| Abstract: | BACKGROUND: Non-alcoholic fatty liver disease (NAFLD) and its progressive form, non-alcoholic steatohepatitis (NASH), are endemic in obesity. We aimed to evaluate the diagnostic accuracy and reproducibility of a simple intraoperative visual liver score to stratify the risk of NASH and NAFLD in obesity and determine the need for liver biopsy. METHODS: This is a prospective cohort study of obese adults undergoing bariatric surgery. The surgical team used a visual liver score to evaluate liver colour, size and surface. This was compared to histology from an intraoperative liver biopsy. RESULTS: There were 152 participants, age 44.6 ± 12 years, BMI 45 ± 8.3 kg/m2. Prevalence of NAFLD was 70.4%, with 12.1% NASH and 26.4% borderline NASH. Single-visual components were less accurate than total composite score. Steatosis was most accurately identified (significant steatosis: AUROC 0.746, p < 0.05; severe steatosis: AUROC 0.855, p < 0.05). NASH was identified with moderate accuracy (AUROC 0.746, p = 0.001), with sensitivity 75% for a score ≥ 2. Stratification into low (≤ 1) and high-risk (≥ 4) scores accurately identified patients who should or should not have an intraoperative biopsy. Most patients with a normal-appearing liver did not have disease (94.4%). The structured visual assessment was quick and interobserver agreement was reasonable (κ = 0.53, p < 0.05). CONCLUSIONS: A simple, structured tool based on liver appearance can be a useful and reliable tool for NAFLD risk stratification and identification of patients who would most and least benefit from a biopsy. A normal liver appearance reliably excludes significant liver disease, avoiding the need for liver biopsy in patients otherwise at high clinical risk of NASH. |
