UK-PBC Risk Score Research Articles

O021 : Non-invasive screening of large esophageal varices using endoscopic capsule and/or liver fibrosis tests

Background and Aims: The non-invasive screening of large esophageal varices (LEV) might be useful in primary prevention of variceal bleeding in cirrhosis. The main objective was to compare strategies for the diagnosis of LEV by using predictive values and direct costs. Methods: 330 cirrhotic patients were included in a prospective multicenter study. Diagnostic tools were upper gastro-intestinal endoscopy (UGIE, reference), esophageal capsule endoscopy (ECE), liver stiffness measurement (LSM) by Fibroscan and several blood markers and liver fibrosis tests. Results: 287 patients had per protocol analysis defined by UGIE and ECE available. Child–Pugh class distribution was: A: 61%, B: 20%, C: 19%. Etiology was: alcohol: 45%, viral hepatitis: 27%, NASH: 5%, mixed: 11%, others: 6%. Among the 35 diagnostic predictors tested in the per protocol population, 10 had an AUROC >0.7 for LEV: Child–Pugh score (0.72), bilirubin (0.72), AST/ALT (0.74), LSM (0.74), hyaluronate (0.77), QuantiMeter (0.77), FibroMeter (0.77), CirrhoMeter (0.77), Elasto-FibroMeter (0.78), ECE (0.89). The most performant combination by binary logistic regression was either with ECE: ECE + AST/ALT (score 1, AUROC: 0.92) or without ECE: prothrombin index + AST/ALT (score 2, AUROC: 0.77), LSM having no independent role. Then, we defined 5 strategies based on noninvasive tools with high negative predictive value (NPV ≥95%) and/or positive predictive value (PPV ≥90%) for LEV in a first step; then, UGEI was used in the indeterminate zone. The 5 strategies were: S1: ECE for NPV and PPV, S2: LSM for NPV and ECE for PPV, S3: LSM for NPV and score 2 (ECE + blood markers) for PPV, S4: score 1 (blood markers) for NPV and PPV, S5: score 2 (ECE + blood markers) for NPV and PPV. The rank of strategies for decreased request to UGEI was: S5: 12.9%, S1: 28.9%, S2: 41.7%, S3:42.2%, S4: 47.5% (p < 0.001 between all and S5 vs others). But the rank for direct costs per patient was different: S4: 64€, S2: 393€, S3: 399€, S5: 632€, S1: 645€, reference UGEI cost being 114€. Conclusions: This is the first study comparing capsule endoscopy and fibrosis tests in the screening of large esophageal varices. By using capsule and blood markers, one can avoid 87% of endoscopies in cirrhotic patients but this strategy has currently high direct costs (x6). By contrast, simple blood markers can avoid 48% of endoscopies and divide the direct costs by around 50%. These results should be validated in an independent population, compared to spleen stiffness and add indirect costs. O022 THE UK-PBC RISK SCORE: DERIVATION AND VALIDATION OF A RISK SCORE TO PREDICT LIVER EVENTS IN THE UK-PBC RESEARCH COHORT M. Carbone, S.J. Sharp, M.A. Heneghan, J.M. Neuberger, G.M. Hirschfield, A.K. Burroughs, D. Thorburn, A. Bathgate, M. Aldersley, C. Adgey, P. Trembling, K. Williamson, L. Jopson, R.T. Lim, N.J. Wareham, H.J. Cordell, G.J. Alexander, J.E. Jones, R.N. Sandford, G.F. Mells, UK-PBC Consortium. Division of Gastroenterology and Hepatology, Addenbrooke’s Hospital, Academic Department of Medical Genetics, MRC Epidemiology Unit, University of Cambridge, Cambridge, Institute of Liver Studies, King’s College Hospital, London, Liver Unit, Queen Elizabeth Hospital, NIHR Biomedical Research Unit and Centre for Liver Research, University of Birmingham, Birmingham, The Sheila Sherlock Liver Centre, The Royal Free Hospital, London, Liver Unit, Royal Infirmary Hospital, Edinburgh, Liver Unit, St James’s Hospital, Leeds, Liver unit, Royal Victoria Hospital, Belfast, Liver Unit, Barts and The London NHS Trust, London, Translational Gastroenterology Unit, John Radcliffe Hospital, Oxford, Institute of Cellular Medicine, Institute of Genetic Medicine, Newcastle university, Newcastle, United Kingdom E-mail: mc767@medschl.cam.ac.uk

The UK-PBC risk score: Derivation and validation of a risk score to predict liver events in the UK-PBC research cohort

The UK-PBC risk score: Derivation and validation of a risk score to predict liver events in the UK-PBC research cohort