946 Factors Leading to Discordance Between Vibration-Controlled Transient Elastography and Liver Biopsy in Patients With Non-Alcoholic Fatty Liver Disease

Abstract

INTRODUCTION: Vibration-Controlled Transient Elastography (VCTE) or Fibroscan® is used for the non-invasive assessment of liver fibrosis. However, liver biopsy remains the gold standard for diagnosis due to potential inaccuracies of VCTE including, the patient's body mass index and interobserver variance. Therefore, we examined simple lab parameters and patient related factors that might contribute to this discordance, in patients with Non-Alcoholic Fatty Liver Disease (NAFLD). METHODS: Patients were identified using appropriate ICD 9 and 10 coding for NAFLD. We included patients 18 years and older who underwent a VCTE (both the M and XL probe) and liver biopsy within 6 months of each other. Electronic medical records were reviewed manually to extract data. Variables of interest included VCTE liver stiffness measurement (LSM), VCTE fibrosis score and biopsy fibrosis stage according to the NASH CRN Scoring System. Only valid VCTE results (IQR < 30 and with 10 valid readings) were included. Discordance was defined a difference of more than 2 fibrosis stages between VCTE and biopsy results. Mann-Whitney U test was used to compare continuous variables, Fischer’s exact test to compare categorical variables and multivariate analysis for logistic regression. RESULTS: 109 patients with fatty liver on imaging and elevated transaminases without other types of liver disease underwent a Fibroscan and a subsequent biopsy. In 52% of patients (57/109) the Fibroscan overestimated fibrosis by 2 stages. In univariate analysis, discordance was associated with BMI (36 vs 32.9), ALT (45 vs 72), Fibroscan IQR (17.24 vs 13.5), FIB-4 Score (1.65 v/s 2.25), and NAFLD Activity Score (2.33 vs 3.36), (P value < 0.05 for all). In multivariate analysis, BMI was the only factor independently associated with discordance (P < 0.05). CONCLUSION: Although multiple factors may lead to discordant results for the stage of fibrosis between VCTE and liver biopsy, our results indicate that a higher BMI is the major factor for this discordance with the XL probe. These data also suggest that a stepwise approach combining a noninvasive scoring system and VCTE should be explored as an alternative approach to improve the accuracy of noninvasive methods and minimize the need for liver biopsy in NAFLD patients.