Current clinical indications and midterm outcomes of surgical portosystemic shunts in adults

Abstract

Background: Portosystemic shunts were originally described for the management of bleeding varices due to cirrhosis. However, in the era of transjugular intrahepatic portosystemic shunts and liver transplant, these procedures are increasingly rare. We reviewed indications, operative details and outcomes of patients undergoing surgical portosystemic shunt procedures. Methods: Clinical and procedural details of consecutive patients undergoing creation of a portosystemic shunt from Jan 1997 to August 2018 were reviewed retrospectively. Endpoints were mortality, primary patency, secondary patency and shunt-related complications. Clinical success was defined as postoperative resolution of the primary symptoms for which the shunt was performed. Univariate Cox regression analysis was performed to search for risk factors for graft thrombosis (loss of primary patency). Subsequently, we compared patency and survival in subgroups: Group I: patients with cirrhosis vs those with non-cirrhotic portal hypertension and Group II: patients with computed tomography signs of portomesenteric venous thrombosis (PMVT) vs those without PMVT. Results: Surgical portosystemic shunts were performed in 99 patients (54 female, mean age 46 years). Indications were refractory gastrointestinal bleeding in 75; ascites in 24; secondary to PMVT in 63 (portal n = 46, hepatic n = 15, splenic n = 17, superior mesenteric vein n = 16) and cirrhosis in 36. PMVT was unprovoked in 45 patients (thrombophilia identified in 17). Comorbidities included hypertension (21), diabetes (16), hyperlipidemia (8), coronary artery disease (5) and smoking (16); average MELD score was 8. Shunts were nonselective in 45 (portocaval n = 25, mesocaval n = 17, mesorenal n= 2, rex n = 1) and selective in 50 (distal n = 6, central n = 48). Procedural success was 99%; average length of stay was 8 (3–122) days. Complications occurred in 55% [hematoma (23), bowel ischemia in (3), encephalopathy in (19), new PMVT (10)]. Early mortality was 2% (1 pulmonary embolism, 1 coagulopathy). Early graft thrombosis occurred in 9; 7 underwent re-intervention (1 endovascular, 6 open- patency restored in 3). Symptom relief was not achieved in 6 with early graft thrombosis. It was complete in 76; partial in 6 (3 patent, 2 thrombosed) and recurrent in 3 (graft thrombosis). Therapeutic anticoagulation was prescribed in 35. The median clinical follow-up was 3.2 (1-20) years and imaging follow-up 1 (1–20) years. On Kaplan-Meier analysis, overall survival was 90%, 73% and 49%; primary patency 84%, 82% and 82%; secondary patency 90%, 85% and 82% at 1, 5 and 10 years respectively. The risk of loss of primary patency [HR of 9.22 (95% CI 1.22–70.27, p = 0.032] but not secondary patency was higher in Group II (Fig 1) with no difference in age adjusted survival in either group. No other predictive factors for shunt thrombosis were identified. Conclusion: Surgical portosystemic shunts are safe and effective for symptom relief in selected patients with portal hypertension. The risk of graft thrombosis is 9-times higher in patients with PMVT and is not influenced by use of therapeutic anticoagulation. Overall survival is similar in those with/without cirrhosis or PMVT. Rethrombosis is associated with symptom recurrence; open re-intervention is advisable for symptom relief when suitable targets are available.