an ALT of 938 IU/L, and bilirubin of 1.3 mg%. All serology for acute or chronic viral hepatitis was negative as was testing for metabolic disorders. However, several auto-antibodies were present including an ANA of 1:640 speckled- pattern, an anti-smooth muscle Ab of 1:20, an elevated thyro- globulin antibody (4 IU/L ) and thyroid microglobulin antibody (339.7 IU/mL). It was noted that he had a nearly undetectable TSH (0.03 IU/L)and an elevated total T3 (179.12 ng/dL). An eosinophil count of 15% was present. One week later his AST peaked at 695, and his ALT at 1,256. His bilirubin remained normal, but he had a slight prolongation of the INR to 1.4. Over the next 5 months his transaminases and INR normalized. An endocrinology consultation was obtained, and concluded that the patient’s subclinical hyerthyroid state was consistent with autoimmune thyroiditis based on results of a radionucleotide thyroid scan. Discussion: Drug-induced autoimmune hepatitis is unusual, with no prior reports to our knowledge of pyrazinamide or rifampin-related cases. These drugs also have not been implicated in autoimmune thyroiditis. In this case, we suspect the combination of the two drugs was responsible for both autoimmune hepatitis and thyroiditis secondary to either a hypersensitivity or immunoallergic reaction, that cleared after treatment was withdrawn. 284 PREDICTORS OF PERIPHERAL VENOUS THROMBOEMBOLISM IN CIRRHOSIS Patrick G. Northup, M.D., Agata Volk-Bednarz, M.D., Carl L. Berg, M.D.*. University of Virginia Health System, Charlottesville, VA. Purpose: Clinical experience suggests that cirrhosis patients are at risk for peripheral venous thromboembolic events (VTE) such as deep vein throm- bosis or pulmonary embolism yet there is a near absence of published literature in this area. This study was designed to define the incidence of VTE in hospitalized cirrhosis patients, to determine if underlying coagu- lopathy due to chronic liver disease offers protection from VTE, and to elucidate possible predictors for VTE in this patient population. Methods: A case-control design with historical controls was used. All inpatients between the years 1993 and 2001 with diagnoses of both VTE and cirrhosis were matched with equal numbers of controls with cirrhosis only. Cases and controls were matched for age, gender, and race as well as for known risk factors for VTE including significant comorbidities and occurrence of surgical procedures. Demographic, laboratory, and outcomes data were compared between cases and controls. Results: Upon review of more than 21,000 admissions of patients with cirrhosis, 247 admissions involving 220 separate patients were identified with VTE events. The corresponding incidence rate of VTE was 1.2 percent. These 220 patients with VTE were matched with 220 controls. There was no statistical difference between cases and controls in age, gender, race, or comorbidities. Cases and controls also did not differ significantly in level of endogenous coagulopathy as measured by admis- sion INR, platelet count, or severity of liver dysfunction as measured by admission Model of End-stage Liver Disease (MELD) score. On admis- sion, patients with VTE events were found to have significantly lower serum albumin (2.76 vs. 3.05 g/dL, p0.001), higher serum creatinine (1.83 vs. 1.51 mg/dL, p=0.012). A linear inverse relationship was found correlating risk of VTE and serum albumin level. Conclusions: VTE events occur in about one percent of admitted patients with cirrhosis. Endogenous coagulopathy appears not to significantly pro- tect the patient from VTE. Serum albumin is inversely proportional to the risk for VTE and high serum creatinine influences this risk as well. Serum albumin may be a surrogate marker for levels of endogenous anticoagulants such as antithrombin III, protein C, and protein S and those levels may fluctuate with renal function. Further prospective studies investigating anticoagulant and procoagulant protein levels and safety of medical anti- coagulation in patients with chronic liver disease are warranted. 285 ANALYSIS OF P53 EXPRESSION, MICROSATELLITE INSTABILITY AND PROLIFERATION RATE IN HEPATOCELLULAR CARCINOMAS FROM MOZAMBIQUE Jose Ferro, M.D., Jose Alexandre Sarmento, M.D.*, Fatima Carneiro, Ph.D., Tavarela Veloso, Ph.D. University Eduardo Mondlane, Maputo, Mozambique; Hospital S. Joa ˜o, Porto, Portugal and Universidade do Porto, Porto, Portugal. Purpose: The incidence of hepatocellular carcinoma (HCC) in Mozam- bique is one of the highest in the world and the prevalence of hepatitis B virus infection (HBV) is also very high in this country. In the present study we analysed the clinicopathological and some molecular features of 65 HCCs arising in patients from Mozambique . Methods: Cases were classified according to WHO classification and HBV infection was analysed by immunohistochemistry, using antibodies against surface and core antigens of the virus. We analysed the expression of p53 and MIB-1 by immunohistochemistry and searched for microsatellite by the analysis of a mononucleotide repeat marker (BAT-26). Results: Mean age of the patients was 43.615.9 and the male/female ratio was 1.5/1. By histology, the majority of tumors displayed a trabecular pattern (71%). Signs of HBV infection were detected in 33,8% of the cases. Cirrhosis was observed in non-neoplastic liver in 52% of the cases. We compared cases developing in liver cirrhosis with those arising in non- cirrhotic liver. Despite the similarity of these two groups regarding age and gender of the patients, as well as the presence of signs of HBV infection, cases developing in non-cirrhotic liver displayed a rate of immunoexpres- sion of p53 (19,2%) that was significantly higher (p=0.017) than the observed in the cases associated with liver cirrhosis (6.8%). Proliferative index, evaluated by MIB-1 expression, was also significantly higher (p=0.0007) in the former group (16,1%3.4%) than in the later (2.7%1.1%). Microsatellite instability was only observed in HCCs de- veloped in non-cirrhotic liver (26.3%) but the size of the sample is too small to allow a definite conclusion on this issue. Conclusions: The ”mollecular“ profile of HCCs developing in non-cir- rhotic liver is distinct from that obseved in HCCs occuring on a background of liver cirrhosis, thus suggesting a different pathogenesis of the two types of HCC. It remains to be clarified if these differences are related with specific etiologic factors (namely HBV, alone or in association with afla- toxins). 286 COMPATIBLE ABO MISMATCH AND LIVER TRANSPLANTATION—A SINGLE TRANSPLANT CENTER’S EXPERIENCE Murat Aladaq, M.D., Sajid Jalil, M.D., Harlan I. Wright, M.D., Ahmet O. Gurakar, M.D.*, Bakr M. Nour, M.D., Anthony Sebastian, M.D. Nazih Zuhdi Transplantation Institute, Oklahoma City, OK. Purpose: Ideally, organ grafting is usually carried out between recipient and donor of identical blood group to minimize the influence of blood group specific antigens on graft and patients survival. The shortage of suitable donor organs and the clinical urgency can lead in to implantation of grafts from ABO mismatched donors. In liver transplantation, it has been suggested that patient and graft survival is diminished with the implantation of ABO incompatible grafts. Indeed, 1-year graft survival in this setting has been reported to range between 25% and 75% less than that of ABO- identical or compatible grafts. We reviewed and compared our experience with transplantation of ABO identical and compatible mismatched livers. Methods: 520 orthotopic liver transplantations were performed between November 1992 and May 2003 in our institution. 55 of them were ABO compatible mismatched transplants. We retrospectively reviewed the data and compared patient and graft survival. Results: Overall 1 month, 1 year, 5 year and 10 year patient survival among mismatched and identical groups were 90.9%–97.1%, 88.5%– 89.8%, S97 AJG – September, Suppl., 2003 Abstracts