Endocrine Abstracts (2017) 50 P315 | DOI: 10.1530/endoabs.50.P315

Advanced NAFLD is common in bariatric surgical patients, poorly staged by non-invasive NAFLD biomarkers and is associated with adverse outcome

Niall Dempster1, Ioannis Gerogiannis2, Rachel Franklin1, Michael Watson3, Lisa Rickers2, Caroline Fletcher2, Eleanor Jenkins2, Bruno Sgromo2, Richard Gillies2, Jeremy Cobbold3, William Rosenberg4, Leanne Hodson1, Jeremy Tomlinson1 & John Ryan3

1Oxford Centre for Diabetes, Endocrinology and Metabolism, Churchill Hospital, Oxford, UK; 2Oxford Bariatric Service, Churchill Hospital, Oxford, UK; 3Hepatology Unit, John Radcliffe Hospital, Oxford, UK; 4Hepatology Unit, Royal Free Hospital, London, UK.

Introduction: Non-Alcoholic Fatty Liver Disease (NAFLD) is reported to be common in patients undergoing bariatric surgery, but the diagnostic accuracy of non-invasive NAFLD biomarkers to stage disease and track progress longitudinally has not been assessed.

Methods: 274 patients undergoing bariatric surgery were included in the study. Intra-operative liver biopsies were taken from 163 patients and histologically graded using the NAFLD Activity Score and Kleiner classifications. Established non-invasive biomarkers were measured (Enhanced Liver Fibrosis (ELF), AST/ALT, APRI, BARD, Fib-4, NAFLD Fibrosis Score (NFS)) and ROC curve analysis was used to determine the utility of each test. Biomarkers and metabolic health markers (including HbA1c) were recorded prior to surgery (mean 17.2+/−20.3 days), 6 and 12 months post-operatively and compared using repeated measures ANOVA.

Results: Steatosis was present in 84.6%, advanced fibrosis in 23.1% and non-alcoholic steatohepatitis (NASH) in 11.9% of intra-operative biopsies. All biomarkers performed poorly in identifying advanced fibrosis (AUROC=0.60-0.69). ELF best predicted cirrhosis (AUROC=0.73), however NICE-recommended ELF cut-off thresholds failed to predict 94% of cases of advanced fibrosis and 80% of patients with cirrhosis.

There was rapid post-operative weight reduction with 67.7% excess weight loss 12 months post-operatively (P=0.00). Metabolic health markers improved during the 12 month post-operative period (including HbA1c 6.2% to 5.3% P=0.00, total cholesterol:HDL ratio 4.1 to 3.3 P=0.00). In contrast, NAFLD biomarkers worsened (including AST/ALT 0.9 to 1.2 P=0.01; Fib-4 1.0 to 1.1 P=0.03).

With a minimum 1 year of post-operative follow-up (n=215), hepatic decompensation incidence was 2.3%, occurring a median of 342 days post-operatively. Decompensated liver disease was associated with intra-operative biopsy-proven cirrhosis (P=0.01).

Conclusion: Advanced NAFLD is common in bariatric surgical patients and intra-operative cirrhosis is associated with hepatic decompensation in the first post-operative year. Currently available non-invasive biomarkers poorly predict histological NAFLD severity and there is a disconnect between improvement in metabolic variables and non-invasive NAFLD assessments post-operatively. Improved biomarkers and an increased awareness of the prevalence of advanced NAFLD in patients undergoing bariatric surgery are required.