Background: The factors predictive of the presence or the absence of esophageal varices in hepatitis C virus (HCV) and advanced fibrosis have not been defined.
Objectives: To define the prevalence of esophageal varices and the factors that are positively and negatively with such varices in hepatitis C and advanced fibrosis.
Design: A prospective study of esophageal varices and associated risk factors in subjects with hepatitis C and advanced fibrosis.
Setting: Prerandomization data from the HALT-C (hepatitis C long-term antiviral treatment against cirrhosis) clinical trial.
Patients and intervention: Subjects with bridging fibrosis or cirrhosis, who were virologic nonresponders to treatment with pegylated interferon alpha 2a and ribavirin, underwent endoscopy.
Results: Sixteen percent of subjects with bridging fibrosis (95/598) and 39% of subjects with cirrhosis (164/418) had varices (P < .0001); 2% of subjects with bridging fibrosis (13/598) and 11% of those with cirrhosis (48/418) had medium or large varices. Subjects with bridging fibrosis and varices had a significantly lower platelet count and higher bilirubin and international normalized ratio (INR) compared with those without varices, suggesting that the biopsy may have underestimated the severity of fibrosis. A platelet count >150,000/mm(3) was associated with a negative predictive value of 99% for esophageal varices. By logistic regression modeling, African American race and female sex were protective, whereas a lower platelet count and higher bilirubin and INR predicted varices (c statistic, 0.758).
Conclusions: The risk of having varices increases with decreasing platelet counts, increasing bilirubin, and INR. The probability of having medium or large varices at platelet counts >150,000/mm(3) is negligible in this population.