OBJECTIVE: Hepatocellular carcinoma (HCC) is an uncommon but serious occurrence in autoimmune hepatitis. Our objective was to determine predictors for this neoplasm to improve screening strategies. METHODS: Two hundred twenty-seven patients underwent hepatic ultrasonography and serum alpha fetoprotein determinations at 6-12-month intervals. RESULTS: Nine patients developed HCC (4%), and each had cirrhosis > or =73 months prior to the malignancy (mean, 110 +/- 7 months). By univariate Cox analysis, features at accession associated with a higher risk of HCC were: male gender (Hazard Ratio [HR] 7.0, 95% Confidence Interval [CI] 1.87-26.1, P= 0.004), history of blood transfusion (HR 5.6, 95% CI 1.51-21.1, P= 0.01), thrombocytopenia (HR 7.3, 95% CI 1.89-28.3, P= 0.004), ascites (HR 23.8, 95% CI 4.65-121.8, P= 0.0001), esophageal varices (HR 7.9, 95% CI 1.96-31.8, P= 0.004), and any sign of portal hypertension (HR 19.1, 95% CI 3.91-93.3, P= 0.0003). Features after accession associated with a higher risk of malignancy were: treatment for > or =3 yr (HR 7.6, 95% CI 1.25-18.2, P= 0.02), worsening laboratory tests during corticosteroid therapy (HR 7.6, 95% CI 1.81-32.1, P= 0.006), and cirrhosis for > or =10 yr (HR 8.4, 95% CI 1.69-41.9, P= 0.009). CONCLUSIONS: Male gender, features of portal hypertension, history of blood transfusions, immunosuppressive treatment for > or =3 yr, treatment failure, and cirrhosis of > or =10 yr duration identify patients at risk for HCC. These risk factors should focus screening in autoimmune hepatitis.
OBJECTIVE:Hepatocellular carcinoma (HCC) is an uncommon but serious occurrence in autoimmune hepatitis. Our objective was to determine predictors for this neoplasm to improve screening strategies. METHODS: Two hundred twenty-seven patients underwent hepatic ultrasonography and serum alpha fetoprotein determinations at 6-12-month intervals. RESULTS: Nine patients developed HCC (4%), and each had cirrhosis > or =73 months prior to the malignancy (mean, 110 +/- 7 months). By univariate Cox analysis, features at accession associated with a higher risk of HCC were: male gender (Hazard Ratio [HR] 7.0, 95% Confidence Interval [CI] 1.87-26.1, P= 0.004), history of blood transfusion (HR 5.6, 95% CI 1.51-21.1, P= 0.01), thrombocytopenia (HR 7.3, 95% CI 1.89-28.3, P= 0.004), ascites (HR 23.8, 95% CI 4.65-121.8, P= 0.0001), esophageal varices (HR 7.9, 95% CI 1.96-31.8, P= 0.004), and any sign of portal hypertension (HR 19.1, 95% CI 3.91-93.3, P= 0.0003). Features after accession associated with a higher risk of malignancy were: treatment for > or =3 yr (HR 7.6, 95% CI 1.25-18.2, P= 0.02), worsening laboratory tests during corticosteroid therapy (HR 7.6, 95% CI 1.81-32.1, P= 0.006), and cirrhosis for > or =10 yr (HR 8.4, 95% CI 1.69-41.9, P= 0.009). CONCLUSIONS: Male gender, features of portal hypertension, history of blood transfusions, immunosuppressive treatment for > or =3 yr, treatment failure, and cirrhosis of > or =10 yr duration identify patients at risk for HCC. These risk factors should focus screening in autoimmune hepatitis.