Reliability of Non-invasive Liver Fibrosis Assessment Tools Versus Biopsy in Pre- and Post-bariatric Surgery Patients with Non-alcoholic Fatty Liver Disease.

Abstract

Liver biopsy (LBx) remains the gold standard to assess fibrosis in non-alcoholic fatty liver disease (NAFLD). Biochemical markers are also useful, but their reliability is not clear in patients with morbid obesity. We assessed the performance of six non-invasive fibrosis assessment tools before and after bariatric surgery (BSx) using LBx. This is a cross-sectional and prospective cohort study. LBx was performed at the time of BSx and 12-month post-operatively and assessed using the Brunt system. Clinical and biochemical measurements were collected at the same time points and six non-invasive fibrosis assessment tools were calculated. One hundred seventy patients had BSx; 79.4% female; age was 46.6 ± 9.8years, and BMI was 48.6 ± 7.5kg/m2. From liver histology, 88% had F0-F2 and 11.2% F3-F4. At BSx, aspartate aminotransferase to platelet ratio index (APRI) and FIB-4 had better accuracy (0.86 and 0.88) with specificity of 96.6% and 94.0% and negative predictive values (NPV) of 88.9% and 93.7%. However, sensitivity (6.7% and 40.0%) and positive predictive values (PPV) (20.0% and 46.2%) were low. Twelve months post-surgery (n = 54), 88.9% of patients had F0-F2 and 11.1% had F3-F4. Fibrosis-4 index (FIB-4) and NAFLD fibrosis score (NFS) had the best accuracy (0.79 and 0.77) with specificity of 83.7% and 86.9% and NPV of 92.3% and 86.9%. However, sensitivity (25% and 0%) and PPV (12.5% and 0%) were low. Overall, FIB-4, APRI, and NFS showed similar performances with higher accuracy, specificity, and NPV. Sensitivity and PPV were low. These tests are more useful at excluding advanced fibrosis.