s / Pancreatology examination with bile from biliary tract or duodenum has been useful for the diagnosis of microlithiasis. Aims: We evaluated the reliability of bile samples collected directly from the biliary tract during ERCP for polarized microscopic examination. Patients m to identify factors favoring necrosis liquefaction. Patients & methods: Fifty-eight patients managed by a step-up approach or open necrosectomy in a tertiary referral center (2008 – 2012) were assigned to 3 groups, according to necrosis status at the intervention time: G1 – 26patients with solid necrosis; G2 – 28patients semisolid walled-off necrosis (WOPN); G3 – 4patients liquefied WOPN. Groups outcome were compared. Logistic regression was used to identify factors favoring necrosis liquefaction. Results: Median timing of intervention was 27days (G11⁄421, G21⁄431, G31⁄437days). Twenty patients died (34.48%): G11⁄453.8%, G21⁄421.4% and G31⁄40%. Open necrosectomy was performed in all G1 patients. Percutaneous catheter drainage (PCD) was primarily used in G2 and G3, with 41% and 100% success. When PCD failed, VARD or open necrosectomy followed. Mean hospital stay was G11⁄489days, G21⁄477days and G31⁄451days. Factors favoring necrosis liquefactions were: time from pancreatitis onset, age > 55 years, hypertriglyceridemic etiology, high amylase content, antibiotic prophylaxis > 2 weeks. Conclusion: Interventional debridement should be postponed in solid necrosis, where enzymatic debridement may be tried as suggested by one of the found factors (high amylase content). Antibiotic prophylaxis > 2 weeks favors liquefaction only by postponing necrosectomy. Time is the strongest factor favoring necrosis evolution to semisolid and liquefied WOPN status, when a step-up approach offers better survival and shorter hospital stay. PII-48 Abstract id: 161. Predictive markers for severe acute pancreatitis: A comparative prospective study within a representative cohort Hanna Sternby, Hannes Hartman, Dorthe Johansen, Stefan Appelros, Henrik Thorlacius, Sara Regn er. Department of Surgery, Sk ane University Hospital Malm€ o{, Lund University, Sweden Introduction: In order to predict severity of Acute Pancreatitis (AP) a large number of biomarkers have been studied retrospectively and in different cohorts. Aims: In this study promising biomarkers will be studied prospectively in a representative cohort of patients with AP aiming to enable comparisons of their predictive capacity in a clinical setting. Patients & methods: Patients with AP admitted to the Malm€ o University Hospital were consecutively included in the study. Blood samples were obtained on admission and daily up to 72 hours. Socioeconomic factors and information of importance for etiology and severity of AP were recorded in a database. Cut off values were set by reviewing the literature for the predefined biomarkers. Results: 202 patients with AP, 111 men and 91 women, with a median age of 66 years (19-97) were included in the study. 13,4% of the patients had severe AP according to the Atlanta -92 criteria while the corresponding figures for the Revised Atlanta-12 were 5,4%. 15, 3% were classified as Moderately Severe AP. Etiology was biliary in 51,5%, alcohol in 15,8%, unknown in 23,3% and of other origin in 9,4%. CRP on Day 3 had a median value of 377 mg/l in patients with SAP and 111 mg/l in patients with mild AP. Conclusion: The clinical data shows that the cohort has a representative composition in order to obtain reliable results concerning patients with AP. Preliminary data shows a significant difference in IL-6 and IL-8 between mild AP and severe AP. Further blood samples will be analysed during the first half of 2013. PII-49 Abstract id: 68. New classifications of severity of acute pancreatitis: Validation of determinant-based and revision of the Atlanta classification Enrique de-Madaria, Nelly G. Acevedo-Piedra, Neftali Moya-Hoyo, M onica Rey-Riveiro, Inmaculada Lopez-Font, Juan Martinez, Martinez Lluis. Unidad de Patologia Pancreatica, Hospital General Universitario de Alicante, Spain Introduction: The Atlanta classification was published in 1993; definitions were given regarding local and systemic complications, as well as severity of acute pancreatitis (AP). After 2 decades we have learned new concepts about the natural history of AP that have prompted a revision of the Atlanta classification. Two new classifications have been proposed. Aims: Our aim was to validate both new classifications. Patients & methods: We analyzed a prospective database which included every adult patient with AP admitted to a third level hospital between December 2007 and January 2013. Every CT scan was retrospectively reviewed according to new definitions. The classifications were validated in terms of outcomes: length of hospital stay, need for ICU admission, nutritional support, invasive treatment and mortality. Results: We analyzed 543 episodes of AP. The most frequent etiology was gallstones (59.5%), followed by alcohol (13.6%). Pancreatic necrosis was present in 66 (12.2%) of the patients, peripancreatic fat necrosis in 109 (20.1%), acute necrotic collections in 106 (19.5%), walled-off necrosis in 61 (11.2%), peripancreatic fluid collections in 98 (18%) and pseudocysts in 19 (3.5%). Transient organ failure was present in 31 patients (5.7%) and persistent organ failure in 21 (3.9%). Sixteen (2.9%) patients died. Determinant-based and Revised Atlanta classifications were associated to statistically and clinically relevant differences in length of hospital stay, need 13 (2013) S2–S98 S67