4 6.0 (ALT-N)   10.8 9.0 (ALT-N)       7.5 (ALT 1-5x)     12.0 (ALT 1-5x)   Metformin cost Sensitivity (%) 95 90 90 26 51 66 Specificity (%) 5 63 66 92 96 92   PPV(%) 40 62 64 69 91 85 NPV(%) 58 90 90 65 74 80   LR+ 1.0 2.45 2.65 3.36 14.0

8.47 LR- 1.06 0.16 0.16 0.40 0.51 0.37   No. of biopsy correctly avoided 7/142(5%) 90/142(63%) 94/142(66%) 25/96(26%) 49/96(51%) 63/96(66%)   No. of incorrect diagnosis 5/142(4%) 10/142(7%) 10/142(7%) 11/96(11%) 5/96(5%) 11/96(11%) Validation Sensitivity (%) 97 79 79 32 41 65 Specificity (%) 14 51 61 86 94 86 PPV(%) 43 52 57 61 82 76 NPV(%) 88 79 82 66 71 79 LR+ 1.12 1.62 2.02 2.36 7.0 4.71 LR- 0.21 0.40 0.34 0.78 0.62 0.41 No. of biopsy correctly avoided 7/51(14%) 26/51(51%) 31/51(61%)

11/34(32%) 14/34(41%) 22/34(65%) No. of incorrect diagnosis 1/51(2%) 7/51(14%) 7/51(14%) 7/34(21%) 3/34(9%) 7/34(21%) Disclosures: Grace LH Wong – Advisory Committees or Review Panels: Otsuka, Gilead; Speaking and Teaching: Echosens, Furui Vincent W. Wong – Advisory Committees or Review Panels: Otsuka, Roche Pharmaceuticals, Selleck ITF2357 Gilead, Abbott; Speaking and Teaching: Bristol-Myers Squibb, Novartis Pharmaceuticals, Echosens Henry Lik-Yuen Chan – Advisory Committees or Review Panels: Gilead, Vertex, Bristol-Myers Squibb, Abbott, Novartis Pharmaceutical, Roche, MSD Background / Aims: The common causes of portal hypertension in India are cirrhosis, non cirrhotic portal hypertension (NCPF) and extra hepatic portal hypertension (EHPVO). The present gold standard to differentiate NCPF from cirrhosis is liver biopsy; there is paucity of data to differentiate these by non invasive means. We therefore assessed the reliability of LS to differentiate cirrhosis from NCPF. Methods: LS was measured, using 5-MHz US transducer probe mounted on the axis of a vibrator (Fibro Scan; Echosens), Ergoloid in 34 consecutive biopsy-proven NCPF, 44 EHPVO and 41 Child A cirrhosis patients. 43 healthy subjects were evaluated as controls. EHPVO was diagnosed if there was portal cavernoma with or without splenic vein cavernoma, and no evidence of liver disease. Patients with NCPH were

excluded if they had hepatitis B or C, alcohol consumption, recent variceal bleed, portal biliopathy or previous shunt surgery. Institutional ethics committee approval and informed consent were obtained. ROC curves were plotted to assess the sensitivity and specificity of LS in differentiating NCPF from Child A cirrhosis. Results LS was higher in patients with EHPVO and NCPF than in healthy controls (p = 0.008). LS was similar in EHPVO patients who presented with bleed as compared to those who did not (6.23 [1.4] kPa vs.6.32 [1.2] kPa, p=0.50). LS was similar in NCPF patients who had variceal bleed at presentation as compared to those who had not (7.6 [2.7] vs 7.1 [3.3], p=0.41). NCPF patients who presented with bleed were older (37 [23-70] y vs.