and liver cirrhosis. Lethal intraabdominal hemorrhage from a gonadal vein after normal pregnancy 4 days postpartum in a quinquipara [in German]. Zentralbl Gynakol 1974;96:946–9.
8. Ingraham CR, Padia SA, Johnson GE, Easterling TR, Liou IW, Kanal KM, et al. Transjugular intrahepatic portosystemic shunt placement during pregnancy: a case series of five patients. Cardiovasc Intervent Radiol 2015;38:1205–10.