[Clinical significance of ultrasound combined with serological indexes for predicting nonalcoholic fatty liver disease in patients with chronic hepatitis B with normal or slightly elevated alanine aminotransferase]

Abstract

Objective: To investigate and establish the related factors of non-invasive score model for prediction of non-alcoholic fatty liver disease in chronic hepatitis B patients with normal or mildly elevated alanine aminotransferase (ALT). Methods: A total of 128 cases with chronic hepatitis B who had undergone liver biopsy were included. According to the presence or absence of hepatocyte steatosis on the pathological results of liver biopsy, they were divided into a fatty infiltration and a non-fatty infiltration group. Patients' demographic characteristics, laboratory test indexes, and pathological test results were collected. Univariate and multivariate logistic regression analysis combined with clinical screening variables were used to establish a predictive model. The prediction efficiency of the new model was evaluated by the receiver operating curve, and the difference between the accuracy of the new model and ultrasound in the diagnosis of fatty liver was compared by Delong's-test. Result: Multivariate regression analysis showed that serum triglyceride, serum uric acid and platelets were highly correlated with intrahepatic steatosis (P<0.05). The regression equation triglyceride-uric acid-platelet (TUP)-1=-8.195+0.011×uric acid+1.439×triglyceride+0.012×platelet count was established by combining the above variables. Tthe equation TUP-2=-7.527+0.010×uric acid+1.309×triglyceride+0.012×platelet count+1.397×fatty liver (ultrasound) was established (yes=1; no=0) after incorporating the results of abdominal ultrasound. The diagnostic value of TUP-1 and TUP-2 models for fatty liver was better than that of ultrasound alone and there was no statistically significant difference in diagnostic value between TUP-1 and TUP-2 models (Z=1.453, P=0.146). Conclusion: Compared with abdominal ultrasonography alone, the new model is more effective in diagnosing fatty liver and has good application value.