Long-Term Azathioprine Use and the Development of Non-Cirrhotic Portal Hypertension

Abstract

Introduction: Non-cirrhotic portal hypertension (NCPH) is caused by intra- and extrahepatic etiologies. NCPH presents with signs of portal hypertension, especially variceal bleeding and splenomegaly. Extrahepatic portal vein thrombosis, nonalcoholic steatohepatitis, and primary sclerosing cholangitis are the most prevalent causes in the Western world. In rare cases, immunosuppressive drugs, particularly azathioprine (AZA), 6-mercaptopurine, and 6-thioguanine (TG), have been implicated in nodular regenerative hyperplasia, an idiopathic form of NCPH. Here we describe a case of NCPH with obstructive portal venopathy possibly induced by long-term AZA use. Case: A 44-year-old male with type I diabetes and Crohn's Disease was referred to our hepatology clinic for management of presumed cirrhosis. At the age of 40, after 20 years on AZA, EGD demonstrated grade II esophageal varices. Ultrasound revealed increased liver echogenicity and marked heterogeneity, as well as portal vein dilatation and splenomegaly. Beta-blocker therapy was initiated and AZA was discontinued. Adalimumab was started but was switched to vedolizumab one year later due to coccidioidomycosis infection. Labs were significant for thrombocytopenia, leukopenia, and microcytic anemia. MRI on presentation to our facility demonstrated splenomegaly and patent portal veins but no signs of cirrhosis. Ultrasound-guided liver biopsy demonstrated dense portal fibrosis and focal obliteration of portal vein branches. Portal vein branches were absent from portal tracts, and outlet venules were dilated (Figure 1). There was no evidence for clinical, radiologic, or histological cirrhosis or fatty liver disease.2250 Figure 1. Absence of portal veins from portal tracts with dilatations of the outlet venules (“herniated” portal veins, arrows), consistent with obstructive portal venopathy (A. H&E stain x200; B. Trichrome stain x100).Discussion: While the pathogenesis of AZA-induced NCPH remains uncertain, there is evidence that glutathione depletion, which is toxic to sinusoidal endothelial cells, leads to veno-occlusive disease. Whether occlusive venopathy is attributable to 6-TG toxicity alone remains unclear. Portal fibrosis and obstructive venopathy on histology in the absence of vascular occlusion on imaging suggest that portal hypertension in our patient was heralded by intrahepatic changes. Damage to endothelium by 6-TG represents one possible mechanism; however, it is just as feasible that an autoimmune process led to NCPH in this patient. We thus propose further research into possible associations between AZA and NCPH.