Liver inflammation and fibrosis may impair the ability of sonography to identify steatosis. We determined the accuracy of sonography in grading steatosis in patients with chronic hepatitis B compared to liver biopsy.Methods
We conducted was a single-center retrospective study of all nontransplanted patients with chronic hepatitis B undergoing sonography and liver biopsy between 2004 and 2014 (n = 109). Steatosis was graded by sonography as none, mild, moderate, or severe. Liver histologic analysis graded steatosis (0, <5%; 1, <33%; 2, <66%; or 3, ≥66%) and staged fibrosis (F0–F4). Severe steatosis was defined as grade 2 or 3. Clinical variables within 6 months of liver biopsy were collected, and the association with steatosis was analyzed by univariate logistic regression.Results
Patients were predominantly Asian (83%), male (62%), and hepatitis B e antigen negative (62%). Twenty-nine percent of patients were obese; 9% had diabetes mellitus; 23% had hypertension; and 31% had dyslipidemia. Forty-four percent of patients had steatosis on liver biopsy; 8% had severe steatosis. The presence of any steatosis on sonography correctly identified any steatosis on liver biopsy in 29 of 48 patients (60%). The absence of steatosis on sonography ruled out severe steatosis on biopsy (specificity, 100%). Severe steatosis on sonography correctly predicted the presence of severe steatosis on liver biopsy (89%; P < .001); however, it was not accurate at distinguishing between steatosis grades. Predictors of biopsy-proven steatosis on univariate analysis included diabetes (P < .001), hypertension (P = .03), hypercholesterolemia (P = .02), and body mass index (P < .001).Conclusions
Sonography had excellent accuracy in identifying patients with steatosis on biopsy. Abdominal sonography can be used to predict clinically important steatosis in patients with chronic hepatitis B.