P16 Non-invasive detection of oesophageal varices: comparison of non-invasive assessment of systemic haemodynamics with laboratory parameters and predictive scores

Introduction Endoscopic screening for varices (OV) is advised in cirrhosis, repeated every 1–3 years, with primary prophylaxis given to large OV. This is costly to endoscopy units, unpleasant for patients and multiple procedures may affect compliance. Cirrhosis is characterised by a hyperdynamic circulation; novel tools make non-invasive assessment possible. Aim Toassess accuracy of non-invasive tests of systemic haemodynamics (Cardiac output and index (CO, CI), systemic vascular resistance (SVR), baroreceptor sensitivity (BRS) to detect OV and compare with other non-invasive methods (Child–Pugh score (CPS), albumin, platelet count, prothrombin time (PT), AST/ALT ratio, platelet count/spleen diameter ratio). Method Prospective study of 29 cirrhotic patients. Systemic haemodynamics were assessed non-invasively with the Finometer®, hepatic venous pressure gradient (HVPG) assessed portal pressure and gastroscopy for variceal size (none/small (absent), medium/large (large)). Results 69% male, median age 47 (42–55) years, CPS 6 (Class A 18, B 10, C 1) and MELD 10 (8–13). Prevalence OV 79%, large 38%. Significant differences in haemodynamics were seen between patients classified as absent or large OV (CO 5.6 vs 8.0lpm, CI 3.0 vs 4.5 l/min/m2, SVR 1.17 vs 0.77MU, HVPG 14 vs 19 mm Hg, BRS 5.8 vs 3.2 ms/mm Hg, CPS 5 vs 7, respectively). Comparisons summarised in the Abstract P16 Table 1. At a cutoff of 7.15lpm, CO predicted large OV with 73% sensitivity, 78% specificity and correctly classified 76% of patients. At a cutoff of 3.66 l//min/m2, CI predicted large OV with 82% sensitivity, 83% specificity and correctly classified 79% of patients. This compares to 78% correctly classified using HVPG, 76% CPS and 59% PT. Conclusion Non-invasive assessment of systemic haemodynamics appears a promising technique to identify cirrhotic patients at risk of having large oesophageal varices. Larger prospective validation studies need to be performed. Standard laboratory tests and predictive scores (except Child–Pugh score) are not reliable to accurately predict large oesophageal varices.